Search Results

Search found 1771 results on 71 pages for 'knowing me knowing you'.

Page 63/71 | < Previous Page | 59 60 61 62 63 64 65 66 67 68 69 70  | Next Page >

  • Getting Started with Prism (aka Composite Application Guidance for WPF and Silverlight)

    - by dotneteer
    Overview Prism is a framework from the Microsoft Patterns and Practice team that allow you to create WPF and Silverlight in a modular way. It is especially valuable for larger projects in which a large number of developers can develop in parallel. Prism achieves its goal by supplying several services: · Dependency Injection (DI) and Inversion of control (IoC): By using DI, Prism takes away the responsibility of instantiating and managing the life time of dependency objects from individual components to a container. Prism relies on containers to discover, manage and compose large number of objects. By varying the configuration, the container can also inject mock objects for unit testing. Out of the box, Prism supports Unity and MEF as container although it is possible to use other containers by subclassing the Bootstrapper class. · Modularity and Region: Prism supplies the framework to split application into modules from the application shell. Each module is a library project that contains both UI and code and is responsible to initialize itself when loaded by the shell. Each window can be further divided into regions. A region is a user control with associated model. · Model, view and view-model (MVVM) pattern: Prism promotes the user MVVM. The use of DI container makes it much easier to inject model into view. WPF already has excellent data binding and commanding mechanism. To be productive with Prism, it is important to understand WPF data binding and commanding well. · Event-aggregation: Prism promotes loosely coupled components. Prism discourages for components from different modules to communicate each other, thus leading to dependency. Instead, Prism supplies an event-aggregation mechanism that allows components to publish and subscribe events without knowing each other. Architecture In the following, I will go into a little more detail on the services provided by Prism. Bootstrapper In a typical WPF application, application start-up is controls by App.xaml and its code behind. The main window of the application is typically specified in the App.xaml file. In a Prism application, we start a bootstrapper in the App class and delegate the duty of main window to the bootstrapper. The bootstrapper will start a dependency-injection container so all future object instantiations are managed by the container. Out of box, Prism provides the UnityBootstrapper and MefUnityBootstrapper abstract classes. All application needs to either provide a concrete implementation of one of these bootstrappers, or alternatively, subclass the Bootstrapper class with another DI container. A concrete bootstrapper class must implement the CreateShell method. Its responsibility is to resolve and create the Shell object through the DI container to serve as the main window for the application. The other important method to override is ConfigureModuleCatalog. The bootstrapper can register modules for the application. In a more advance scenario, an application does not have to know all its modules at compile time. Modules can be discovered at run time. Readers to refer to one of the Open Modularity Quick Starts for more information. Modules Once modules are registered with or discovered by Prism, they are instantiated by the DI container and their Initialize method is called. The DI container can inject into a module a region registry that implements IRegionViewRegistry interface. The module, in its Initialize method, can then call RegisterViewWithRegion method of the registry to register its regions. Regions Regions, once registered, are managed by the RegionManager. The shell can then load regions either through the RegionManager.RegionName attached property or dynamically through code. When a view is created by the region manager, the DI container can inject view model and other services into the view. The view then has a reference to the view model through which it can interact with backend services. Service locator Although it is possible to inject services into dependent classes through a DI container, an alternative way is to use the ServiceLocator to retrieve a service on demard. Prism supplies a service locator implementation and it is possible to get an instance of the service by calling: ServiceLocator.Current.GetInstance<IServiceType>() Event aggregator Prism supplies an IEventAggregator interface and implementation that can be injected into any class that needs to communicate with each other in a loosely-coupled fashion. The event aggregator uses a publisher/subscriber model. A class can publishes an event by calling eventAggregator.GetEvent<EventType>().Publish(parameter) to raise an event. Other classes can subscribe the event by calling eventAggregator.GetEvent<EventType>().Subscribe(EventHandler, other options). Getting started The easiest way to get started with Prism is to go through the Prism Hands-On labs and look at the Hello World QuickStart. The Hello World QuickStart shows how bootstrapper, modules and region works. Next, I would recommend you to look at the Stock Trader Reference Implementation. It is a more in depth example that resemble we want to set up an application. Several other QuickStarts cover individual Prism services. Some scenarios, such as dynamic module discovery, are more advanced. Apart from the official prism document, you can get an overview by reading Glen Block’s MSDN Magazine article. I have found the best free training material is from the Boise Code Camp. To be effective with Prism, it is important to understands key concepts of WPF well first, such as the DependencyProperty system, data binding, resource, theme and ICommand. It is also important to know your DI container of choice well. I will try to explorer these subjects in depth in the future. Testimony Recently, I worked on a desktop WPF application using Prism. I had a wonderful experience with Prism. The Prism is flexible enough even in the presence of third party controls such as Telerik WPF controls. We have never encountered any significant obstacle.

    Read the article

  • SPARC M7 Chip - 32 cores - Mind Blowing performance

    - by Angelo-Oracle
    The M7 Chip Oracle just announced its Next Generation Processor at the HotChips HC26 conference. As the Tech Lead in our Systems Division's Partner group, I had a front row seat to the extraordinary price performance advantage of Oracle current T5 and M6 based systems. Partner after partner tested  these systems and were impressed with it performance. Just read some of the quotes to see what our partner has been saying about our hardware. We just announced our next generation processor, the M7. This has 32 cores (up from 16-cores in T5 and 12-cores in M6). With 20 nm technology  this is our most advanced processor. The processor has more cores than anything else in the industry today. After the Sun acquisition Oracle has released 5 processors in 4 years and this is the 6th.  The S4 core  The M7 is built using the foundation of the S4 core. This is the next generation core technology. Like its predecessor, the S4 has 8 dynamic threads. It increases the frequency while maintaining the Pipeline depth. Each core has its own fine grain power estimator that keeps the core within its power envelop in 250 nano-sec granularity. Each core also includes Software in Silicon features for Application Acceleration Support. Each core includes features to improve Application Data Integrity, with almost no performance loss. The core also allows using part of the Virtual Address to store meta-data.  User-Level Synchronization Instructions are also part of the S4 core. Each core has 16 KB Instruction and 16 KB Data L1 cache. The Core Clusters  The cores on the M7 chip are organized in sets of 4-core clusters. The core clusters share  L2 cache.  All four cores in the complex share 256 KB of 4 way set associative L2 Instruction Cache, with over 1/2 TB/s of throughput. Two cores share 256 KB of 8 way set associative L2 Data Cache, with over 1/2 TB/s of throughput. With this innovative Core Cluster architecture, the M7 doubles core execution bandwidth. to maximize per-thread performance.  The Chip  Each  M7 chip has 8 sets of these core-clusters. The chip has 64 MB on-chip L3 cache. This L3 caches is shared among all the cores and is partitioned into 8 x 8 MB chunks. Each chunk is  8-way set associative cache. The aggregate bandwidth for the L3 cache on the chip is over 1.6TB/s. Each chip has 4 DDR4 memory controllers and can support upto 16 DDR4 DIMMs, allowing for 2 TB of RAM/chip. The chip also includes 4 internal links of PCIe Gen3 I/O controllers.  Each chip has 7 coherence links, allowing for 8 of these chips to be connected together gluelessly. Also 32 of these chips can be connected in an SMP configuration. A potential system with 32 chips will have 1024 cores and 8192 threads and 64 TB of RAM.  Software in Silicon The M7 chip has many built in Application Accelerators in Silicon. These features will be exposed to our Software partners using the SPARC Accelerator Program.  The M7  has built-in logic to decompress data at the speed of memory access. This means that applications can directly work on compressed data in memory increasing the data access rates. The VA Masking feature allows the use of part of the virtual address to store meta-data.  Realtime Application Data Integrity The Realtime Application Data Integrity feature helps applications safeguard against invalid, stale memory reference and buffer overflows. The first 4-bits if the Pointer can be used to store a version number and this version number is also maintained in the memory & cache lines. When a pointer accesses memory the hardware checks to make sure the two versions match. A SEGV signal is raised when there is a mismatch. This feature can be used by the Database, applications and the OS.  M7 Database In-Memory Query Accelerator The M7 chip also includes a In-Silicon Query Engines.  These accelerate tasks that work on In-Memory Columnar Vectors. Oracle In-Memory options stores data in Column Format. The M7 Query Engine can speed up In-Memory Format Conversion, Value and Range Comparisons and Set Membership lookups. This engine can work on Compressed data - this means not only are we accelerating the query performance but also increasing the memory bandwidth for queries.  SPARC Accelerated Program  At the Hotchips conference we also introduced the SPARC Accelerated Program to provide our partners and third part developers access to all the goodness of the M7's SPARC Application Acceleration features. Please get in touch with us if you are interested in knowing more about this program. 

    Read the article

  • What Counts For a DBA: Simplicity

    - by Louis Davidson
    Too many computer processes do an apparently simple task in a bizarrely complex way. They remind me of this strip by one of my favorite artists: Rube Goldberg. In order to keep the boss from knowing one was late, a process is devised whereby the cuckoo clock kisses a live cuckoo bird, who then pulls a string, which triggers a hat flinging, which in turn lands on a rod that removes a typewriter cover…and so on. We rely on creating automated processes to keep on top of tasks. DBAs have a lot of tasks to perform: backups, performance tuning, data movement, system monitoring, and of course, avoiding being noticed.  Every day, there are many steps to perform to maintain the database infrastructure, including: checking physical structures, re-indexing tables where needed, backing up the databases, checking those backups, running the ETL, and preparing the daily reports and yes, all of these processes have to complete before you can call it a day, and probably before many others have started that same day. Some of these tasks are just naturally complicated on their own. Other tasks become complicated because the database architecture is excessively rigid, and we often discover during “production testing” that certain processes need to be changed because the written requirements barely resembled the actual customer requirements.   Then, with no time to change that rigid structure, we are forced to heap layer upon layer of code onto the problematic processes. Instead of a slight table change and a new index, we end up with 4 new ETL processes, 20 temp tables, 30 extra queries, and 1000 lines of SQL code.  Report writers then need to build reports and make magical numbers appear from those toxic data structures that are overly complex and probably filled with inconsistent data. What starts out as a collection of fairly simple tasks turns into a Goldbergian nightmare of daily processes that are likely to cause your dinner to be interrupted by the smartphone doing the vibration dance that signifies trouble at the mill. So what to do? Well, if it is at all possible, simplify the problem by either going into the code and refactoring the complex code to simple, or taking all of the processes and simplifying them into small, independent, easily-tested steps.  The former approach usually requires an agreement on changing underlying structures that requires countless mind-numbing meetings; while the latter can generally be done to any complex process without the same frustration or anger, though it will still leave you with lots of steps to complete, the ability to test each step independently will definitely increase the quality of the overall process (and with each step reporting status back, finding an actual problem within the process will be definitely less unpleasant.) We all know the principle behind simplifying a sequence of processes because we learned it in math classes in our early years of attending school, starting with elementary school. In my 4 years (ok, 9 years) of undergraduate work, I remember pretty much one thing from my many math classes that I apply daily to my career as a data architect, data programmer, and as an occasional indentured DBA: “show your work”. This process of showing your work was my first lesson in simplification. Each step in the process was in fact, far simpler than the entire process.  When you were working an equation that took both sides of 4 sheets of paper, showing your work was important because the teacher could see every step, judge it, and mark it accordingly.  So often I would make an error in the first few lines of a problem which meant that the rest of the work was actually moving me closer to a very wrong answer, no matter how correct the math was in the subsequent steps. Yet, when I got my grade back, I would sometimes be pleasantly surprised. I passed, yet missed every problem on the test. But why? While I got the fact that 1+1=2 wrong in every problem, the teacher could see that I was using the right process. In a computer process, the process is very similar. We take complex processes, show our work by storing intermediate values, and test each step independently. When a process has 100 steps, each step becomes a simple step that is tested and verified, such that there will be 100 places where data is stored, validated, and can be checked off as complete. If you get step 1 of 100 wrong, you can fix it and be confident (that if you did your job of testing the other steps better than the one you had to repair,) that the rest of the process works. If you have 100 steps, and store the state of the process exactly once, the resulting testable chunk of code will be far more complex and finding the error will require checking all 100 steps as one, and usually it would be easier to find a specific needle in a stack of similarly shaped needles.  The goal is to strive for simplicity either in the solution, or at least by simplifying every process down to as many, independent, testable, simple tasks as possible.  For the tasks that really can’t be done completely independently, minimally take those tasks and break them down into simpler steps that can be tested independently.  Like working out division problems longhand, have each step of the larger problem verified and tested.

    Read the article

  • Subscribable World Cup 2010 Calendar

    - by jamiet
    I bang on quite a lot on this blog about ways in which data can get published over the web and one of the most interesting ways, in my opinion, of publishing data in a structured manner that is well understood is to use the iCalendar specification. There isn’t much information in the world that doesn’t have some concept of “when” so iCalendar is a great way of distributing that information. You have probably used iCalendar at some point without even knowing about it. All files with a .ics suffix are iCalendar format files and that is why you can happily import them into Outlook, Hotmail Calendar, Google Calendar etc… where they can be parsed and have the semantic data (when, where and who) extracted from them. Importing of iCalendar format data is really only half the trick though; in my opinion the real value of iCalendar-formatted calendar is the ability to subscribe to them. Subscribing has a simple benefit over importing but that single benefit is of massive importance: a subscriber to an iCalendar calendar can periodically check to see if any updates have been made and, if they have, automatically update the local copy. The real benefit to the user is the productivity gain – a single update to an iCalendar means that all subscribers are automatically made aware of the change and there is zero effort on the part of the subscriber; as my former colleague Howard van Rooijen is fond of saying, “work smarter not harder” – nowhere is this edict more ably demonstrated than subscribing versus importing of calendars. If you want to read some more thoughts about iCalendar then go and read my past blog post Calendar syndication - My big hope for 2009's breakthrough technology or better still go and seek out Jon Udell who speaks very authoritatively on the issue of iCalendar. With this subject of iCalendar on my mind I was interested to discover (via Steve Clayton’s blog post Download the world cup fixtures) that the BBC had made a .ics file available containing all of the matches in the upcoming World Cup. As you can probably guess this was a file that was made available so that it could be imported into your calendar of choice. It had one obvious downside though, right now nobody knows who is going to be playing in the knock-out stages so the calendar looks like this: with no teams being named after 25th June. How much more useful would this calendar have been if the BBC had made it possible to subscribe to the calendar instead, thus the calendar could be updated with the teams for the knock out stages when they are known and every subscriber would have a permanently up-to-date record of all the fixtures in their calendar. Better still, the calendar could be updated with match results as well or perhaps even post a match report from the BBC sport pages; when calendars are made subscribable a sea of opportunity opens up for distribution of information. So with that in mind I have decided to go one better than the BBC. I have imported their .ics into a brand new Hotmail calendar and made it publicly available at the following URLs: HTML http://cid-dc1ed121af0476be.calendar.live.com/calendar/World+Cup+2010/index.html iCalendar webcal://cid-dc1ed121af0476be.calendar.live.com/calendar/World+Cup+2010/calendar.ics The link you’re really interested in is the second one - click on that and it should open up in your calendar software of choice. Or, if you want to view it in an online calendar such as Hotmail Calendar or Google Calendar, copy and paste that URL into the appropriate place. Some people have told me they’re having trouble with the iCalendar link in which case hit the HTML link and then click “View ICS” at the resultant web page: I shall endeavour to keep the calendar updated throughout the World Cup and even if I don’t you’re no worse off than if you had imported the BBC’s .ics file so why not give it a try? If I do keep it up to date then you will have a permanent record of the 2010 World Cup available in your calendar. Forever. If you have your calendar synced to your smartphone then you’ll be carrying match reports around with you without you having to do a single thing. Surely that’s worth a quick click isn’t it?   If you have any thoughts let me have them in the comments below. Thanks for reading. @Jamiet Share this post: email it! | bookmark it! | digg it! | reddit! | kick it! | live it!

    Read the article

  • SSAS Maestro Training in July 2012 #ssasmaestro #ssas

    - by Marco Russo (SQLBI)
    A few hours ago Chris Webb blogged about SSAS Maestro and I’d like to propagate the news, adding also some background info. SSAS Maestro is the premier certification on Analysis Services that selects the best experts in Analysis Services around the world. In 2011 Microsoft organized two rounds of training/exams for SSAS Maestros and up to now only 11 people from the first wave have been announced – around 10% of attendees of the course! In the next few days the new Maestros from the second round should be announced and this long process is caused by many factors that I’m going to explain. First, the course is just a step in the process. Before the course you receive a list of topics to study, including the slides of the course. During the course, students receive a lot of information that might not have been included in the slides and the best part of the course is class interaction. Students are expected to bring their experience to the table and comparing case studies, experiences and having long debates is an important part of the learning process. And it is also a part of the evaluation: good questions might be also more important than good answers! Finally, after the course, students have their homework and this may require one or two months to be completed. After that, a long (very long) evaluation process begins, taking into account homework, labs, participation… And for this reason the final evaluation may arrive months later after the course. We are going to improve and shorten this process with the next courses. The first wave of SSAS Maestro had been made by invitation only and now the program is opening, requiring a fee to participate in order to cover the cost of preparation, training and exam. The number of attendees will be limited and candidates will have to send their CV in order to be admitted to the course. Only experienced Analysis Services developers will be able to participate to this challenging program. So why you should do that? Well, only 10% of students passed the exam until now. So if you need 100% guarantee to pass the exam, you need to study a lot, before, during and after the course. But the course by itself is a precious opportunity to share experience, create networking and learn mission-critical enterprise-level best practices that it’s hard to find written on books. Oh, well, many existing white papers are a required reading *before* the course! The course is now 5 days long, and every day can be *very* long. We’ll have lectures and discussions in the morning and labs in the afternoon/evening. Plus some more lectures in one or two afternoons. A heavy part of the course is about performance optimization, capacity planning, monitoring. This edition will introduce also Tabular models, and don’t expect something you might find in the SSAS Tabular Workshop – only performance, scalability monitoring and optimization will be covered, knowing Analysis Services is a requirement just to be accepted! I and Chris Webb will be the teachers for this edition. The course is expensive. Applying for SSAS Maestro will cost around 7000€ plus taxes (reduced to 5000€ for students of a previous SSAS Maestro edition). And you will be locked in a training room for the large part of the week. So why you should do that? Well, as I said, this is a challenging course. You will not find the time to check your email – the content is just too much interesting to think you can be distracted by something else. Another good reason is that this course will take place in Italy. Well, the course will take place in the brand new Microsoft Innovation Campus, but in general we’ll be able to provide you hints to get great food and, if you are willing to attach one week-end to your trip, there are plenty of places to visit (and I’m not talking about the classic Rome-Florence-Venice) – you might really need to relax after such a week! Finally, the marking process after the course will be faster – we’d like to complete the evaluation within three months after the course, considering that 1-2 months might be required to complete the homework. If at this point you are not scared: registration will open in mid-April, but you can already write to [email protected] sending your CV/resume and a short description of your level of SSAS knowledge and experience. The selection process will start early and you may want to put your admission form on top of the FIFO queue!

    Read the article

  • Delving into design patterns, and what that means for the Oracle user experience

    - by Kathy.Miedema
    By Kathy Miedema, Oracle Applications User Experience George Hackman, Senior Director, Applications User Experiences The Oracle Applications User Experience team has some exciting things happening around Fusion Applications design patterns. Because we’re hoping to have some new offerings soon (stay tuned with VoX to see what’s in the pipeline around Fusion Applications design patterns), now is a good time to talk more about what design patterns can do for the individual user as well as the entire company. George Hackman, Senior Director of Operations User Experience, says the first thing to note is that user experience is not just about the user interface. It’s about understanding how people do things, observing them, and then finding the patterns that emerge. The Applications UX team develops those patterns and then builds them into Oracle applications. What emerges, Hackman says, is a consistent, efficient user experience that promotes a productive workplace. Creating design patterns What is a design pattern in the context of enterprise software? “Every day, people use technology to get things done,” Hackman says. “They navigate a virtual world that reaches from enterprise to consumer apps, and from desktop to mobile. This virtual world is constantly under construction. New areas are being developed and old areas are being redone. As this world is being built and remodeled, efficient pathways and practices emerge. “Oracle's user experience team watches users navigate this world. We measure their productivity and ask them about their satisfaction. We take the most efficient, most productive pathways from the enterprise and consumer world and turn them into Oracle's user experience patterns.” Hackman describes the process as combining all of the best practices from every part of a user’s world. Members of the user experience team observe, analyze, design, prototype, and measure each work task to find the best possible pattern for a particular work flow. As the team builds the patterns, “we make sure they are fully buildable using Oracle technology,” Hackman said. “So customers know they can use these patterns. There’s no need to make something up from scratch, not knowing whether you can even build it.” Hackman says that creating something on a computer is a good example of a user experience pattern. “People are creating things all the time,” he says. “On the consumer side, they are creating documents. On the enterprise side, they are creating expense reports. On a mobile phone, they are creating contacts. They are using different apps like iPhone or Facebook or Gmail or Oracle software, all doing this creation process.” The Applications UX team starts their process by observing how people might create something. “We observe people creating things. We see the patterns, we analyze and document, then we apply them to our products. It might be different from phone to web browser, but we have these design patterns that create a consistent experience across platforms, and across products, too. The result for customers Oracle constantly improves its part of the virtual world, Hackman said. New products are created and existing products are upgraded. Because Oracle builds user experience design patterns, Oracle's virtual world becomes both more powerful and more familiar at the same time. Because of design patterns, users can navigate with ease as they embrace the latest technology – because it behaves the way they expect it to. This means less training and faster adoption for individual users, and more productivity for the business as a whole. Hackman said Oracle gives customers and partners access to design patterns so that they can build in the virtual world using the same best practices. Customers and partners can extend applications with a user experience that is comfortable and familiar to their users. For businesses that are integrating different Oracle applications, design patterns are key. The user experience created in E-Business Suite should be similar to the user experience in Fusion Applications, Hackman said. If a user is transitioning from one application to the other, it shouldn’t be difficult for them to do their work. With design patterns, it isn’t. “Oracle user experience patterns are the building blocks for the virtual world that ensure productivity, consistency and user satisfaction,” Hackman said. “They are built for the enterprise, but incorporate the best practices from across the virtual world. They empower productivity and facilitate social interaction. When you build with patterns, you get all the end-user benefits of less training / retraining from the finished product. You also get faster / cheaper development.” What’s coming? You can already access design patterns to help you build Dashboards with OBIEE here. And we promised you at the beginning that we had something in the pipeline on Fusion Applications design patterns. Look for the announcement about when they are available here on VoX.

    Read the article

  • A little primer on using TFS with a small team

    - by johndoucette
    The scenario; A small team of 3 developers mostly in maintenance mode with traditional ASP.net, classic ASP, .Net integration services and utilities with the company’s third party packages, and a bunch of java-based Coldfusion web applications all under Visual Source Safe (VSS). They are about to embark on a huge SharePoint 2010 new construction project and wanted to use subversion instead VSS. TFS was a foreign word and smelled of “high cost” and of an “over complicated process”. Since they had no preconditions about the old TFS versions (‘05 & ‘08), it was fun explaining how simple it was to install a TFS server and get the ball rolling, with or without all the heavy stuff one sometimes associates with such a huge and powerful application management lifecycle product. So, how does a small team begin using TFS? 1. Start by using source control and migrate current VSS source trees into TFS. You can take the latest version or migrate the entire version history. It’s up to you on whether you want a clean start or need quick access to all the version notes and history of the bits. 2. Since most shops are mainly in maintenance mode with existing applications, begin using bug workitems for everything. When you receive an issue/bug from your current tracking system, manually enter the workitem in TFS right through Visual Studio. You can automate the integration to the current tracking system later or replace it entirely. Believe me, this thing is powerful and can handle even the largest of help desks. 3. With new construction, begin work with requirements and task workitems and follow the traditional sprint-based development lifecycle. Obviously, some minor training will be needed, but don’t fear, this is very intuitive and MSDN has a ton of lesson based labs and videos. 4. For the java developers, use the new Team Explorer Everywhere 2010 plugin (recently known as Teamprise). There is a seamless interface in Eclipse, but also a good command-line utility for other environments such as Dreamweaver. 5. Wait to fully integrate the whole workitem/project management/testing process until your team is familiar with the integrated workitems for bugs and code. After a while, you will see the team wanting more transparency into the work they are all doing and naturally, everyone will want workitems to help them organize the chaos! 6. Management will be limited in the value of the reports until you have a fully blown implementation of project planning, construction, build, deployment and testing. However, there are some basic “bug rate” reports and current backlog listings that can provide good information. Some notable explanations of TFS; Work Item Tracking and Project Management - A workitem represents the unit of work within the system which enables tracking of all activities produced by a user, whether it is a developer, business user, project manager or tester. The properties of a workitem such as linked changesets (checked-in code), who updated the data and when, the states and reasons for change, are all transitioned to a data warehouse within TFS for reporting purposes. A workitem can be defines as a "bug", "requirement", test case", or a "change request". They drive the work effort by the individual assigned to it and also provide a key role in defining what needs to be done. Workitems are the things the team needs to do to accomplish a goal. Test Case Management - Starting with a workitem known as a "test case", a tester (or developer) can now author and manage test cases within a formal test plan subsystem. Although TFS supports the test case workitem type, there is a new product known as the VS Test Professional 2010 which allows a tester to facilitate manual tests including fast forwarding steps in the process to arrive at the assertion point quickly. This repeatable process provides quick regression tests and can be conducted by the business user to ensure completeness during UAT. In addition, developers no longer can provide a response to a bug with the line "cannot reproduce". With every test run, attachments including the recorded session, captured environment configurations and settings, screen shots, intellitrace (debugging history), and in some cases if the lab manager is being used, a snapshot of the tested environment is available. Version Control - A modern system allowing shared check-in/check-out, excellent merge conflict resolution, Shelvesets (personal check-ins), branching/merging visualization, public workspaces, gated check-ins, security hierarchy capabilities, and changeset/workitem tracking. Knowing what was done with the code by any developer has become much easier to picture and resolve issues. Team Build - Automate the compilation process whether you need it to be whenever a developer checks-in code, periodically such as nightly builds for testers in the morning, or manual builds to be deployed into production. Each build can run through pre-determined tests, perform code analysis to see if the developer conforms to the team standards, and reject the build if either fails. Project Portal & Reporting - Provide management with a dashboard with insight into the project(s). "Where are we" in each step of the way including past iterations and the current burndown rate. Enabling this feature is easy as it seamlessly interfaces with existing SharePoint implementations.

    Read the article

  • Interview with Geoff Bones, developer on SQL Storage Compress

    - by red(at)work
    How did you come to be working at Red Gate? I've been working at Red Gate for nine months; before that I had been at a multinational engineering company. A number of my colleagues had left to work at Red Gate and spoke very highly of it, but I was happy in my role and thought, 'It can't be that great there, surely? They'll be back!' Then one day I visited to catch up them over lunch in the Red Gate canteen. I was so impressed with what I found there, that, three days later, I'd applied for a role as a developer. And how did you get into software development? My first job out of university was working as a systems programmer on IBM mainframes. This was quite a while ago: there was a lot of assembler and loading programs from tape drives and that kind of stuff. I learned a lot about how computers work, and this stood me in good stead when I moved over the development in the 90s. What's the best thing about working as a developer at Red Gate? Where should I start? One of the great things as a developer at Red Gate is the useful feedback and close contact we have with the people who use our products, either directly at trade shows and other events or through information coming through the product managers. The company's whole ethos is built around assisting the user, and this is in big contrast to my previous development roles. We aim to produce tools that people really want to use, that they enjoy using, and, as a developer, this is a great thing to aim for and a great feeling when we get it right. At Red Gate we also try to cut out the things that distract and stop us doing our jobs. As a developer, this means that I can focus on the code and the product I'm working on, knowing that others are doing a first-class job of making sure that the builds are running smoothly and that I'm getting great feedback from the testers. We keep our process light and effective, as we want to produce great software more than we want to produce great audit trails. Tell us a bit about the products you are currently working on. You mean HyperBac? First let me explain a bit about what HyperBac is. At heart it's a compression and encryption technology, but with a few added features that open up a wealth of really exciting possibilities. Right now we have the HyperBac technology in just three products: SQL HyperBac, SQL Virtual Restore and SQL Storage Compress, but we're only starting to develop what it can do. My personal favourite is SQL Virtual Restore; for example, I love the way you can use it to run independent test databases that are all backed by a single compressed backup. I don't think the market yet realises the kind of things you do once you are using these products. On the other hand, the benefits of SQL Storage Compress are straightforward: run your databases but use only 20% of the disk space. Databases are getting larger and larger, and, as they do, so does your ROI. What's a typical day for you? My days are pretty varied. We have our daily team stand-up meeting and then sometimes I will work alone on a current issue, or I'll be pair programming with one of my colleagues. From time to time we give half a day up to future planning with the team, when we look at the long and short term aims for the product and working out the development priorities. I also get to go to conferences and events, which is unusual for a development role and gives me the chance to meet and talk to our customers directly. Have you noticed anything different about developing tools for DBAs rather than other IT kinds of user? It seems to me that DBAs are quite independent minded; they know exactly what the problem they are facing is, and often have a solution in mind before they begin to look for what's on the market. This means that they're likely to cherry-pick tools from a range of vendors, picking the ones that are the best fit for them and that disrupt their environments the least. When I've met with DBAs, I've often been very impressed at their ability to summarise their set up, the issues, the obstacles they face when implementing a tool and their plans for their environment. It's easier to develop products for this audience as they give such a detailed overview of their needs, and I feel I understand their problems.

    Read the article

  • Must-see sessions at TCUK11

    - by Roger Hart
    Technical Communication UK is probably the best professional conference I've been to. Last year, I spoke there on content strategy, and this year I'll be co-hosting a workshop on embedded user assistance. Obviously, I'd love people to come along to that; but there are some other sessions I'd like to flag up for anybody thinking of attending. Tuesday 20th Sept - workshops This will be my first year at the pre-conference workshop day, and I'm massively glad that our workshop hasn't been scheduled along-side the one I'm really interested in. My picks: It looks like you're embedding user assistance. Would you like help? My colleague Dom and I are presenting this one. It's our paen to Clippy, to the brilliant idea he represented, and the crashing failure he was. Less precociously, we'll be teaching embedded user assistance, Red Gate style. Statistics without maths: acquiring, visualising and interpreting your data This doesn't need to do anything apart from what it says on the tin in order to be gold dust. But given the speakers, I suspect it will. A data-informed approach is a great asset to technical communications, so I'd recommend this session to anybody event faintly interested. The speakers here have a great track record of giving practical, accessible introductions to big topics. Go along. Wednesday 21st Sept - day one There's no real need to recommend the keynote for a conference, but I will just point out that this year it's Google's Patrick Hofmann. That's cool. You know what else is cool: Focus on the user, the rest follows An intro to modelling customer experience. This is a really exciting area for tech comms, and potentially touches on one of my personal hobby-horses: the convergence of technical communication and marketing. It's all part of delivering customer experience, and knowing what your users need lets you help them, sell to them, and delight them. Content strategy year 1: a tale from the trenches It's often been observed that content strategy is great at banging its own drum, but not so hot on compelling case studies. Here you go, folks. This is the presentation I'm most excited about so far. On a mission to communicate! Skype help their users communicate, but how do they communicate with them? I guess we'll find out. Then there's the stuff that I'm not too excited by, but you might just be. The standards geeks and agile freaks can get together in a presentation on the forthcoming ISO standards for agile authoring. Plus, there's a session on VBA for tech comms. I do have one gripe about day 1. The other big UK tech comms conference, UA Europe, have - I think - netted the more interesting presentation from Ellis Pratt. While I have no doubt that his TCUK case study on producing risk assessments will be useful, I'd far rather go to his talk on game theory for tech comms. Hopefully UA Europe will record it. Thursday 22nd Sept - day two Day two has a couple of slots yet to be confirmed. The rumour is that one of them will be the brilliant "Questions and rants" session from last year. I hope so. It's not ranting, but I'll be going to: RTFMobile: beyond stating the obvious Ultan O'Broin is an engaging speaker with a lot to say, and mobile is one of the most interesting and challenging new areas for tech comms. Even if this weren't a research-based presentation from a company with buckets of technology experience, I'd be going. It is, and you should too. Pattern recognition for technical communicators One of the best things about TCUK is the tendency to include sessions that tackle the theoretical and bring them towards the practical. Kai and Chris delivered cracking and well-received talks last year, and I'm looking forward to seeing what they've got for us on some of the conceptual underpinning of technical communication. Developing an interactive non-text learning programme Annoyingly, this clashes with Pattern Recognition, so I hope at least one of the streams is recorded again this year. The idea of communicating complex information without words us fascinating and this sounds like a great example of this year's third stream: "anything but text". For the localization and DITA crowds, there's rich pickings on day two, though I'm not sure how many of those sessions I'm interested in. In the 13:00 - 13:40 slot, there's an interesting clash between Linda Urban on re-use and training content, and a piece on minimalism I'm sorely tempted by. That's my pick of #TCUK11. I'll be doing a round-up blog after the event, and probably talking a bit more about it beforehand. I'm also reliably assured that there are still plenty of tickets.

    Read the article

  • Some mail details about Orange Mauritius

    Being an internet service provider is not easy after all for a lot of companies. Luckily, there are quite some good international operators in this world. For example Orange Mauritius aka Mauritius Telecom aka Wanadoo(?) aka MyT here in Mauritius. The local circumstances give them a quasi-monopol position on fixed lines for telephony and therefore cable-based DSL internet connectivity. So far, not bad but as usual... the details. Just for the records, I am only using the services of Orange for mobile but friends and customers are bound, eh stuck, with other services of Orange Mauritius. And usually, being the IT guy, they get in touch with me to complain about problems or to ask questions on either their ADSL / MyT connection, mail services or whatever. Most of those issues are user-related and easily to solve by tweaking the configuration of their computer a little bit but sometimes it's getting weird. Using Orange ADSL... somewhere else Now, let's imagine we are an Orange ADSL customer for ages and we are using their mail services with our very own mail address like "[email protected]". We configured our mail client like Thunderbird, Outlook Express, Outlook or Windows Mail as publicly described, and we are able to receive and send emails like a champion. No problems at all, the world is green. Did I mention that we have a laptop? Ok, let's take our movable piece of information technology and visit a friend here on the island. Not surprising, he is also customer of Orange, so we can read and answer emails. But Orange is not the online internet service provider and one day, we happen to hang out with someone that uses Emtel via WiMAX or UMTS.. And the fun starts... We can still receive and read emails from our Orange mail account and the IT world is still bright but try to send mails to someone outside the domain "@intnet.mu" or "@orange.mu". Your mail client will deny sending mail with SMTP message 5.1.0 "blah not allowed". First guess, there is problem with the mail client, maybe magically the configuration changed over-night. But no it is still working at home... So, there is for sure a problem with the guy's internet connection. At least, it is his fault not to have Orange internet services, so it can not work properly... The Orange Mail FAQ After some more frustation we finally checkout the Orange Mail FAQ to see whether this (obviously?) common problem has been described already. Sorry, but those FAQ entries are even more confusing as it is not really clear how to handle this scenario. Best of all is that most of the entries are still refering to use servers of the domain "intnet.mu". I mean Orange will disable those systems in favour of the domain "orange.mu" in the near future and does not amend their FAQs. Come on, guys! Ok, settings for POP3 are there. Hm, what about the secure version POP3S? No signs at all... Even changing your mail client to use password encryption with STARTTLS is not allowed at all. Use "bow.intnet.mu" for incoming mail... Ahhh, pretty obvious host name. I mean, at least something like pop.intnet.mu or pop3.intnet.mu would have been more accurate. Funny of all, the hostname "pop.orange.mu" is accessible to receive your mail account. Alright, checking SMTP options for authentication or other like POP-before-SMTP or whatever well-known and established mechanism to send emails are described. I guess that spotting a whale or shark in Mauritian waters would be easier. Trial and error on SMTP settings reveal that neither STARTTLS or any other connection / password encryption is available. Using SSL/TLS on SMTP only reveals that there is no service answering your request. Calling customer service So, we have to bite into the bitter apple and get in touch with Orange customer service and complain/explain them our case and ask for advice. After some hiccups, we finally manage to get hold of someone competent in mail services and we receive the golden spoon of mail configuration made by Orange Mauritius: SMTP hostname: smtpauth.intnet.mu And the world of IT is surprisingly green again. Customer satisfaction? Dear Orange Mauritius, what's the problem with this information? Are you scared of mail spammer? Why isn't there any case in your FAQs? Ok, talking about your FAQs - simply said: they are badly outdated! Configure your mail client to use server name based in the domain intnet.mu but specify your account username with orange.mu as domain part. Although, that there are servers available on the domain orange.mu after all. So, why don't you provide current information like this: POP3 server name: pop.orange.muSMTP server name: smtp.orange.muSMTP authenticated: smtpauth.orange.mu It's not difficult, is it? In my humble opinion not really and you would provide clean, consistent and up-to-date information for your customers. This would produce less frustation and so less traffic on your customer service lines. Which after all, would improve the total user experience and satisfaction level on both sides. Without knowing these facts. Now, imagine you would take your laptop abroad and have to use other internet service providers to be able to be online... Calling your customer service would be unnecessary expensive!

    Read the article

  • When does "proper" programming no longer matter?

    - by Kai Qing
    I've been a full time programmer for about 8 years now. Web based mostly, ranging in weird jobs for clients. Never anything I "want" to do. So my experience is limited to what I've been contracted to do, having no real incentive to master anything in particular. So here's my scenario and ultimately what I wonder about... I've been building an android game in my spare time. It's using the libgdx library so quite a bit of the heavy lifting is done for me. I don't read much of the docs cause unless it's in tutorial format I will just not care, and ultimately most of my questions have already been asked on stackoverflow. I get along fine and my game works as expected... Suspiciously well, even. So much so that I wonder why one should bother to be "proper" when coding if the end result is ultimately the same. To be more specific, I used a hashtable because I wanted something close to an associative array. Human readable key values. In other places to achieve similar things, I use a vector. I know libgdx has vector2 and vector3 classes, but I've never used them. When I come across weird problems and search stackoverflow for help, I see a lot of people just reaming the questions that use a certain datatype when another one is technically "proper." Like using an ArrayList because it does not require defined bounds versus re-defining an int[] with new known boundaries. Or even something trivial like this: for(int i = 0; i < items.length; i ++) { // do something } I know it evaluates item.length on every iteration. I just don't care. I know items will never be more than 15 to 20 items. So why bother caring if I evaluate items.length on every iteration? So I wonder - why does everyone get all up in arms over this? Who cares if I use a less efficient datatype to get the job done? I ran some tests to see how the app performs using the lazy, get it done fast and don't look back method I just described versus the proper, follow the tutorial and use the exact data types suggested by the community. The results: Same thing. Average 45 fps. I opened every app on the phone and galaxy tab. Same deal. No difference. My game is pretty graphic intensive. It's not like it's just a simple thing. I expected it to perform kind of badly since I don't care to optimize image assets or... well, you probably get the idea. I'm making the game for fun. As a joke, really. But in doing so I'm working outside the normal scope of my job, which is to always follow the rules and do it the right way. So to say, I am without bounds here and this has caused me to wonder why I ever really care to be "proper" So I guess my question to you is this: Is there a threshold when it no longer matters to be proper? Is there a lasting, longer term consequence to the lazy, get it done and don't look back route? Is it ok to say - "so long as it gets the job done, I don't care?" Disclaimer: When I program my game, I am almost always drunk. I do it to remember why I got into this stuff to begin with because the monotony of client based web work will make you hate being a programmer. I'm having a blast and my game is not crashing, tests well, performs well, looks good on all devices so far and has no noticeable negative impact on any of my testing devices. I expected failure because I was being so drunkenly careless with my code, but to my surprise, it had no noticeable impact. I am now starting to question the need to be careful. Help me regain the ability to care! ... or explain why it's not a bad thing to not care. Secondary disclaimer: I am aware of the benefits of maintainability. For myself and others. Agreed. But it's not like someone happening across my inefficient int[] loop won't know what it does. As an experienced programmer those kinds of things are just clear on sight. I document the complex stuff for myself knowing I was drunk and will probably need a reminder. Those notes would clarify any confusion for someone who might ever gaze upon my ridiculous game - though the reality is that either I maintain it myself or it fades into time. I'm ok with that. But if it doesn't slow the device down, or crash, then crossing the t's and dotting the i's might actually require more time than it's worth.

    Read the article

  • Summit reflections

    - by Rob Farley
    So far, my three PASS Summit experiences have been notably different to each other. My first, I wasn’t on the board and I gave two regular sessions and a Lightning Talk in which I told jokes. My second, I was a board advisor, and I delivered a precon, a spotlight and a Lightning Talk in which I sang. My third (last week), I was a full board director, and I didn’t present at all. Let’s not talk about next year. I’m not sure there are many options left. This year, I noticed that a lot more people recognised me and said hello. I guess that’s potentially because of the singing last year, but could also be because board elections can bring a fair bit of attention, and because of the effort I’ve put in through things like 24HOP... Yeah, ok. It’d be the singing. My approach was very different though. I was watching things through different eyes. I looked for the things that seemed to be working and the things that didn’t. I had staff there again, and was curious to know how their things were working out. I knew a lot more about what was going on behind the scenes to make various things happen, and although very little about the Summit was actually my responsibility (based on not having that portfolio), my perspective had moved considerably. Before the Summit started, Board Members had been given notebooks – an idea Tom (who heads up PASS’ marketing) had come up with after being inspired by seeing Bill walk around with a notebook. The plan was to take notes about feedback we got from people. It was a good thing, and the notebook forms a nice pair with the SQLBits one I got a couple of years ago when I last spoke there. I think one of the biggest impacts of this was that during the first keynote, Bill told everyone present about the notebooks. This set a tone of “we’re listening”, and a number of people were definitely keen to tell us things that would cause us to pull out our notebooks. PASSTV was a new thing this year. Justin, the host, featured on the couch and talked a lot of people about a lot of things, including me (he talked to me about a lot of things, I don’t think he talked to a lot people about me). Reaching people through online methods is something which interests me a lot – it has huge potential, and I love the idea of being able to broadcast to people who are unable to attend in person. I’m keen to see how this medium can be developed over time. People who know me will know that I’m a keen advocate of certification – I've been SQL certified since version 6.5, and have even been involved in creating exams. However, I don’t believe in studying for exams. I think training is worthwhile for learning new skills, but the goal should be on learning those skills, not on passing an exam. Exams should be for proving that the skills are there, not a goal in themselves. The PASS Summit is an excellent place to take exams though, and with an attitude of professional development throughout the event, why not? So I did. I wasn’t expecting to take one, but I was persuaded and took the MCM Knowledge Exam. I hadn’t even looked at the syllabus, but tried it anyway. I was very tired, and even fell asleep at one point during it. I’ll find out my result at some point in the future – the Prometric site just says “Tested” at the moment. As I said, it wasn’t something I was expecting to do, but it was good to have something unexpected during the week. Of course it was good to catch up with old friends and make new ones. I feel like every time I’m in the US I see things develop a bit more, with more and more people knowing who I am, who my staff are, and recognising the LobsterPot brand. I missed being a presenter, but I definitely enjoyed seeing many friends on the list of presenters. I won’t try to list them, because there are so many these days that people might feel sad if I don’t mention them. For those that I managed to see, I was pleased to see that the majority of them have lifted their presentation skills since I last saw them, and I happily told them as much. One person who I will mention was Paul White, who travelled from New Zealand to his first PASS Summit. He gave two sessions (a regular session and a half-day), packed large rooms of people, and had everyone buzzing with enthusiasm. I spoke to him after the event, and he told me that his expectations were blown away. Paul isn’t normally a fan of crowds, and the thought of 4000 people would have been scary. But he told me he had no idea that people would welcome him so well, be so friendly and so down to earth. He’s seen the significance of the SQL Server community, and says he’ll be back. It’ll be good to see him there. Will you be there too?

    Read the article

  • How Estimates Became Quotes

    - by Lee Brandt
    It’s our fault. Well, not completely, but we haven’t helped the situation any. All of what follows comes from my own experiences which, from talking to lots of other developers about it, seems to be pretty much par for the course. Where We Started When we first started estimating, we estimated pretty clearly. We would try to imagine something we’d done that was similar to the project being estimated and we’d toss it about in our heads a bit and see how much bigger or smaller we thought this new thing was, and add or subtract accordingly. We wouldn’t spend too much time on it, because we wanted to get to writing the software. Eventually, we’d come across some huge problem that there was now way we could’ve known about ahead of time. Either we didn’t see this thing or, we didn’t realize that this particular version of a problem would be so… problematic. We usually call this “not knowing what we don’t know”. It’s unavoidable. We just can’t know. Until we wade in and start putting some code together, there are just some things we won’t know… and some things we don’t even know that we don’t know. Y’know? So what happens? We go over budget. Project managers scream and dance the dance of the stressed-out project manager, and there is nothing we can do (or could’ve done) about it. We didn’t know. We thought about it for a bit and we didn’t see this herculean task sitting in the middle of our nice quiet project, and it has bitten us in the rear end. We now know how to handle this in the future, though. We will take some more time to pick around the requirements and discover all those things we don’t know. We’ll do some prototyping, we’ll read some blogs about similar projects, we’ll really grill the customer with questions during the requirements gathering phase. We’ll keeping asking “what else?” until the shove us down the stairs. We’ll take our time and uncover it all. We Learned, But Good The next time comes, and you know what happens? We do it. We grill the customer for weeks and prototype and read and research and we estimate everything down to the last button on the last form. Know what that gets us? It gets us three months of wasted time, and our estimate will still be off. Possibly off by a factor of four. WTF, mate? No way we could be surprised by something! We uncovered every particle. We turned every stone. How is it we still came across unknowns? Because we STILL didn’t know what we didn’t know. How could we? We didn’t know to ask. The worst part is, we’ve now convinced the product that this is NOT an estimate. It is a solid number based on massive research and an endless number of questions that they answered. There is absolutely now way you don’t know everything there is to know about this project now. No way there is anything you haven’t uncovered. And their faith in that “Esti-Quote” goes through the roof. When the project goes over this time, they might even begin to question whether or not you know what you’re doing. Who could blame them? You drilled them for weeks about every little thing, and when they complained about all the questions, you told them you wanted to uncover everything so there would be no surprises. SO we set them up to faile Guess, Then Plan We had a chance. At the beginning we could have just said, “That’s just a gut-feeling estimate, based on my past experience with similar projects. There could still be surprises.” If we spend SOME time doing SOME discovery and then bounce that against our own past experiences, we can come up with a fairly healthy estimate. We can then help the product owner understand that an estimate is a guess. Sure, it’s an educated guess, but it is still a guess. If we get it right it will be almost completely luck. Then, we help them to plan the development by taking that guess (yes, they still need the guess for planning purposes) and start measuring early and often to see if we still think we are right. We should adjust the estimate and alert the product owner as soon as we see problems (bad news does not age well) and we should be able to see any problems immediately if we are constantly measuring our pace. In lean software, we start with that guess and begin measuring cycle times immediately. Then we can make projections based on those cycle times and compare them to our estimate. This constant feedback is the best way to ensure that there are no surprises at the END of the project. There sill still be surprises, but we’ll see them sooner and have a better understanding of how they will affect our overall timeline. What do you think?

    Read the article

  • Application Development: Python or Java (or PHP)

    - by luckysmack
    I'm looking to get into application development, such as Facebook or Android apps and games. I am doing this for fun and to learn. Once my skills are to par I would like to have some side income from the apps, but I'm not banking on living off that (just so you know where I'm coming from and know what my end goals are). Currently I know and am familiar with PHP and frameworks such as cakephp and yii. However, I have been wanting to learn another language to broaden my horizons and to become a better developer. So I have narrowed it down to 2 languages. Python, and Java (I can already hear people cringing at the difference in the languages I have chosen, but I have some reasons). Python: closer to PHP that Java. Cross platformability. Also great as a general scripting language and has many file system level benefits that PHP does not. Cleaner syntax, readability, blah blah and the list goed on. Python will work great for cross platform apps and can be run on many OS's and is supported by Facebook for app development. But there is no support on Android (for full fledged apps). Java: a much stronger typed language, very robust community and corporate backing. Knowing Java is also good for personal marketability for enterprises, if you're into that. The main benefit here is that Java can write apps natively for Android and the apps can be ported for web versions to play on Facebook. So while I have seen many developers prefer Java over the two, Java has this significant advantage, where I can market my apps in both markets and in the future build more potential income. But like I said it is for fun. While money isn't the goal, it would still be nice. PHP: I'm putting this here because I know it already, and I'm sure a case could be made for it. It obviously works great for Facebook but like Python does not do so well on android. While it's mostly the realm of 'application development' that appeals to me, I do find Android apps fairly interesting and something that has a ton of potential to. But then again Facebook has a ton more users and the apps can also potentially be more immersive (desktop vs. mobile). So this is why I'm kinda stuck on what route to choose. Python for Facebook and web apps, with likely faster development to production times, or Java which can be developed for any of the platforms to make apps. Side note: I'm not really trying to get into 3D development, mostly 2D. And I also want to make an app with real-time play (websockets, etc). Someone mentioned node, js to me for that but Python seems to be more globally versatile for my goals. So, to anyone that does Facebook or Android development in either language: what do you suggest? Any input is valuable and I do appreciate it. And sorry for being long winded. EDIT: as mentioned in one of the answers, my primary goal is gaming. Although I do have some plans for non gaming apps such as general web based and desktop based ones. But gaming is my main goal with the possibility of income. EDIT: Another consideration could be Jython. Writing Python code which is converted into Java bytecode. This would allow the ability to do Android apps using Python. I could be wrong though, I'm still looking into it. Update 1-26-11: I recently acquired a new job which required I learn .NET using C#. Im sure some of you are cringing already but I really like the whole system and how it all works together between desktop and web development. But, as I am still interested in Python very much, and after some research I have decided I will learn Python as well as the IronPython implementation for .NET. But (again: I know...) since .NET is mostly a Windows thing and not as cross-compatible as I like, I will be learning Mono which is a cross platform implementation of .NET where I can use what I learn at work using C# and what I want to learn, Python/IronPython. So while learning and writing C#/.NET @ work I will be learning Python - Mono - Iron Python for what I want to do personally. And the benefit of them all being very closely related will help me out a lot, I think. What do you guys think? I almost feel like that should be another question, but there's not much of a question. Either way, you guys gave very helpful input.

    Read the article

  • The Linux powered LAN Gaming House

    - by sachinghalot
    LAN parties offer the enjoyment of head to head gaming in a real-life social environment. In general, they are experiencing decline thanks to the convenience of Internet gaming, but Kenton Varda is a man who takes his LAN gaming very seriously. His LAN gaming house is a fascinating project, and best of all, Linux plays a part in making it all work.Varda has done his own write ups (short, long), so I'm only going to give an overview here. The setup is a large house with 12 gaming stations and a single server computer.The client computers themselves are rack mounted in a server room, and they are linked to the gaming stations on the floor above via extension cables (HDMI for video and audio and USB for mouse and keyboard). Each client computer, built into a 3U rack mount case, is a well specced gaming rig in its own right, sporting an Intel Core i5 processor, 4GB of RAM and an Nvidia GeForce 560 along with a 60GB SSD drive.Originally, the client computers ran Ubuntu Linux rather than Windows and the games executed under WINE, but Varda had to abandon this scheme. As he explains on his site:"Amazingly, a majority of games worked fine, although many had minor bugs (e.g. flickering mouse cursor, minor rendering artifacts, etc.). Some games, however, did not work, or had bad bugs that made them annoying to play."Subsequently, the gaming computers have been moved onto a more conventional gaming choice, Windows 7. It's a shame that WINE couldn't be made to work, but I can sympathize as it's rare to find modern games that work perfectly and at full native speed. Another problem with WINE is that it tends to suffer from regressions, which is hardly surprising when considering the difficulty of constantly improving the emulation of the Windows API. Varda points out that he preferred working with Linux clients as they were easier to modify and came with less licensing baggage.Linux still runs the server and all of the tools used are open source software. The hardware here is a Intel Xeon E3-1230 with 4GB of RAM. The storage hanging off this machine is a bit more complex than the clients. In addition to the 60GB SSD, it also has 2x1TB drives and a 240GB SDD.When the clients were running Linux, they booted over PXE using a toolchain that will be familiar to anyone who has setup Linux network booting. DHCP pointed the clients to the server which then supplied PXELINUX using TFTP. When booted, file access was accomplished through network block device (NBD). This is a very easy to use system that allows you to serve the contents of a file as a block device over the network. The client computer runs a user mode device driver and the device can be mounted within the file system using the mount command.One snag with offering file access via NBD is that it's difficult to impose any security restrictions on different areas of the file system as the server only sees a single file. The advantage is perfomance as the client operating system simply sees a block device, and besides, these security issues aren't relevant in this setup.Unfortunately, Windows 7 can't use NBD, so, Varda had to switch to iSCSI (which works in both server and client mode under Linux). His network cards are not compliant with this standard when doing a netboot, but fortunately, gPXE came to the rescue, and he boostraps it over PXE. gPXE is also available as an ISO image and is worth knowing about if you encounter an awkward machine that can't manage a network boot. It can also optionally boot from a HTTP server rather than the more traditional TFTP server.According to Varda, booting all 12 machines over the Gigabit Ethernet network is surprisingly fast, and once booted, the machines don't seem noticeably slower than if they were using local storage. Once loaded, most games attempt to load in as much data as possible, filling the RAM, and the the disk and network bandwidth required is small. It's worth noting that these are aspects of this project that might differ from some other thin client scenarios.At time of writing, it doesn't seem as though the local storage of the client machines is being utilized. Instead, the clients boot into Windows from an image on the server that contains the operating system and the games themselves. It uses the copy on write feature of LVM so that any writes from a client are added to a differencing image allocated to that client. As the administrator, Varda can log into the Linux server and authorize changes to the master image for updates etc.SummaryOverall, Varda estimates the total cost of the project at about $40,000, and of course, he needed a property that offered a large physical space in order to house the computers and the gaming workstations. Obviously, this project has stark differences to most thin client projects. The balance between storage, network usage, GPU power and security would not be typical of an office installation, for example. The only letdown is that WINE proved to be insufficiently compatible to run a wide variety of modern games, but that is, perhaps, asking too much of it, and hats off to Varda for trying to make it work.

    Read the article

  • Some New .NET Downloads and Resources

    - by Kevin Grossnicklaus
    Last week I was fortunate enough to spend time in Redmond on Microsoft’s campus for the 2011 Microsoft MVP Summit.  It was great to hang out with a number of old friends and get the opportunity to talk tech with the various product teams up at Microsoft.  The weather wasn’t exactly sunny but Microsoft always does a great job with the Summit and everyone had a blast (heck, I even got to run the bases at SafeCo field) While much of what we saw is covered under NDA, there a ton of great things in the pipeline from Microsoft and many things that are already available (or just became so) that I wasn’t necessarily aware of.  The purpose of this post is to share some of the info I learned on resources and tools available to .NET developers today.  Please let me know if you have any questions (or if you know of something else cool which might benefit others). Enjoy! Visual Studio 2010 SP1 Microsoft has issued the RTM release of Visual Studio 2010 SP1.  You can download the full SP1 on MSDN as of today (March 10th to the general public) and take advantage of such things as: Silverlight 4 is included in the box (as opposed to a separate install) Silverlight 4 Profiling WCF RIA Services SP1 Intellitrace for 64-bit and SharePoint ASP.NET now easily supports IIS Express and SQL CE Want a description of all that’s new beyond the above biased list (which arguably only contains items I think are important)?  Check out this KB article. Portable Library Tools CTP Without much fanfare Microsoft has released a CTP of a new add-in to Visual Studio 2010 which simplifies code sharing between projects targeting different runtimes (i.e. Silverlight, WPF, Win7 Phone, XBox).   With this Add-In installed you can add a new project of type “Portable Library” and specify which platforms you wish to target.  Once that is done, any code added to this library will be limited to use only features which are common to all selected frameworks.  Other projects can now reference this portable library and be provided assemblies custom built to their environment.  This greatly simplifies the current process of sharing linked files between platforms like WPF and Silverlight.  You can find out more about this CTP and how it works on this great blog post. Visual Studio Async CTP Microsoft has also released a CTP of a set of language and framework enhancements to provide a much more powerful asynchronous programming model.   Due to the focus on async programming in all types of platforms (and it being the ONLY option in Silverlight and Win7 phone) a move towards a simpler and more understandable model is always a good thing. This CTP (called Visual Studio Async CTP) can be downloaded here.  You can read more about this CTP on this blog post. MSDN Code Samples Gallery Microsoft has also launched new code samples gallery on their MSDN site: http://code.msdn.microsoft.com/.   This site allows you to easily search for small samples of code related to a particular technology or platform.  If a sample of code you are looking for is not found, you can request one via the site and other developers can see your request and provide a sample to the site to suit your needs.  You can also peruse requested samples and, if you find a scenario where you can provide value, upload your own sample for the benefit of others.  Samples are packaged into the VS .vsix format and include any necessary references/dependencies.  By using .vsix as the deployment mechanism, as samples are installed from the site they are kept in your Visual Studio 2010 Samples Gallery and kept for your future reference. If you get a chance, check out the site and see how it is done.  Although a somewhat simple concept, I was very impressed with their implementation and the way they went about trying to suit a need.  I’ll definitely be looking there in the future as need something or want to share something. MSDN Search Capabilities Another item I learned recently and was not aware of (that might seem trivial to some) is the power of the MSDN site’s search capabilities.  Between the Code Samples Gallery described above and the search enhancements on MSDN, Microsoft is definitely investing in their platform to help provide developers of all skill levels the tools and resources they need to be successful. What do I mean by the MSDN search capability and why should you care? If you go to the MSDN home page (http://msdn.microsoft.com) and use the “Search MSDN with Big” box at the very top of the page you will see some very interesting results.  First, the search actually doesn’t just search the MSDN library it searches: MSDN Library All Microsoft Blogs CodePlex StackOverflow Downloads MSDN Magazine Support Knowledgebase (I’m not sure it even ends there but the above are all I know of) Beyond just searching all the above locations, the results are formatted very nicely to give some contextual information based on where the result came from.  For example, if a keyword search returned results from CodePlex, each row in the search results screen would include a large amount of information specific to CodePlex such as: Looking at the above results immediately tells you everything from the page views to the CodePlex ratings.  All in all, knowing that this much information is indexed and available from a single search location will lead me to utilize this as one of my initial searches for development information.

    Read the article

  • Disaster Recovery Discovery

    - by Rodney Landrum
    Last weekend I joined several of my IT staff on a mission to perform a DR test in our remote CoLo center in a large South East city of the US. Can I be more obtuse? The goal was simple for me as the sole DBA in a throng of Windows, Storage, Network and SAN admins – restore the databases and make them work. There were 4 applications that back ended to 7 SQL Server databases on 4 different SQL Server instances. We would maintain the original server names, but beyond that it was fair game. We had time to prepare so I was able to script out or otherwise automate the recovery process. I used sp_help_revlogin for three of the servers, a bit of a cheat actually because restoring the Master database on the target DR servers was the specified course of action according to the DR procedures ( the caveat “IF REQUIRED” left it open to interpretation. I really wanted to avoid the step of restoring Master for a number of reasons but mainly because I did not want to deal with issues starting SQL Services afterward. Having to account for the location of TempDB and the version conflicts of the resource DBs were just two of the battles I chose not to fight. Not to mention other system database location problems that might arise and prevent SQL from starting.  I was going to have to restore all of the user databases anyway, so I would not really gain any benefit, outside of logins, for taking the time to restore the source Master database over the newly installed one on the fresh server. What I wanted was the ability to restore the Master database as a user database, call it Master_Mine, from a backup on the source system and then use that restored database to script the SQL Logins and passwords on the DR systems. While I did not attempt this on the trip, the thought stuck in my mind and this past week I succeeded at scripting user accounts and passwords using only a restored copy of the Master database. Granted there were several challenges to overcome.  Also, as is usual for any work like this the usual disclaimers apply:  This is not something that I would imagine Microsoft would condone or support and this was really only an experiment for me to learn if it was even possible. While I have tested the process with success, I do not know that I would use this technique in a documented procedure because future updates for SQL Server will render this technique non-functional. I thought at first, incorrectly of course, that I could use sp_help_revlogin on a restored copy of the master database I named Master_Mine.   Since sp_help_revlogin uses system schema objects, sys.syslogins and sys.server_principals, this was not going to work because all results would come from the main Master database. To test this I added a SQL login via SSMS, backed up Master, restored  it as Master_Mine, and then deleted the login.  Even though the test account I created should presumably still be in the Master_Mine database, I should be able to get to it and script out its creation with its password hash so that I would not need to know the password, but any applications that stored that password would not have to be altered in the DR scenario. They would just work as expected. Once I realized that would not work I began looking deeper.  Knowing that sys.syslogins and sys.server_principals are system views, their underlying code should be available with sp_helptext, right? They were. And this led me to discover the two tables sys.sysxlgns and sys.sysprivs, where the data I needed was stored. These tables existed in both the real Master and the restored copy, Master_Mine.  I used this information to tweak the sp_help_revlogin stored procedure to use these tables instead to create the logins cursor used in sp_help_revlogin. For the password hash,  sp_help_revlogin uses the function LoginProperty() which takes a user name and option ‘passwordhash’ to return the hash for the user. Unfortunately, it requires the login to exist in the Master database. This would not work. So another slight modification I had to make was to pull the password hash itself (pwdhash from sys.sysxlgns) into the logins cursor and comment out the section of sp_help_revlogin that uses LoginProperty. Instead, I pass the pwdhash value as the variable @PWD_varbinary to the sp_hexadecimal stored procedure which is also created by and used within the code provided by Microsoft in the link above for sp_help_revlogin. The final challenge: sys.sysxlgns and sys.server_principals are visible only within a Dedicated Administrator Connection (DAC) query window in SSMS or within SQLCDMD.  To open a DAC connection you have to be logged in on the SQL Server itself, via RDP in my case,  and you preface the server name in the query connection with ADMIN:, so that the server connection looks like ADMIN:ServerName. From there you can create the modified stored procedure in the restored copy of a Master database from a source system as whatever name you like, and then run the modified stored procedure. I named my new stored procedure usp_help_revlogin_MyMaster. Upon execution I was happy to see the logins and password hashes that I needed to apply from the source Master database without having to restore over the new Master system database and without the need to access the original server (assuming it was down due to whatever disaster put it in that state). You will note that I am not providing full code samples here of the modifications. I will say that it was a slight bit of work and anyone who needed to do this for whatever reason, could fairly easily roll their own solution with the information provided herein.  My goal, as I said was to prove that this could be done and provide another option if required to ease the burden of getting SQL Servers up and available in an emergency situation where alternatives may be more challenging or otherwise unavailable.  

    Read the article

  • Creating ADF Faces Comamnd Button at Runtime

    - by Frank Nimphius
    Normal 0 false false false EN-US X-NONE X-NONE /* Style Definitions */ table.MsoNormalTable {mso-style-name:"Table Normal"; mso-tstyle-rowband-size:0; mso-tstyle-colband-size:0; mso-style-noshow:yes; mso-style-priority:99; mso-style-qformat:yes; mso-style-parent:""; mso-padding-alt:0in 5.4pt 0in 5.4pt; mso-para-margin:0in; mso-para-margin-bottom:.0001pt; mso-pagination:widow-orphan; font-size:11.0pt; font-family:"Calibri","sans-serif"; mso-ascii-font-family:Calibri; mso-ascii-theme-font:minor-latin; mso-fareast-font-family:"Times New Roman"; mso-fareast-theme-font:minor-fareast; mso-hansi-font-family:Calibri; mso-hansi-theme-font:minor-latin; mso-bidi-font-family:"Times New Roman"; mso-bidi-theme-font:minor-bidi;} In ADF Faces, the command button is an instance of RichCommandButton and can be created at runtime. While creating the button is not difficult at all, adding behavior to it requires knowing about how to dynamically create and add an action listener reference. The example code below shows two methods: The first method, handleButtonPress is a public method exposed on a managed bean. public void handleButtonPress(ActionEvent event){   System.out.println("Event handled");   //optional: partially refresh changed components if command   //issued as a partial submit } The second method is called in response to a user interaction or on page load and dynamically creates and adds a command button. When the button is pressed, the managed bean method – the action handler – defined above is called. The action handler is referenced using EL in the created MethodExpression instance. If the managed bean is in viewScope, backingBeanScope or pageFlowsScope, then you need to add these scopes as a prefix to the EL (as you would when configuring the managed bean reference at design time) //Create command button and add it as a child to the parent component that is passed as an //argument to this method private void reateCommandButton(UIComponent parent){   RichCommandButton edit = new RichCommandButton();   //make the request partial   edit.setPartialSubmit(true);   edit.setText("Edit");                             //compose the method expression to invoke the event handler   FacesContext fctx = FacesContext.getCurrentInstance();   Application application = fctx.getApplication();   ExpressionFactory elFactory = application.getExpressionFactory();   ELContext elContext = facesCtx.getELContext();   MethodExpression methodExpressio = null;   //Make sure the EL expression references a valid managed bean method. Ensure   //the bean scope is properly addressed    methodExpression =  elFactory.createMethodExpression(                              elContext,"#{myRequestScopeBean.handleButtonPress}",                             Object.class,new Class[] {ActionEvent.class});   //Create the command buttonaction listener reference   MethodExpressionActionListener al = null;          al= new MethodExpressionActionListener(methodExpression);    edit.addActionListener(al);     //add new command button to parent component and PPR the component for     //the button to show    parent.getChildren().add(edit);    AdfFacesContext adfFacesContext = AdfFacesContext.getCurrentInstance();     adfFacesContext.addPartialTarget(parent);  }

    Read the article

  • yield – Just yet another sexy c# keyword?

    - by George Mamaladze
    yield (see NSDN c# reference) operator came I guess with .NET 2.0 and I my feeling is that it’s not as wide used as it could (or should) be.   I am not going to talk here about necessarity and advantages of using iterator pattern when accessing custom sequences (just google it).   Let’s look at it from the clean code point of view. Let's see if it really helps us to keep our code understandable, reusable and testable.   Let’s say we want to iterate a tree and do something with it’s nodes, for instance calculate a sum of their values. So the most elegant way would be to build a recursive method performing a classic depth traversal returning the sum.           private int CalculateTreeSum(Node top)         {             int sumOfChildNodes = 0;             foreach (Node childNode in top.ChildNodes)             {                 sumOfChildNodes += CalculateTreeSum(childNode);             }             return top.Value + sumOfChildNodes;         }     “Do One Thing” Nevertheless it violates one of the most important rules “Do One Thing”. Our  method CalculateTreeSum does two things at the same time. It travels inside the tree and performs some computation – in this case calculates sum. Doing two things in one method is definitely a bad thing because of several reasons: ·          Understandability: Readability / refactoring ·          Reuseability: when overriding - no chance to override computation without copying iteration code and vice versa. ·          Testability: you are not able to test computation without constructing the tree and you are not able to test correctness of tree iteration.   I want to spend some more words on this last issue. How do you test the method CalculateTreeSum when it contains two in one: computation & iteration? The only chance is to construct a test tree and assert the result of the method call, in our case the sum against our expectation. And if the test fails you do not know wether was the computation algorithm wrong or was that the iteration? At the end to top it all off I tell you: according to Murphy’s Law the iteration will have a bug as well as the calculation. Both bugs in a combination will cause the sum to be accidentally exactly the same you expect and the test will PASS. J   Ok let’s use yield! That’s why it is generally a very good idea not to mix but isolate “things”. Ok let’s use yield!           private int CalculateTreeSumClean(Node top)         {             IEnumerable<Node> treeNodes = GetTreeNodes(top);             return CalculateSum(treeNodes);         }             private int CalculateSum(IEnumerable<Node> nodes)         {             int sumOfNodes = 0;             foreach (Node node in nodes)             {                 sumOfNodes += node.Value;             }             return sumOfNodes;         }           private IEnumerable<Node> GetTreeNodes(Node top)         {             yield return top;             foreach (Node childNode in top.ChildNodes)             {                 foreach (Node currentNode in GetTreeNodes(childNode))                 {                     yield return currentNode;                 }             }         }   Two methods does not know anything about each other. One contains calculation logic another jut the iteration logic. You can relpace the tree iteration algorithm from depth traversal to breath trevaersal or use stack or visitor pattern instead of recursion. This will not influence your calculation logic. And vice versa you can relace the sum with product or do whatever you want with node values, the calculateion algorithm is not aware of beeng working on some tree or graph.  How about not using yield? Now let’s ask the question – what if we do not have yield operator? The brief look at the generated code gives us an answer. The compiler generates a 150 lines long class to implement the iteration logic.       [CompilerGenerated]     private sealed class <GetTreeNodes>d__0 : IEnumerable<Node>, IEnumerable, IEnumerator<Node>, IEnumerator, IDisposable     {         ...        150 Lines of generated code        ...     }   Often we compromise code readability, cleanness, testability, etc. – to reduce number of classes, code lines, keystrokes and mouse clicks. This is the human nature - we are lazy. Knowing and using such a sexy construct like yield, allows us to be lazy, write very few lines of code and at the same time stay clean and do one thing in a method. That's why I generally welcome using staff like that.   Note: The above used recursive depth traversal algorithm is possibly the compact one but not the best one from the performance and memory utilization point of view. It was taken to emphasize on other primary aspects of this post.

    Read the article

  • Project of Projects with team Foundation Server 2010

    - by Martin Hinshelwood
    It is pretty much accepted that you should use Areas instead of having many small Team Projects when you are using Team Foundation Server 2010. I have implemented this scenario many times and this is the current iteration of layout and considerations. If like me you work with many customers you will find that you get into a grove for how to set these things up to make them as easily understandable for everyone, while giving the best functionality. The trick is in making it as intuitive as possible for both you and the developers that need to work with it. There are five main places where you need to have the Product or Project name in prominence of any other value. Area Iteration Source Code Work Item Queries Build Once you decide how you are doing this in each of these places you need to keep to it religiously. Evan if you have one source code file to keep, make sure it is in the right place. This makes your developers and others working with the format familiar with where everything should go, as well as building up mussel memory. This prevents the neat system degenerating into a nasty mess. Areas Areas are traditionally used to separate out parts of your product / project so that you can see how much effort has gone into each. Figure: The top level areas are for reporting and work item separation There are massive advantages of using this method. You can: move work from one project to another rename a project / product It is far more likely that a project or product gets renamed than a department. Tip: If you have many projects, over 100, you should consider categorising them here, but make sure that the actual project name always sits at the same level so you know which is which. Figure: Always keep things that are the same at the same level Note: You may use these categories only at the Area/Iteration level to make it easier to select on drop down lists. You may not want to use them everywhere. On the other hand, for consistency it would be better to. Iterations Iterations are usually used to some sort of time based consideration. Here I am splitting into Iterations with periodic releases. Figure: Each product needs to be able to have its own cadence The ability to have each project run at its own pace and to enable them to have their own release schedule is often of paramount importance and you don’t want to fix your 100+ projects to all be released on the same date. Source Code Having a good structure for your source even if you are not branching or having multiple products under the same structure is always a good idea. Figure: Separate out your products source You need to think about both your branches as well as the structure of your source. All your code should be under “Source” and everything you need to build your solution including Build Scripts and 3rd party tools should be under your “Main” (branch) folder. This should them be branched by “Quality”, “Release” or both to get the most out of your branching structure. The important thing is to make sure you branch (or be able to branch) everything you need to build, test and deploy your application to an environment. That environment may be development, test or even production, but I can’t stress the importance of having everything your need. Note: You usually will not be able to install custom software on your build server. Store any *.dll’s or *.exe’s that you need under the “Tools\Tool1” folder. Note: Consult the Branching Guidance for Team Foundation Server 2010 for more on branching Figure: Adding category may be a necessary evil Even if you have to have a couple of categories called “Default”, it is better than not knowing the difference between a folder, Product and Branch. Work Item Queries Queries are used to load lists of Work Items out of TFS so you can see what work you have. This means that you want to also separate queries out by Product / project to make it easier to Figure: Again you have the same first level structure Having Folders also in Work Item Tracking we do the same thing. We put all the queries under a folder named for the Product / Project and change each query to have “AreaPath=[TeamProject]\[ProductX]” in the query instead of the standard “Project=@Project”. Tip: Don’t have a folder with new queries for each iteration. Instead have a single “Current” folder that has queries that point to the current iteration. Just change the queries as you move from one iteration to another. Tip: You can ctrl+drag the “Product1” folder to create your “Product2” folder. Builds You may have many builds both for individual products but also for different quality's. This can be further complicated by having some builds that action “Gated Check-In” and others that are specifically for “Release”, “Test” or another purpose. Figure: There are no folders, yet, for the builds so you need a good naming convention Its a pity that there are no folders under builds, some way to categorise would be nice. In lue of that at the moment you can use a functional naming convention that at least allows you to find what you want. Conclusion It is really easy to both achieve and to stick to this format if you take the time to do it. Unless you have 1000+ builds or 100+ Products you are unlikely run into any issues. Even then there are things you can do to mitigate the issues and I have describes some of them above. Let me know if you can think of any other things to make this easier.

    Read the article

  • Silverlight Relay Commands

    - by George Evjen
    v\:* {behavior:url(#default#VML);} o\:* {behavior:url(#default#VML);} w\:* {behavior:url(#default#VML);} .shape {behavior:url(#default#VML);} Normal 0 false false false false EN-US X-NONE X-NONE /* Style Definitions */ table.MsoNormalTable {mso-style-name:"Table Normal"; mso-tstyle-rowband-size:0; mso-tstyle-colband-size:0; mso-style-noshow:yes; mso-style-priority:99; mso-style-parent:""; mso-padding-alt:0in 5.4pt 0in 5.4pt; mso-para-margin:0in; mso-para-margin-bottom:.0001pt; mso-pagination:widow-orphan; font-size:11.0pt; font-family:"Calibri","sans-serif"; mso-ascii-font-family:Calibri; mso-ascii-theme-font:minor-latin; mso-hansi-font-family:Calibri; mso-hansi-theme-font:minor-latin; mso-bidi-font-family:"Times New Roman"; mso-bidi-theme-font:minor-bidi;} I am fairly new at Silverlight development and I usually have an issue that needs research every day. Which I enjoy, since I like the idea of going into a day knowing that I am  going to learn something new. The issue that I am currently working on centers around relay commands. I have a pretty good handle on Relay Commands and how we use them within our applications. <Button Command="{Binding ButtonCommand}" CommandParameter="NewRecruit" Content="New Recruit" /> Here in our xaml we have a button. The button has a Command and a CommandParameter. The command binds to the ButtonCommand that we have in our ViewModel RelayCommand _buttonCommand;         /// <summary>         /// Gets the button command.         /// </summary>         /// <value>The button command.</value>         public RelayCommand ButtonCommand         {             get             {                 if (_buttonCommand == null)                 {                     _buttonCommand = new RelayCommand(                         x => x != null && x.ToString().Length > 0 && CheckCommandAvailable(x.ToString()),                         x => ExecuteCommand(x.ToString()));                 }                 return _buttonCommand;             }         }   In our relay command we then do some checks with a lambda expression. We check if the command  parameter is null, is the length greater than 0 and we have a CheckCommandAvailable method that will tell  us if the button is even enabled. After we check on these three items we then pass the command parameter to an action method. This is all pretty straight forward, the issue that we solved a few days ago centered around having a control that needed to use a Relay Command and this control was a nested control and was using a different DataContext. The example below illustrates how we handled this scenario. In our xaml usercontrol we had to set a name to this control. <Controls3:RadTileViewItem x:Class="RecruitStatusTileView"     xmlns="http://schemas.microsoft.com/winfx/2006/xaml/presentation"     xmlns:x="http://schemas.microsoft.com/winfx/2006/xaml"     xmlns:d="http://schemas.microsoft.com/expression/blend/2008"     xmlns:mc="http://schemas.openxmlformats.org/markup-compatibility/2006"      xmlns:Controls1="clr-namespace:Telerik.Windows.Controls;assembly=Telerik.Windows.Controls"      xmlns:Controls2="clr-namespace:Telerik.Windows.Controls;assembly=Telerik.Windows.Controls.Input"      xmlns:Controls3="clr-namespace:Telerik.Windows.Controls;assembly=Telerik.Windows.Controls.Navigation"      mc:Ignorable="d" d:DesignHeight="400" d:DesignWidth="800" Header="{Binding Title,Mode=TwoWay}" MinimizedHeight="100"                             x:Name="StatusView"> Here we are using a telerik RadTileViewItem. We set the name of this control to “StatusView”. In our button control we set our command parameters and commands different than the example above. <HyperlinkButton Content="{Binding BigBoardButtonText, Mode=TwoWay}" CommandParameter="{Binding 'Position.PositionName'}" Command="{Binding ElementName=StatusView, Path=DataContext.BigBoardCommand, Mode=TwoWay}" /> This hyperlink button lives in a ListBox control and this listbox has an ItemSource of PositionSelectors. The Command Parameter is binding to the Position.Position property of that PositionSelectors object. This again is pretty straight forward again. What gets a bit tricky is the Command property in the hyperlink. It is binding to the element name we created in the user control (StatusView) Because this hyperlink is in a listbox and is in the item template it doesn’t have a direct handle on the DataContext that the RadTileViewItem has so we have to make sure it does. We do that by binding to the element name of status view then set the path to DataContext.BigBoardCommand. BigBoardCommand is the name of the RelayCommand in the view model. private RelayCommand _bigBoardCommand = null;         /// <summary>         /// Gets the big board command.         /// </summary>         /// <value>The big board command.</value>         public RelayCommand BigBoardCommand         {             get             {                 if (_bigBoardCommand == null)                 {                     _bigBoardCommand = new RelayCommand(x => true, x => AddToBigBoard(x.ToString()));                 }                 return _bigBoardCommand;             }         } From there we check for true again and then call the action and pass in the parameter that we had as the command parameter. What we are working on now is a bit trickier than this second example. In the above example we are only creating this TileViewItem with this name “StatusView” once. In another part of our application we are generating multiple TileViewItems, so we cannot set the name in the control as we cant have multiple controls with the same name. When we run the application we get an error that reads that the value is out of expected range. My searching has led me to think we cannot have multiple controls with the same name. This is today’s problem and Ill post the solution to this once it is found.

    Read the article

  • Ransomware: Why This New Malware is So Dangerous and How to Protect Yourself

    - by Chris Hoffman
    Ransomware is a type of malware that tries to extort money from you. One of the nastiest examples, CryptoLocker, takes your files hostage and holds them for ransom, forcing you to pay hundreds of dollars to regain access. Most malware is no longer created by bored teenagers looking to cause some chaos. Much of the current malware is now produced by organized crime for profit and is becoming increasingly sophisticated. How Ransomware Works Not all ransomware is identical. The key thing that makes a piece of malware “ransomware” is that it attempts to extort a direct payment from you. Some ransomware may be disguised. It may function as “scareware,” displaying a pop-up that says something like “Your computer is infected, purchase this product to fix the infection” or “Your computer has been used to download illegal files, pay a fine to continue using your computer.” In other situations, ransomware may be more up-front. It may hook deep into your system, displaying a message saying that it will only go away when you pay money to the ransomware’s creators. This type of malware could be bypassed via malware removal tools or just by reinstalling Windows. Unfortunately, Ransomware is becoming more and more sophisticated. One of the latest examples, CryptoLocker, starts encrypting your personal files as soon as it gains access to your system, preventing access to the files without knowing the encryption key. CryptoLocker then displays a message informing you that your files have been locked with encryption and that you have just a few days to pay up. If you pay them $300, they’ll hand you the encryption key and you can recover your files. CryptoLocker helpfully walks you through choosing a payment method and, after paying, the criminals seem to actually give you a key that you can use to restore your files. You can never be sure that the criminals will keep their end of the deal, of course. It’s not a good idea to pay up when you’re extorted by criminals. On the other hand, businesses that lose their only copy of business-critical data may be tempted to take the risk — and it’s hard to blame them. Protecting Your Files From Ransomware This type of malware is another good example of why backups are essential. You should regularly back up files to an external hard drive or a remote file storage server. If all your copies of your files are on your computer, malware that infects your computer could encrypt them all and restrict access — or even delete them entirely. When backing up files, be sure to back up your personal files to a location where they can’t be written to or erased. For example, place them on a removable hard drive or upload them to a remote backup service like CrashPlan that would allow you to revert to previous versions of files. Don’t just store your backups on an internal hard drive or network share you have write access to. The ransomware could encrypt the files on your connected backup drive or on your network share if you have full write access. Frequent backups are also important. You wouldn’t want to lose a week’s worth of work because you only back up your files every week. This is part of the reason why automated back-up solutions are so convenient. If your files do become locked by ransomware and you don’t have the appropriate backups, you can try recovering them with ShadowExplorer. This tool accesses “Shadow Copies,” which Windows uses for System Restore — they will often contain some personal files. How to Avoid Ransomware Aside from using a proper backup strategy, you can avoid ransomware in the same way you avoid other forms of malware. CryptoLocker has been verified to arrive through email attachments, via the Java plug-in, and installed on computers that are part of the Zeus botnet. Use a good antivirus product that will attempt to stop ransomware in its tracks. Antivirus programs are never perfect and you could be infected even if you run one, but it’s an important layer of defense. Avoid running suspicious files. Ransomware can arrive in .exe files attached to emails, from illicit websites containing pirated software, or anywhere else that malware comes from. Be alert and exercise caution over the files you download and run. Keep your software updated. Using an old version of your web browser, operating system, or a browser plugin can allow malware in through open security holes. If you have Java installed, you should probably uninstall it. For more tips, read our list of important security practices you should be following. Ransomware — CryptoLocker in particular — is brutally efficient and smart. It just wants to get down to business and take your money. Holding your files hostage is an effective way to prevent removal by antivirus programs after it’s taken root, but CryptoLocker is much less scary if you have good backups. This sort of malware demonstrates the importance of backups as well as proper security practices. Unfortunately, CryptoLocker is probably a sign of things to come — it’s the kind of malware we’ll likely be seeing more of in the future.     

    Read the article

  • Windows for IoT, continued

    - by Valter Minute
    Originally posted on: http://geekswithblogs.net/WindowsEmbeddedCookbook/archive/2014/08/05/windows-for-iot-continued.aspxI received many interesting feedbacks on my previous blog post and I tried to find some time to do some additional tests. Bert Kleinschmidt pointed out that pins 2,3 and 10 of the Galileo are connected directly to the SOC, while pin 13, the one used for the sample sketch is controlled via an I2C I/O expander. I changed my code to use pin 2 instead of 13 (just changing the variable assignment at the beginning of the code) and latency was greatly reduced. Now each pulse lasts for 1.44ms, 44% more than the expected time, but ways better that the result we got using pin 13. I also used SetThreadPriority to increase the priority of the thread that was running the sketch to THREAD_PRIORITY_HIGHEST but that didn't change the results. When I was using the I2C-controlled pin I tried the same and the timings got ways worse (increasing more than 10 times) and so I did not commented on that part, wanting to investigate the issua a bit more in detail. It seems that increasing the priority of the application thread impacts negatively the I2C communication. I tried to use also the Linux-based implementation (using a different Galileo board since the one provided by MS seems to use a different firmware) and the results of running the sample blink sketch modified to use pin 2 and blink the led for 1ms are similar to those we got on the same board running Windows. Here the difference between expected time and measured time is worse, getting around 3.2ms instead of 1 (320% compared to 150% using Windows but far from the 100.1% we got with the 8-bit Arduino). Both systems were not under load during the test, maybe loading some applications that use part of the CPU time would make those timings even less reliable, but I think that those numbers are enough to draw some conclusions. It may not be worth running a full OS if what you need is Arduino compatibility. The Arduino UNO is probably the best Arduino you can find to perform this kind of development. The Galileo running the Linux-based stack or running Windows for IoT is targeted to be a platform for "Internet of Things" devices, whatever that means. At the moment I don't see the "I" part of IoT. We have low level interfaces (SPI, I2C, the GPIO pins) that can be used to connect sensors but the support for connectivity is limited and the amount of work required to deliver some data to the cloud (using a secure HTTP request or a message queuing system like APMQS or MQTT) is still big and the rich OS underneath seems to not provide any help doing that.Why should I use sockets and can't access all the high level connectivity features we have on "full" Windows?I know that it's possible to use some third party libraries, try to build them using the Windows For IoT SDK etc. but this means re-inventing the wheel every time and can also lead to some IP concerns if used for products meant to be closed-source. I hope that MS and Intel (and others) will focus less on the "coolness" of running (some) Arduino sketches and more on providing a better platform to people that really want to design devices that leverage internet connectivity and the cloud processing power to deliver better products and services. Providing a reliable set of connectivity services would be a great start. Providing support for .NET would be even better, leaving native code available for hardware access etc. I know that those components may require additional storage and memory etc. So making the OS componentizable (or, at least, provide a way to install additional components) would be a great way to let developers pick the parts of the system they need to develop their solution, knowing that they will integrate well together. I can understand that the Arduino and Raspberry Pi* success may have attracted the attention of marketing departments worldwide and almost any new development board those days is promoted as "XXX response to Arduino" or "YYYY alternative to Raspberry Pi", but this is misleading and prevents companies from focusing on how to deliver good products and how to integrate "IoT" features with their existing offer to provide, at the end, a better product or service to their customers. Marketing is important, but can't decide the key features of a product (the OS) that is going to be used to develop full products for end customers integrating it with hardware and application software. I really like the "hackable" nature of open-source devices and like to see that companies are getting more and more open in releasing information, providing "hackable" devices and supporting developers with documentation, good samples etc. On the other side being able to run a sketch designed for an 8 bit microcontroller on a full-featured application processor may sound cool and an easy upgrade path for people that just experimented with sensors etc. on Arduino but it's not, in my humble opinion, the main path to follow for people who want to deliver real products.   *Shameless self-promotion: if you are looking for a good book in Italian about the Raspberry Pi , try mine: http://www.amazon.it/Raspberry-Pi-alluso-Digital-LifeStyle-ebook/dp/B00GYY3OKO

    Read the article

  • Lessons learnt in implementing Scrum in a Large Organization that has traditional values

    - by MarkPearl
    I recently had the experience of being involved in a “test” scrum implementation in a large organization that was used to a traditional project management approach. Here are some lessons that I learnt from it. Don’t let the Project Manager be the Product Owner First lesson learnt is to identify the correct product owner – in this instance the product manager assumed the role of the product owner which was a mistake. The product owner is the one who has the most to loose if the project fails. With a methodology that advocates removing the role of the project manager from the process then it is not in the interests of the person who is employed as a project manager to be the product owner – in fact they have the most to gain should the project fail. Know the time commitments of team members to the Project Second lesson learnt is to get a firm time commitment of the members on a team for the sprint and to hold them to it. In this project instance many of the issues we faced were with team members having to double up on supporting existing projects/systems and the scrum project. In many situations they just didn’t get round to doing any work on the scrum project for several days while they tried to meet other commitments. Initially this was not made transparent to the team – in stand up team members would say that had done some work but would be very vague on how much time they had actually spent using the blackhole of their other legacy projects as an excuse – putting up a time burn down chart made time allocations transparent and easy to hold the team to. In addition, how can you plan for a sprint without knowing the actual time available of the members – when I mean actual time, the exercise of getting them to go through all their appointments and lunch times and breaks and removing them from their time commitment helps get you to a realistic time that they can dedicate. Make sure you meet your minimum team sizes In a recent post I wrote about the difference between a partnership and a team. If you are going to do scrum in a large organization make sure you have a minimum team size of at least 3 developers. My experience with larger organizations is that people have a tendency to be sick more, take more leave and generally not be around – if you have a team size of two it is so easy to loose momentum on the project – the more people you have in the team (up to about 9) the more the momentum the project will have when people are not around. Swapping from one methodology to another can seem as waste to the customer It sounds bad, but most customers don’t care what methodology you use. Often they have bought into the “big plan upfront”. If you can, avoid taking a project on midstream from a traditional approach unless the customer has not bought into the process – with this particular project they had a detailed upfront planning breakaway with the customer using the traditional approach and then before the project started we moved onto a scrum implementation – this seemed as waste to the customer. We should have managed the customers expectation properly. Don’t play the role of the scrum master if you can’t be the scrum master With this particular implementation I was the “scrum master”. But all I did was go through the process of the formal meetings of scrum – I attended stand up, retrospectives and planning – but I was not hands on the ground. I was not performing the most important role of removing blockages – and by the end of the project there were a number of blockages “cropping up”. What could have been a better approach was to take someone on the team and train them to be the scrum master and be present to coach them. Alternatively actually be on the team on a fulltime basis and be the scrum master. By just going through the meetings of scrum didn’t mean we were doing scrum. So we failed with this one, if you fail look at it from an agile perspective As this particular project drew to a close and it became more and more apparent that it was not going to succeed the failure of it became depressing. Emotions were expressed by various people on the team that we not encouraging and enforced the failure. Embracing the failure and looking at it for what it is instead of taking it as the end of the world can change how you grow from the experience. Acknowledging that it failed and then focussing on learning from why and how to avoid the failure in the future can change how you feel emotionally about the team, the project and the organization.

    Read the article

  • Summit reflections

    - by Rob Farley
    So far, my three PASS Summit experiences have been notably different to each other. My first, I wasn’t on the board and I gave two regular sessions and a Lightning Talk in which I told jokes. My second, I was a board advisor, and I delivered a precon, a spotlight and a Lightning Talk in which I sang. My third (last week), I was a full board director, and I didn’t present at all. Let’s not talk about next year. I’m not sure there are many options left. This year, I noticed that a lot more people recognised me and said hello. I guess that’s potentially because of the singing last year, but could also be because board elections can bring a fair bit of attention, and because of the effort I’ve put in through things like 24HOP... Yeah, ok. It’d be the singing. My approach was very different though. I was watching things through different eyes. I looked for the things that seemed to be working and the things that didn’t. I had staff there again, and was curious to know how their things were working out. I knew a lot more about what was going on behind the scenes to make various things happen, and although very little about the Summit was actually my responsibility (based on not having that portfolio), my perspective had moved considerably. Before the Summit started, Board Members had been given notebooks – an idea Tom (who heads up PASS’ marketing) had come up with after being inspired by seeing Bill walk around with a notebook. The plan was to take notes about feedback we got from people. It was a good thing, and the notebook forms a nice pair with the SQLBits one I got a couple of years ago when I last spoke there. I think one of the biggest impacts of this was that during the first keynote, Bill told everyone present about the notebooks. This set a tone of “we’re listening”, and a number of people were definitely keen to tell us things that would cause us to pull out our notebooks. PASSTV was a new thing this year. Justin, the host, featured on the couch and talked a lot of people about a lot of things, including me (he talked to me about a lot of things, I don’t think he talked to a lot people about me). Reaching people through online methods is something which interests me a lot – it has huge potential, and I love the idea of being able to broadcast to people who are unable to attend in person. I’m keen to see how this medium can be developed over time. People who know me will know that I’m a keen advocate of certification – I've been SQL certified since version 6.5, and have even been involved in creating exams. However, I don’t believe in studying for exams. I think training is worthwhile for learning new skills, but the goal should be on learning those skills, not on passing an exam. Exams should be for proving that the skills are there, not a goal in themselves. The PASS Summit is an excellent place to take exams though, and with an attitude of professional development throughout the event, why not? So I did. I wasn’t expecting to take one, but I was persuaded and took the MCM Knowledge Exam. I hadn’t even looked at the syllabus, but tried it anyway. I was very tired, and even fell asleep at one point during it. I’ll find out my result at some point in the future – the Prometric site just says “Tested” at the moment. As I said, it wasn’t something I was expecting to do, but it was good to have something unexpected during the week. Of course it was good to catch up with old friends and make new ones. I feel like every time I’m in the US I see things develop a bit more, with more and more people knowing who I am, who my staff are, and recognising the LobsterPot brand. I missed being a presenter, but I definitely enjoyed seeing many friends on the list of presenters. I won’t try to list them, because there are so many these days that people might feel sad if I don’t mention them. For those that I managed to see, I was pleased to see that the majority of them have lifted their presentation skills since I last saw them, and I happily told them as much. One person who I will mention was Paul White, who travelled from New Zealand to his first PASS Summit. He gave two sessions (a regular session and a half-day), packed large rooms of people, and had everyone buzzing with enthusiasm. I spoke to him after the event, and he told me that his expectations were blown away. Paul isn’t normally a fan of crowds, and the thought of 4000 people would have been scary. But he told me he had no idea that people would welcome him so well, be so friendly and so down to earth. He’s seen the significance of the SQL Server community, and says he’ll be back. It’ll be good to see him there. Will you be there too?

    Read the article

< Previous Page | 59 60 61 62 63 64 65 66 67 68 69 70  | Next Page >