Search Results

Search found 15423 results on 617 pages for 'uses clause'.

Page 529/617 | < Previous Page | 525 526 527 528 529 530 531 532 533 534 535 536  | Next Page >

  • Oracle Fusion Procurement Designed for User Productivity

    - by Applications User Experience
    Sean Rice, Manager, Applications User Experience Oracle Fusion Procurement Design Goals In Oracle Fusion Procurement, we set out to create a streamlined user experience based on the way users do their jobs. Oracle has spent hundreds of hours with customers to get to the heart of what users need to do their jobs. By designing a procurement application around user needs, Oracle has crafted a user experience that puts the tools that people need at their fingertips. In Oracle Fusion Procurement, the user experience is designed to provide the user with information that will drive navigation rather than requiring the user to find information. One of our design goals for Oracle Fusion Procurement was to reduce the number of screens and clicks that a user must go through to complete frequently performed tasks. The requisition process in Oracle Fusion Procurement (Figure 1) illustrates how we have streamlined workflows. Oracle Fusion Self-Service Procurement brings together billing metrics, descriptions of the order, justification for the order, a breakdown of the components of the order, and the amount—all in one place. Previous generations of procurement software required the user to navigate to several different pages to gather all of this information. With Oracle Fusion, everything is presented on one page. The result is that users can complete their tasks in less time. The focus is on completing the work, not finding the work. Figure 1. Creating a requisition in Oracle Fusion Self-Service Procurement is a consumer-like shopping experience. Will Oracle Fusion Procurement Increase Productivity? To answer this question, Oracle sought to model how two experts working head to head—one in an existing enterprise application and another in Oracle Fusion Procurement—would perform the same task. We compared Oracle Fusion designs to corresponding existing applications using the keystroke-level modeling (KLM) method. This method is based on years of research at universities such as Carnegie Mellon and research labs like Xerox Palo Alto Research Center. The KLM method breaks tasks into a sequence of operations and uses standardized models to evaluate all of the physical and cognitive actions that a person must take to complete a task: what a user would have to click, how long each click would take (not only the physical action of the click or typing of a letter, but also how long someone would have to think about the page when taking the action), and user interface changes that result from the click. By applying standard time estimates for all of the operators in the task, an estimate of the overall task time is calculated. Task times from the model enable researchers to predict end-user productivity. For the study, we focused on modeling procurement business process task flows that were considered business or mission critical: high-frequency tasks and high-value tasks. The designs evaluated encompassed tasks that are currently performed by employees, professional buyers, suppliers, and sourcing professionals in advanced procurement applications. For each of these flows, we created detailed task scenarios that provided the context for each task, conducted task walk-throughs in both the Oracle Fusion design and the existing application, analyzed and documented the steps and actions required to complete each task, and applied standard time estimates to the operators in each task to estimate overall task completion times. The Results The KLM method predicted that the Oracle Fusion Procurement designs would result in productivity gains in each task, ranging from 13 percent to 38 percent, with an overall productivity gain of 22.5 percent. These performance gains can be attributed to a reduction in the number of clicks and screens needed to complete the tasks. For example, creating a requisition in Oracle Fusion Procurement takes a user through only two screens, while ordering the same item in a previous version requires six screens to complete the task. Modeling user productivity has resulted not only in advances in Oracle Fusion applications, but also in advances in other areas. We leveraged lessons learned from the KLM studies to establish products like Oracle E-Business Suite (EBS). New user experience features in EBS 12.1.3, such as navigational improvements to the main menu, a Google-type search using auto-suggest, embedded analytics, and an in-context list of values tool help to reduce clicks and improve efficiency. For more information about KLM, refer to the Measuring User Productivity blog.

    Read the article

  • I spy a Live Framework portal

    - by jamiet
    Those that have followed my blogs for a while may know that I have a slightly banal interest in Windows Live and, more specifically, the Live Services developer platform'; if that doesn’t sound interesting to you then stop reading now. My interest mainly stems from the Live Mesh technology that was announced a couple of years ago and the data synchronisation platform API that underpins it; that platform is called the Live Framework or LiveFX for short. At the Professional Developer’s Conference (PDC) 2008 Microsoft made LiveFX available to the public as a Tech Preview and I spent some time learning to use it and also built a few test apps on it too. In August 2009 an announcement came that that tech preview was getting shut down: "At the Professional Developer Conference 2008, we gave the developer community access to the technical preview of the Live Framework. The Live Framework is core to our vision of providing you with a consistent programming interface. Now we are working to integrate existing services, controls and the Live Framework into the next release of Windows Live. Your feedback continues to help us build the best possible offerings for Windows Live users, for you and for your customers. " Since then news on LiveFX has disappeared save for a throwaway session at PDC09 and I was hoping that news was going to appear at this week’s MIX conference but nothing was forthcoming. Instead though today I stumbled upon an unannounced portal for future LiveFX applications on Microsoft’s Azure portal at http://live.azure.com. Check it out: I consider this to be very good news. This Azure portal was built after the LiveFX tech preview was decommissioned so seeing Live Services existing so prominently alongside Microsoft’s other cloud efforts like Windows Azure and SQL Azure vindicates my early investment in the platform and gives me hope that we’re going to see something get released very very soon. I believe that the potential uses for this platform are extremely compelling and I’m looking forward to trying some out in the near future. I am also expecting LiveFX to have a heavy dependency on the OData protocol that I talked about yesterday in my post OData.org updated - gives clues about future sql azure enhancements so you can tell where my interest in that stems from. In case you’re wondering the projects that you see listed above (Basic List Sample, JT-proj etc…) are projects that I built on the old Tech Preview platform so clearly that stuff has not gone for good which is also good news; not just because it means I’ll have access to the code I wrote before but I also assume it means that LiveFX won’t have changed much since its tech preview incarnation. I know there are other LiveFX buffs out there and hopefully this news reaches some of them. If you are one of them the please put a comment below and let me know your thoughts! @Jamiet Share this post: email it! | bookmark it! | digg it! | reddit! | kick it! | live it!

    Read the article

  • Opening the Internet Settings Dialog and using Windows Default Network Settings via Code

    - by Rick Strahl
    Ran into a question from a client the other day that asked how to deal with Internet Connection settings for running  HTTP requests. In this case this is an old FoxPro app and it's using WinInet to handle the actual HTTP connection. Another client asked a similar question about using the IE Web Browser control and configuring connection properties. Regardless of platform or tools used to do HTTP connections, you can probably configure custom connection and proxy settings in your application to configure http connection settings manually. However, this is a repetitive process for each application requires you to track system information in your application which is undesirable. Often it's much easier to rely on the system wide proxy settings that Windows provides via the Internet Settings dialog. The dialog is a Control Panel applet (inetcpl.cpl) and is the same dialog that you see when you pop up Internet Explorer's Options dialog: This dialog controls the Windows connection properties that determine how the Windows HTTP stack connects to the Internet and how Proxy's are used if configured. Depending on how the HTTP client is configured - it can typically inherit and use these global settings. Loading the Settings Dialog Programmatically The settings dialog is a Control Panel applet with the name of: inetcpl.cpl and you can use any Shell execution mechanism (Run dialog, ShellExecute API, Process.Start() in .NET etc.) to invoke the dialog. Changes made there are immediately reflected in any applications that use the default connection settings. In .NET you can simply do this to bring up the Internet Settings dialog with the Connection tab enabled: Process.Start("inetcpl.cpl",",4"); In FoxPro you can simply use the RUN command to execute inetcpl.cpl: lcCmd = "inetcpl.cpl ,4" RUN &lcCmd Using the Default Connection/Proxy Settings When using WinInet you specify the Http connect type in the call to InternetOpen() like this (FoxPro code here): hInetConnection=; InternetOpen(THIS.cUserAgent,0,; THIS.chttpproxyname,THIS.chttpproxybypass,0) The second parameter of 0 specifies that the default system proxy settings should be used and it uses the settings from the Internet Settings Connections tab. Other connection options for HTTP connections include 1 - direct (no proxies and ignore system settings), 3 - explicit Proxy specification. In most situations a connection mode setting of 0 should work. In .NET HTTP connections by default are direct connections and so you need to explicitly specify a default proxy or proxy configuration to use. The easiest way to do this is on the application level in the config file: <configuration> <system.net> <defaultProxy> <proxy bypassonlocal="False" autoDetect="True" usesystemdefault="True" /> </defaultProxy> </system.net> </configuration> You can do the same sort of thing in code specifying the proxy explicitly and using System.Net.WebProxy.GetDefaultProxy(). So when making HTTP calls to Web Services or using the HttpWebRequest class you can set the proxy with: StoreService.Proxy = WebProxy.GetDefaultProxy(); All of this is pretty easy to deal with and in my opinion is a way better choice to managing connection settings than having to track this stuff in your own application. Plus if you use default settings, most of the time it's highly likely that the connection settings are already properly configured making further configuration rare.© Rick Strahl, West Wind Technologies, 2005-2011Posted in Windows  HTTP  .NET  FoxPro   Tweet (function() { var po = document.createElement('script'); po.type = 'text/javascript'; po.async = true; po.src = 'https://apis.google.com/js/plusone.js'; var s = document.getElementsByTagName('script')[0]; s.parentNode.insertBefore(po, s); })();

    Read the article

  • ASP.NET MVC: Simple view to display contents of DataTable

    - by DigiMortal
    In one of my current projects I have to show reports based on SQL Server views. My code should be not aware of data it shows. It just asks data from view and displays it user. As WebGrid didn’t seem to work with DataTable (at least with no hocus-pocus) I wrote my own very simple view that shows contents of DataTable. I don’t focus right now on data querying questions as this part of my simple generic reporting stuff is still under construction. If the final result is something good enough to share with wider audience I will blog about it for sure. My view uses DataTable as model. It iterates through columns collection to get column names and then iterates through rows and writes out values of all columns. Nothing special, just simple generic view for DataTable. @model System.Data.DataTable @using System.Data; <h2>Report</h2> <table>     <thead>     <tr>     @foreach (DataColumn col in Model.Columns)         {                  <th>@col.ColumnName</th>     }         </tr>     </thead>             <tbody>     @foreach (DataRow row in Model.Rows)         {                 <tr>         @foreach (DataColumn col in Model.Columns)                 {                          <td>@row[col.ColumnName]</td>         }                 </tr>     }         </tbody> </table> In my controller action I have code like this. GetParams() is simple function that reads parameter values from form. This part of my simple reporting system is still under construction but as you can see it will be easy to use for UI developers. public ActionResult TasksByProjectReport() {      var data = _reportService.GetReportData("MEMOS",GetParams());      return View(data); } Before seeing next silver bullet in this example please calm down. It is just plain and simple stuff for simple needs. If you need advanced and powerful reporting system then better use existing components by some vendor.

    Read the article

  • Calling XAI Inbound Services from Oracle BI Publisher

    - by ACShorten
    Note: This technique requires Oracle BI Publisher 1.1.3.4.1 which supports Service Complex Types. Web Services require credentials for authentication. Note: The deafults for the product installation are used in this article. If your site uses alternative values then substitute those alternatives where applicable. Note: Examples shown in this article are examples for illustrative purposes only. When building a report in Oracle BI Publisher it may be necessary to call an XAI Inbound Service to get information via the object rather than directly calling the database tables for various reasons: The CLOB fields used in the Object are accessible for a report. Note: CLOB fields cannot be used as criteria in the current release. Objects can take advantage of algorithms to format or calculate additional data that is not stored in the database directly. For example, Information format strings can automatically generated by the object which gives consistent information between a report and the online screens. To use this facility the following process must be performed: Ensure that the product group, cisusers by default, is enabled for the SPLServiceBean in the console. This allows BI Publisher access to call Web Services directly. To ensure this follow the instructions below: Logon to the Oracle WebLogic server console using an appropriate administrator account. By default the user system or weblogic is provided for this purpose. Navigate to the Security Realms section and select your configured realm. This is set to myrealm by default. In the Roles and Policies section, expand the SPLService section of the Deployments option to reveal the SPLServiceBean roles. If there is no role associated with the SPLServiceBean, create a new EJB role and specify the cisusers role, by default. For example:   Add a Role Condition to the role just created, with a Predicate List of Group and specify cisusers as the Group Argument Name. For example: Save all your changes. The XAI Inbound Services to be used by BI Publisher must be defined prior to using the interface. Refer to the XAI Best Practices (Doc Id: 942074.1) from My Oracle Support or via the online help for more information about this process. Inside BI Publisher create your report, according to the BI Publisher documentation. When specifying the dataset, under the Data Model Report option, specify the following to use an XAI Inbound Service as a data source: Parameter Comment Type Web Service Complex Type true Username Any valid user name within the product. This user MUST have security access to the objects referenced in the XAI Inbound Service Password Authentication password for Username Timeout Timeout, in seconds, set for the Web Service call. For example 60 seconds. WSDL URL Use the WSDL URL on the XAI Inbound Service definition as your WSDL URL. It will be in the following format by default:http://<host>:<port>/<server>/XAIApp/xaiserver/<service>?WSDLwhere: <host> - Host Name of Web Application Server <port> - Port allocated to Web Application Server for product access <server> - Server context for server <service> - XAI Inbound Service Name Note: For customers using secure transmission should substitute https instead of http and use the HTTPS port allocated to the product at installation time. Web Service Select the name of the service that shows in the drop-down menu. If no service name shows up, it means that Publisher could not establish a connection with the server or WSDL name provided in the above URL in order to get the service name. See BI Publisher server log for more information. Method Select the name of the Method that shows in the drop-down menu. A method name should show in the Method drop-down menu once the Web Service name is selected. For example: Additionally, filters can be used from the Web Service that can be generated, required or optional, from the WSDL in the Parameter List. For example:

    Read the article

  • Converting an Oracle VM VirtualBox VM into an Oracle VM Server image

    - by wim.coekaerts
    As we are working on tighter seemless moving of VM's between the 2 products, here are a few simple steps to convert an existing Oracle VM VirtualBox image over. Steps involved to make it easy/straightforward : (1) When creating a VM in Virtualbox, using Oracle Linux as an example, make sure that /etc/fstab only uses labels. Do not use hardcoded device names. instead of an entry /dev/sda1 /u01 ext3 defaults 1 1 use LABEL=foo /u01 ext3 defaults 1 1 for more info on labels : man e2label or use a logical volume /dev/VolGroup00/LVfoo /u01 ext3 defaults 1 1 Doing so will make it easier to have an OS boot up on a different hypervisor with potentially different device names. For instance, the VirtualBox VM might expose a scsi driver while in Oracle VM Server you might end up with an ide disk, this then changes /dev/sda to /dev/hda. (2) If you have a VM created that you want to convert, then shut down the VM in VirtualBox and convert the image files : go the the directory that contains your HardDisk image files (.VirtualBox/HardDisks/* as an example) for each of the virtual disks run the following command : VBoxManage clonehd virtualdiskfilename.vdi system.img --format raw where virtualdiskfilename.vdi is the original VBox VM file (this can also be a vmdk file) and system.img is the name of the virtualdisk for Oracle VM. this can be any filename as well, I typically use system.img to specify the boot disk (as is common for Oracle VM template creation) (3) create a vm.cfg To run a VM converted from VirtualBox, you have to create a vm.cfg for Oracle VM server that creates an HVM guest. The easiest is to use a simple hvm vm.cfg and change it for your vm. I have an example here : acpi = 1 apic = 1 builder = 'hvm' device_model = '/usr/lib/xen/bin/qemu-dm' disk = ['file:system.img,hda,w', 'file:oracle.img,hdb,w',',hdc:cdrom,r',] kernel = '/usr/lib/xen/boot/hvmloader' memory = '1024' name = 'vmname' on_crash = 'restart' on_reboot = 'restart' pae = 1 serial = 'pty' timer_mode = '0' usbdevice = 'tablet' vcpus = 1 vif = ['bridge=xenbr0,type=ioemu'] vif_other_config = [] vnc = 1 vncconsole = 1 vnclisten = '0.0.0.0' vncpasswd = '' vncunused = 1 If you take the above vm.cfg, all you need to do - modify disk = (add your virtual disks in there) - modify memory = (amount of memory your VM needs) - modify name = (enter a name for your VM here) - modify vif = (might want to replace bridge=xenbr0 to the bridge you want to use) if you want more than 1 vcpu or other changes of course you have to make those as well. (4) copy this set of files onto your Oracle VM server or onto a webserver in a subdirectory and import the template through Oracle VM Manager. You can also just start the vm using xm create vm.cfg if you like. And that's it. As I said, we are working on automation around all this but it is relatively trivial to convert VM's over as long as you take the basic issues into account. Primarily the set up of the filesystems and the use of labels in /etc/fstab. There are other potential things to look at, such as network config. If you want to make that part clean then prior to shutting down the VM change /etc/modprobe.conf and/or add the mac address of the VM into the vm.cfg in the vifs line. The good thing, at least with Linux, is that even tho the virtual hardware changes, Linux will deal with it just fine (e1000 vs 8139 realtek, ide vs scsi etc). hope this helps.

    Read the article

  • SQL SERVER – Sends backups to a Network Folder, FTP Server, Dropbox, Google Drive or Amazon S3

    - by pinaldave
    Let me tell you about one of the most useful SQL tools that every DBA should use – it is SQLBackupAndFTP. I have been using this tool since 2009 – and it is the first program I install on a SQL server. Download a free version, 1 minute configuration and your daily backups are safe in the cloud. In summary, SQLBackupAndFTP Creates SQL Server database and file backups on schedule Compresses and encrypts the backups Sends backups to a network folder, FTP Server, Dropbox, Google Drive or Amazon S3 Sends email notifications of job’s success or failure SQLBackupAndFTP comes in Free and Paid versions (starting from $29) – see version comparison. Free version is fully functional for unlimited ad hoc backups or for scheduled backups of up to two databases – it will be sufficient for many small customers. What has impressed me from the beginning – is that I understood how it works and was able to configure the job from a single form (see Image 1 – Main form above) Connect to you SQL server and select databases to be backed up Click “Add backup destination” to configure where backups should go to (network, FTP Server, Dropbox, Google Drive or Amazon S3) Enter your email to receive email confirmations Set the time to start daily full backups (or go to Settings if you need Differential or  Transaction Log backups on a flexible schedule) Press “Run Now” button to test You can get to this form if you click “Settings” buttons in the “Schedule section”. Select what types of backups and how often you want to run them and you will see the scheduled backups in the “Estimated backup plan” list A detailed tutorial is available on the developer’s website. Along with SQLBackupAndFTP setup gives you the option to install “One-Click SQL Restore” (you can install it stand-alone too) – a basic tool for restoring just Full backups. However basic, you can drag-and-drop on it the zip file created by SQLBackupAndFTP, it unzips the BAK file if necessary, connects to the SQL server on the start, selects the right database, it is smart enough to restart the server to drop open connections if necessary – very handy for developers who need to restore databases often. You may ask why is this tool is better than maintenance tasks available in SQL Server? While maintenance tasks are easy to set up, SQLBackupAndFTP is still way easier and integrates solution for compression, encryption, FTP, cloud storage and email which make it superior to maintenance tasks in every aspect. On a flip side SQLBackupAndFTP is not the fanciest tool to manage backups or check their health. It only works reliably on local SQL Server instances. In other words it has to be installed on the SQL server itself. For remote servers it uses scripting which is less reliable. This limitations is actually inherent in SQL server itself as BACKUP DATABASE command  creates backup not on the client, but on the server itself. This tool is compatible with almost all the known SQL Server versions. It works with SQL Server 2008 (all versions) and many of the previous versions. It is especially useful for SQL Server Express 2005 and SQL Server Express 2008, as they lack built in tools for backup. I strongly recommend this tool to all the DBAs. They must absolutely try it as it is free and does exactly what it promises. You can download your free copy of the tool from here. Please share your experience about using this tool. I am eager to receive your feedback regarding this article. Reference: Pinal Dave (http://blog.SQLAuthority.com)   Filed under: PostADay, SQL, SQL Authority, SQL Backup and Restore, SQL Query, SQL Server, SQL Tips and Tricks, SQL Utility, SQLServer, T SQL, Technology

    Read the article

  • Plymouth and GRUB do not show at all

    - by WarriorIng64
    I am using Ubuntu 11.04 64-bit as my only OS on my desktop computer, which used to only run Ubuntu 10.04 LTS until I had the time to upgrade it with a fresh install. It uses integrated NVIDIA graphics (listed as a GeForce 6150SE nForce 430 by the NVIDIA X Server Settings utility) with the current proprietary driver as provided by the Additional Drivers utility, and has a VGA connection to a 1680x1050 Acer monitor. I used to get the (ugly-looking version of) Plymouth graphical boot screen while under 10.04. It didn't look that great, but I was fine with it. Now, it doesn't show on 11.04 at all during boot (I just get an error message in a moving gray box from the monitor saying "Input Not Supported"), and only rarely it will show on shutdown, all garbled up. I could not get GRUB to show during boot while holding down Shift, either (same error message), but pressing Enter while it should be up starts the system normally. A picture of the error message I was getting: Once fully booted, the system still shows the login screen and desktop just fine. Any information on how to troubleshoot this would be appreciated. If there's any hardware-specific stuff I forgot to include here, let me know the relevant commands to run in a comment below. Things that I've tried: Running plymouth in a framebuffer: no effect Booting with nomodeset as my grub boot: option no effect Booting with nomodeset and plymouth in a framebuffer: no effect other than Plymouth showing during shutdown only Following the Softpedia instructions for fixing Plymouth's resolution: Problem mostly solved, except logo does not show in Plymouth during boot, and both grub and Plymouth are slightly off-center #4 above, but with nomodeset removed as a grub boot option: same effect as #4 #5 above, but with vt.handoff=7 added as a grub boot option: same effect as #4 I have added the current contents of /etc/default/grub as requested in the comments: # If you change this file, run 'update-grub' afterwards to update # /boot/grub/grub.cfg. # For full documentation of the options in this file, see: # info -f grub -n 'Simple configuration' GRUB_DEFAULT=0 GRUB_HIDDEN_TIMEOUT=0 GRUB_HIDDEN_TIMEOUT_QUIET=true GRUB_TIMEOUT=10 GRUB_DISTRIBUTOR=`lsb_release -i -s 2> /dev/null || echo Debian` GRUB_CMDLINE_LINUX_DEFAULT="quiet splash video=uvesafb:mode_option=1280x1024-24,mtrr=3,scroll=ywrap" GRUB_CMDLINE_LINUX="" # Uncomment to enable BadRAM filtering, modify to suit your needs # This works with Linux (no patch required) and with any kernel that obtains # the memory map information from GRUB (GNU Mach, kernel of FreeBSD ...) #GRUB_BADRAM="0x01234567,0xfefefefe,0x89abcdef,0xefefefef" # Uncomment to disable graphical terminal (grub-pc only) #GRUB_TERMINAL=console # The resolution used on graphical terminal # note that you can use only modes which your graphic card supports via VBE # you can see them in real GRUB with the command `vbeinfo' GRUB_GFXMODE=1280x1024 # Uncomment if you don't want GRUB to pass "root=UUID=xxx" parameter to Linux #GRUB_DISABLE_LINUX_UUID=true # Uncomment to disable generation of recovery mode menu entries #GRUB_DISABLE_RECOVERY="true" # Uncomment to get a beep at grub start #GRUB_INIT_TUNE="480 440 1" CURRENT STATUS: I forgot to uncomment one line as per "things that I've tried" #4, so I took care of that. I can now see GRUB during startup when I hold Shift and a normal-looking Plymouth during shutdown...but Plymouth during boot is now just a solid purple screen. In each case, it's displayed a little off-center to the left, with a thin black bar running down the right side of the monitor. The error pictured above no longer shows. I'd say this problem is about 2/3 solved now. UPDATE: After Natty started freezing up on me, I decided to dual-boot with Oneiric, which unfortunately shows the same problems. Rather than trying all these workarounds though, I decided to do what I should have done from the start and file a pair of bug reports. LAST UPDATE: Bug 850908 has been confirmed as a legitimate nouveaufb bug. I have overwritten my 11.04 partition with 12.04 LTS, and I can confirm at this time that the issue is present there, as well. I will now flag this question to be closed, yet I hope it was helpful for anyone who experienced similar issues; if you are still having the same problem as me, please go there and mark yourself as affected. Thanks!

    Read the article

  • The Hunger Games for Aspiring IT Professionals

    - by Dain C. Hansen
    Normal 0 false false false EN-US X-NONE X-NONE MicrosoftInternetExplorer4 /* 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:10.0pt; font-family:"Calibri","sans-serif"; mso-bidi-font-family:"Times New Roman";} It seems that no one can escape the buzz around Hunger Games. And who could? Stephen King said it best in his review when he referred to the Collins’ novel as “a violent, jarring speed-rap of a novel that generates nearly constant suspense and may also generate a fair amount of controversy”. So what’s the tie in for IT? Let’s leave the dystopia of District 12 and come back to today’s reality. This is the world of radical IT paradigm shifts that haven’t been seen since Java was introduced in 1995. Everything you learned in school is probably outdated as of Friday. And everything you learned on Friday will probably change when you get to work on Monday. Nevertheless, we’re eager, we’re aspiring, we’re hungry to learn. While the challenges upon us may not rival the venomous bees (or ‘tracker jackers’) seen in this blockbuster, there are certainly obstacles to be found. In preparation, I leave you two pieces of advice - aside from avoiding werewolves… Learn the Cloud If you had asked me what to learn in 1995, I would have said, “Go learn Java”. But now my advice is “Go learn Java and then learn Cloud”. Cloud computing and Java go hand in hand. This is especially true for Oracle’s own Public Cloud which uses Java (via WebLogic 12c) as well as Oracle Database at its core foundation. Understanding the connotations of elasticity, scale, virtualization, and multi-tenancy, (to name just a few) requires a strong foundation in computer science and especially Java to get it right. Without Java, the Cloud is nothing more than a brittle application meagerly deployed on the internet. Get Social and Actively Participate And at all levels. Socializing your ideas internally is dreadfully important. And this means socializing and communicating your good ideas to lines of business, to architects, business analysts, developers, DBAs and Operations. But don’t forget to go external. Stay current by being on the lookout for blogs, tweets, webcasts, papers, podcasts and videos for your technology area. Be not just a subscriber but a participant in these channels as well. Attend industry and vendor sponsored events to learn from the experts – and seek out opportunities to stay connected with those that are smarter than you. You’ll gain more understanding if you participate actively. At the same time you’ll make friends (and allies) and you’ll be glad you did. Tell help you get social and actively participate [while learning the Cloud] here are a couple of pointers for you: See our website on Cloud and Fusion Middleware Subscribe to our regular Fusion Middleware Newsletter Follow us on Twitter and Facebook Find us at one of our key events Meanwhile, happy IT hunger games!

    Read the article

  • Review: ComponentOne Studio for Entity Framework

    - by Tim Murphy
    While I have always been a fan of libraries that improve coding efficiency and reduce code redundancy I have mostly been using ones that were in the public domain.  As part of the Geeks With Blogs Influencers program a got my hands on ComponentOne’s Studio for Entity Framework.  Below are my thought after working with the product for several weeks. My coding preference has always been maintainable code that is reusable across an enterprises protfolio.  Because of this my focus in reviewing this product is less on the RAD components and more on its benefits for layered applications using code first Entity Framework. Before we get into the pros and cons here is a summary of the main feature listed for SEF. Unified Data Context Virtual Data Access More Powerful Data Binding Pros The first thing that I found to my liking is the C1DataSource. It basically manages a cache for your Entity Model context.  Under RAD conditions this is setup automatically when you drop the object on a your design surface.  If you are like me and want to abstract you data management into a library it takes a little more work, but it is still acceptable and gains the same benefits. The second feature that I found beneficial is the definition of views with improved sorting and filtering.  Again the ease of use of these features is greater on the RAD side but no capabilities are missing when manipulating object in code. Linq has become my friend over the last couple of years and it was great to see that ComponentOne had ensured that it remained a first class citizen in their design.  When you look into this product yourself I would suggest taking a dive into LiveLinq which allow the joining of different data source types. As I went through discovering the features of this framework I appreciated the number of examples that they supplied for different uses.  Besides showing how to use SEF with WinForms, WPF and Silverlight they also showed how to accomplish tasks both RAD, code only and MVVM approaches. Cons The only area that I would really like to see improvement is in there level of detail in their documentation.  Specifically I would like to have seen some of the supporting code explained, such as what some supporting object did, in the examples instead of having to go to the programmer’s reference. I did find some times where currently existing projects had some trouble determining scope that the RAD controls were allowed, but I expect this is something that is in part end user related. Summary Overall I found the Studio for Entity Framework capable and well thought out.  If you are already using the Entity Framework this product will fit into your environment with little effort in return for greater flexibility and greater robustness in your solutions. Whether the $895 list price for a standard version works for you will depend on your return on investment. Smaller companies with only a small number of projects may not be able to stomach it, you get a full featured product that is supported by a well established company.  The more projects and the more code you have the greater your return on investment will be. Personally I intend to apply this product to some production systems and will probably have some tips and tricks in the future. del.icio.us Tags: ComponentOne,Studio for Entity Framework,Geeks With Blogs,Influencers,Product Reviews

    Read the article

  • Building vs. Buying a Master Data Management Solution

    - by david.butler(at)oracle.com
    Many organizations prefer to build their own MDM solutions. The argument is that they know their data quality issues and their data better than anyone. Plus a focused solution will cost less in the long run then a vendor supplied general purpose product. This is not unreasonable if you think of MDM as a point solution for a particular data quality problem. But this approach carries significant risk. We now know that organizations achieve significant competitive advantages when they deploy MDM as a strategic enterprise wide solution: with the most common best practice being to deploy a tactical MDM solution and grow it into a full information architecture. A build your own approach most certainly will not scale to a larger architecture unless it is done correctly with the larger solution in mind. It is possible to build a home grown point MDM solution in such a way that it will dovetail into broader MDM architectures. A very good place to start is to use the same basic technologies that Oracle uses to build its own MDM solutions. Start with the Oracle 11g database to create a flexible, extensible and open data model to hold the master data and all needed attributes. The Oracle database is the most flexible, highly available and scalable database system on the market. With its Real Application Clusters (RAC) it can even support the mixed OLTP and BI workloads that represent typical MDM data access profiles. Use Oracle Data Integration (ODI) for batch data movement between applications, MDM data stores, and the BI layer. Use Oracle Golden Gate for more real-time data movement. Use Oracle's SOA Suite for application integration with its: BPEL Process Manager to orchestrate MDM connections to business processes; Identity Management for managing users; WS Manager for managing web services; Business Intelligence Enterprise Edition for analytics; and JDeveloper for creating or extending the MDM management application. Oracle utilizes these technologies to build its MDM Hubs.  Customers who build their own MDM solution using these components will easily migrate to Oracle provided MDM solutions when the home grown solution runs out of gas. But, even with a full stack of open flexible MDM technologies, creating a robust MDM application can be a daunting task. For example, a basic MDM solution will need: a set of data access methods that support master data as a service as well as direct real time access as well as batch loads and extracts; a data migration service for initial loads and periodic updates; a metadata management capability for items such as business entity matrixed relationships and hierarchies; a source system management capability to fully cross-reference business objects and to satisfy seemingly conflicting data ownership requirements; a data quality function that can find and eliminate duplicate data while insuring correct data attribute survivorship; a set of data quality functions that can manage structured and unstructured data; a data quality interface to assist with preventing new errors from entering the system even when data entry is outside the MDM application itself; a continuing data cleansing function to keep the data up to date; an internal triggering mechanism to create and deploy change information to all connected systems; a comprehensive role based data security system to control and monitor data access, update rights, and maintain change history; a flexible business rules engine for managing master data processes such as privacy and data movement; a user interface to support casual users and data stewards; a business intelligence structure to support profiling, compliance, and business performance indicators; and an analytical foundation for directly analyzing master data. Oracle's pre-built MDM Hub solutions are full-featured 3-tier Internet applications designed to participate in the full Oracle technology stack or to run independently in other open IT SOA environments. Building MDM solutions from scratch can take years. Oracle's pre-built MDM solutions can bring quality data to the enterprise in a matter of months. But if you must build, at lease build with the world's best technology stack in a way that simplifies the eventual upgrade to Oracle MDM and to the full enterprise wide information architecture that it enables.

    Read the article

  • WEB203 &ndash; Jump into Silverlight!&hellip; and Become Effective Immediately with Tim Huckaby, Fou

    - by Robert Burger
    Getting ready for the good stuff. Definitely wish there were more Silverlight and WCF RIA sessions, but this is a start.  Was lucky to get a coveted power-enabled seat.  Luckily, due to my trustily slow Verizon data card, I can get these notes out amidst a total Internet outage here.  This is the second breakout session of the day, and is by far standing-room only.  I stepped out before the session started to get a cool Diet COKE and wouldn’t have gotten back in if I didn’t already have a seat. Tim says this is an intro session and that he’s been begging for intro sessions at TechEd for years and that by looking at this audience, he thinks the demand is there.  Admittedly, I didn’t know this was an intro session, or I might have gone elsewhere.  But, it was the very first Silverlight session, so I had to be here. Tim says he will be providing a very good comprehensive reference application at the end of the presentation.  He has just demoed it, and it is a full CRUD-based Sales Manager application based on…  AdventureWorks! Session Agenda What it is / How to get started Declarative Programming Layout and Controls, Events and Commands Working with Data Adding Style to Your Application   Silverlight…  “WPF Light” Why is the download 4.2MB?  Because the direct competitor is a 4.2MB download.  There is no technical reason it is not the entire framework.  It is purely to “be competitive”.   Getting Started Get all of the following downloads from www.silverlight.net/getstarted Install VS2010 or Visual Web Developer Express 2010 Install Silverlight 4 Tools for VS2010 Install Expression Blend 4 Install the Silverlight 4 Toolkit   Reference Application Features Uses MVVM pattern – a way to move data access code that would normally be inline within the UI and placing it in nice data access libraries Images loaded dynamically from the database, converting GIF to PNG because Silverlight does not support GIF. LINQ to SQL is the data access model WCF is the data provider and is using binary message encoding   Declarative Programming XAML replaces code for UI representation Attributes control Layout and Style Event handlers wired-up in XAML Declarative Data Binding   Layout Overview Content rendering flows inside of parent Fixed positioning (Canvas) is seldom used Panels are used to house content Margins and Padding over fixed size   Panels StackPanel – Arranges child elements into a single line oriented horizontally or vertically Grid – A flexible grid are that consists of rows and columns Canvas – An are where positions are specifically fixed WrapPanel (in Toolkit) – Positions child elements in sequential position left to right and top to bottom. DockPanel (in Toolkit) – Positions child controls within a dockable area   Positioning Horizontal and Vertical Alignment Margin – Separates an element from neighboring elements Padding – Enlarges the effective size of an element by a thickness   Controls Overview Not all controls created equal Silverlight, as a subset of WPF, so many WPF controls do not exist in the core Siverlight release Silverlight Toolkit continues to add controls, but are released in different quality bands Plenty of good 3rd party controls to fill the gaps Windows Phone 7 is to have 95% of controls available in Silverlight Core and Toolkit.   Events and Commands Standard .NET Events Routed Events Commands – based on the ICommand interface – logical action that can be invoked in several ways   Adding Style to Your Application Resource Dictionaries – Contains a hash table of key/value pairs.  Silverlight can only use Static Resources whereas WPF can also use Dynamic Resources Visual State Manager Silverlight 4 supports Implicit styles ResourceDictionary.MergedDictionaries combines many different file-based resources   Downloads

    Read the article

  • A debugging experience with "highly compatible" ASP.NET 4.5

    - by Jeff
    I have to admit that I will pretty much upgrade software for no reason other than being on the latest version. I won't do it if it's super expensive (Adobe gets money from me about once every three or four years at best), but particularly with frameworks and stuff generally available as part of my MSDN subscription, I'll be bleeding edge. CoasterBuzz was running on the MVC 4 framework pretty much as soon as they did a "go live" license for it. I didn't really jump in head-first with Windows 8 and Visual Studio 2012, in part because I just wasn't interested in doing the reinstalls for each new version. Turns out there weren't that many revisions anyway. But when the final versions were released a week and a half ago, I jumped in. I saw on one of the Microsoft sites that .Net 4.5 was a "highly compatible in-place update" to the framework. Good enough for me. I was obviously running it by default in Windows 8, and installed it on my production server. I suppose it's "highly compatible," except when it isn't. Three of my sites are running with various flavors of the MVC version of POP Forums. All of them stopped working under ASP.NET 4.5. It was not immediately obvious what the problem might be beyond an exception indicating that there were no repository classes registered with Ninject, which I use for dependency injection in the forums. This was made all the more weird by the fact that it ran fine locally in the dev Web host. My first instinct was to spin up a Windows Server VM on my local box and put the remote debugger on it. (Side note: running multiple VM's on a Retina MacBook Pro with 16 gigs of RAM is pretty much the most awesome thing ever. I can't believe this computer is for real, and not a 50-pound tower under my desk.) What might have been going on in IIS that doesn't happen in Visual Studio? In the debugging process, I realized that I might be looking in the wrong place. POP Forums creates a Ninject container using a method called from a PreApplicationStartMethod attribute, and at that time registers a module (what Ninject uses to map interfaces to implementations) that maps all of the core dependencies. It also creates an instance of an HttpModule that originally hosted the "services" (search indexing, mailer, etc.), but now just records errors. That's all well and good, but the actual repository mapping, where data is actually read or persisted, happens in Application_Start() in global.asax. The idea there is that you can swap out the SqlSingleWebServer repos for something tuned for multiple servers, Oracle or something else. Of course, if I used something like StructureMap, which does convention-based mapping for dependency injection (a class implementing ISettingsRepository called SettingsRepository is automagically mapped), I wouldn't have to worry about it. In any case, the HttpModule, being instantiated before Application_Start() gets to run, would throw because there was no repo mapped where it could get settings from the database. This makes total sense. The fix is sort of a hack, where I don't setup the innards of the HttpModule until a call to its BeginRequest is made. I say it's a hack, because its primary function, logging exceptions, won't work until the app has warmed up. Still, this brings up an interesting question about the race condition, and what changed in 4.5 when it's running in IIS. In ASP.NET 4, it would appear that the code called via the PreApplicationStartMethod was either failing silently, and running again later, or it was getting to that code after Application_Start was called. In any case, weird thing. The real pain point I'm experiencing now is a bug in MVC 4 that is extremely serious because it renders the mobile/alternate view functionality very much broken.

    Read the article

  • Lazy Evaluation &ndash; Why being lazy in F# blows my mind!

    - by MarkPearl
    First of all – shout out to Peter Adams – from the feedback I have gotten from him on the last few posts of F# that I have done – my mind has just been expanded. I did a blog post a few days ago about infinite sequences – I didn’t really understand what was going on with it, and I still don’t really get it – but I am getting closer. In Peter’s last comment he made mention of Lazy Evaluation. I am ashamed to say that up till then I had never heard about lazy evaluation – how can evaluation be lazy? I mean, I know about lazy loading and that makes sense… but surely something is either evaluated or not! Well… a bit of reading today and I have been enlightened to a point – if you do know of any good articles explaining lazy evaluation please send them to me. So what is lazy evaluation and why is it useful? Lazy evaluation is a process whereby the system only computes the values needed and “ignores” the computations not needed. I’m going out on a limb here, but with this explanation in hand, imagine the following C# code… public int CalculatedVal() { int Val1 = 0; int Val2 = 0; for (int Count = 0; Count < 1000000; Count++) { Val1++; } return Val2; }   Normally, even though Val1 is never needed, the system would loop 1000000 times and add 1 to the current value of Val1. Imagine if the system realized this and so just skipped this segment of code and instead did the following…. public int CalculatedVal() { int Val1 = 0; return Val2; }   A massive saving in computation and wasted effort. Now I am pretty sure it isn’t as simple as this but I think this is the basic idea. For a more detailed explanation of lazy evaluation in c#, Pedram Rezei has a wonderful post on lazy evaluation and makes some C# comparisons. I am not going to take any thunder from him by repeating everything he said since I think he did such a good job of explaining it himself. What I am interested in though is how in F# do you tell something to have lazy evalution, and how do you know if something will be eager or lazy by looking at it. I found this post was useful. From reading around F# by default uses eager evaluation unless explicitly told to use lazy evaluation. One exception to this is sequences, which are lazy by default. Now reading about lazy evaluation has helped me understand more about F# coding… From my understanding of F# because of its declarative nature, most of the actual code you are declaring properties and rules – very little code is actually saying do this right now - but when it comes to a “do this” code section, it then evaluates and optimizes code and applies the rules. So props to lazy evaluation and its optimizations…

    Read the article

  • Create Custom Windows Key Keyboard Shortcuts in Windows

    - by Asian Angel
    Nearly everyone uses keyboard shortcuts of some sort on their Windows system but what if you could create new ones for your favorite apps or folders? You might just be amazed at how simple it can be with just a few clicks and no programming using WinKey. WinKey in Action During the installation process you will see this window that gives you a good basic idea of just what can be accomplished with this wonderful little app. As soon as the installation process has finished you will see the “Main App Window”. It provides a simple straightforward listing of all the keyboard shortcuts that it is currently managing. Note: WinKey will automatically add an entry to the “Startup Listing” in your “Start Menu” during installation. To see the regular built-in Windows keyboard shortcuts that it is managing click “Standard Shortcuts” to select it and then click on “Properties”. For those who are curious WinKey does have a “System Tray Icon” that can be disabled if desired. Now onto creating those new keyboard shortcuts… For our example we decided to create a keyboard shortcut for an app rather than a folder. To create a shortcut for an app click on the small “Paper Icon” as shown here. Once you have done that browse to the appropriate folder and select the exe file. The second step will be choosing which keyboard shortcut you would like to associate with that particular app. You can use the drop-down list to choose from a listing of available keyboard combinations. For our example we chose “Windows Key + A”. The final step is choosing the “Run Mode”. There are three options available in the drop-down list…choose the one that best suits your needs. Here is what our example looked like once finished. All that is left to do at this point is click “OK” to finish the process. And just like that your new keyboard shortcut is now listed in the “Main App Window”. Time to try out your new keyboard shortcut! One quick use of our new keyboard shortcut and Iron Browser opened right up. WinKey really does make creating new keyboard shortcuts as simple as possible. Conclusion If you have been wanting to create new keyboard shortcuts for your favorite apps and folders then it really does not get any simpler than with WinKey. This is definitely a recommended app for anyone who loves “get it done” software. Links Download WinKey at Softpedia Similar Articles Productive Geek Tips Show Keyboard Shortcut Access Keys in Windows VistaCreate a Keyboard Shortcut to Access Hidden Desktop Icons and FilesKeyboard Ninja: 21 Keyboard Shortcut ArticlesAnother Desktop Cube for Windows XP/VistaHow-To Geek on Lifehacker: Control Your Computer with Shortcuts & Speed Up Vista Setup TouchFreeze Alternative in AutoHotkey The Icy Undertow Desktop Windows Home Server – Backup to LAN The Clear & Clean Desktop Use This Bookmarklet to Easily Get Albums Use AutoHotkey to Assign a Hotkey to a Specific Window Latest Software Reviews Tinyhacker Random Tips DVDFab 6 Revo Uninstaller Pro Registry Mechanic 9 for Windows PC Tools Internet Security Suite 2010 Recycle ! Find That Elusive Icon with FindIcons Looking for Good Windows Media Player 12 Plug-ins? Find Out the Celebrity You Resemble With FaceDouble Whoa ! Use Printflush to Solve Printing Problems

    Read the article

  • A temporary disagreement

    - by Tony Davis
    Last month, Phil Factor caused a furore amongst some MVPs with an article that attempted to offer simple advice to developers regarding the use of table variables, versus local and global temporary tables, in their code. Phil makes clear that the table variables do come with some fairly major limitations.no distribution statistics, no parallel query plans for queries that modify table variables.but goes on to suggest that for reasonably small-scale strategic uses, and with a bit of due care and testing, table variables are a "good thing". Not everyone shares his opinion; in fact, I imagine he was rather aghast to learn that there were those felt his article was akin to pulling the pin out of a grenade and tossing it into the database; table variables should be avoided in almost all cases, according to their advice, in favour of temp tables. In other words, a fairly major feature of SQL Server should be more-or-less 'off limits' to developers. The problem with temp tables is that, because they are scoped either in the procedure or the connection, it is easy to allow them to hang around for too long, eating up precious memory and bulking up the shared tempdb database. Unless they are explicitly dropped, global temporary tables, and local temporary tables created within a connection rather than within a stored procedure, will persist until the connection is closed or, with connection pooling, until the connection is reused. It's also quite common with ASP.NET applications to have connection leaks, as Bill Vaughn explains in his chapter in the "SQL Server Deep Dives" book, meaning that the web page exits without closing the connection object, maybe due to an error condition. This will then hang around in the heap for what might be hours before picked up by the garbage collector. Table variables are much safer in this regard, since they are batch-scoped and so are cleaned up automatically once the batch is complete, which also means that they are intuitive to use for the developer because they conform to scoping rules that are closer to those in procedural code. On the surface then, an ideal way to deal with issues related to tempdb memory hogging. So why did Phil qualify his recommendation to use Table Variables? This is another of those cases where, like scalar UDFs and table-valued multi-statement UDFs, developers can sometimes get into trouble with a relatively benign-looking feature, due to way it's been implemented in SQL Server. Once again the biggest problem is how they are handled internally, by the SQL Server query optimizer, which can make very poor choices for JOIN orders and so on, in the absence of statistics, especially when joining to tables with highly-skewed data. The resulting execution plans can be horrible, as will be the resulting performance. If the JOIN is to a large table, that will hurt. Ideally, Microsoft would simply fix this issue so that developers can't get burned in this way; they've been around since SQL Server 2000, so Microsoft has had a bit of time to get it right. As I commented in regard to UDFs, when developers discover issues like with such standard features, the database becomes an alien planet to them, where death lurks around each corner, and they continue to avoid these "killer" features years after the problems have been eventually resolved. In the meantime, what is the right approach? Is it to say "hammers can kill, don't ever use hammers", or is it to try to explain, as Phil's article and follow-up blog post have tried to do, what the feature was intended for, why care must be applied in its use, and so enable developers to make properly-informed decisions, without requiring them to delve deep into the inner workings of SQL Server? Cheers, Tony.

    Read the article

  • Adventures in MVVM &ndash; ViewModel Location and Creation

    - by Brian Genisio's House Of Bilz
    More Adventures in MVVM In this post, I am going to explore how I prefer to attach ViewModels to my Views.  I have published the code to my ViewModelSupport project on CodePlex in case you'd like to see how it works along with some examples.  Some History My approach to View-First ViewModel creation has evolved over time.  I have constructed ViewModels in code-behind.  I have instantiated ViewModels in the resources sectoin of the view. I have used Prism to resolve ViewModels via Dependency Injection. I have created attached properties that use Dependency Injection containers underneath.  Of all these approaches, I continue to find issues either in composability, blendability or maintainability.  Laurent Bugnion came up with a pretty good approach in MVVM Light Toolkit with his ViewModelLocator, but as John Papa points out, it has maintenance issues.  John paired up with Glen Block to make the ViewModelLocator more generic by using MEF to compose ViewModels.  It is a great approach, but I don’t like baking in specific resolution technologies into the ViewModelSupport project. I bring these people up, not to name drop, but to give them credit for the place I finally landed in my journey to resolve ViewModels.  I have come up with my own version of the ViewModelLocator that is both generic and container agnostic.  The solution is blendable, configurable and simple to use.  Use any resolution mechanism you want: MEF, Unity, Ninject, Activator.Create, Lookup Tables, new, whatever. How to use the locator 1. Create a class to contain your resolution configuration: public class YourViewModelResolver: IViewModelResolver { private YourFavoriteContainer container = new YourFavoriteContainer(); public YourViewModelResolver() { // Configure your container } public object Resolve(string viewModelName) { return container.Resolve(viewModelName); } } Examples of doing this are on CodePlex for MEF, Unity and Activator.CreateInstance. 2. Create your ViewModelLocator with your custom resolver in App.xaml: <VMS:ViewModelLocator x:Key="ViewModelLocator"> <VMS:ViewModelLocator.Resolver> <local:YourViewModelResolver /> </VMS:ViewModelLocator.Resolver> </VMS:ViewModelLocator> 3. Hook up your data context whenever you want a ViewModel (WPF): <Border DataContext="{Binding YourViewModelName, Source={StaticResource ViewModelLocator}}"> This example uses dynamic properties on the ViewModelLocator and passes the name to your resolver to figure out how to compose it. 4. What about Silverlight? Good question.  You can't bind to dynamic properties in Silverlight 4 (crossing my fingers for Silverlight 5), but you CAN use string indexing: <Border DataContext="{Binding [YourViewModelName], Source={StaticResource ViewModelLocator}}"> But, as John Papa points out in his article, there is a silly bug in Silverlight 4 (as of this writing) that will call into the indexer 6 times when it binds.  While this is little more than a nuisance when getting most properties, it can be much more of an issue when you are resolving ViewModels six times.  If this gets in your way, the solution (as pointed out by John), is to use an IndexConverter (instantiated in App.xaml and also included in the project): <Border DataContext="{Binding Source={StaticResource ViewModelLocator}, Converter={StaticResource IndexConverter}, ConverterParameter=YourViewModelName}"> It is a bit uglier than the WPF version (this method will also work in WPF if you prefer), but it is still not all that bad.  Conclusion This approach works really well (I suppose I am a bit biased).  It allows for composability from any mechanisim you choose.  It is blendable (consider serving up different objects in Design Mode if you wish... or different constructors… whatever makes sense to you).  It works in Cider.  It is configurable.  It is flexible.  It is the best way I have found to manage View-First ViewModel hookups.  Thanks to the guys mentioned in this article for getting me to something I love using.  Enjoy.

    Read the article

  • Oracle’s New Approach to Cloud-based Applications User Experiences

    - by Oracle OpenWorld Blog Team
    By Misha Vaughan It was an exciting Oracle OpenWorld this year for customers and partners, as they got to see what their input into the Oracle user experience research and development process has produced for cloud-delivered applications. The result of all this engagement and listening is a focus on simplicity, mobility, and extensibility. These were the core themes across Oracle OpenWorld sessions, executive roundtables, and analyst briefings given by Jeremy Ashley, Oracle's vice president of user experience. The highlight of every meeting with a customer featured the new simplified UI for Oracle’s cloud applications.    Attendees at some sessions and events also saw a vision of what is coming next in the Oracle user experience, and they gave direct feedback on whether this would help solve their business problems.  What did attendees think of what they saw this year? Rebecca Wettemann of Nucleus Research was part of  an analyst briefing on next-generation user experiences from Oracle. Here’s what she told CRM Buyer in an interview just after the event:  “Many of the improvements are incremental, which is not surprising, as Oracle regularly updates its application,” Rebecca Wettemann, vice president of Nucleus Research, told CRM Buyer. "Still, there are distinct themes to this latest set of changes. One is usability. Oracle Sales Cloud, for example, is designed to have zero training for onboarding sales reps, which it does," she explained. "It is quite impressive, actually—the intuitive nature of the application and the design work they have done with this goal in mind. The software uses as few buttons and fields as possible," she pointed out. "The sales rep doesn't have to ask, 'what is the next step?' because she can see what it is."  What else did we hear? Oracle OpenWorld is a time when we can take a broader pulse of our customers’ and partners’ concerns. This year we heard some common user experience themes on the following: · A desire to continue to simplify widely used self-service tasks · A need to understand how customers or partners could take some of the UX lessons learned on simplicity and mobility into their own custom areas and projects  · The continuing challenge of needing to support bring-your-own-device and corporate-provided mobile devices to end users · A desire to harmonize user experiences across platforms for specific business-use cases  What does this mean for next year? Well, there were a lot of things we could only show to smaller groups of customers in our Oracle OpenWorld usability labs and HQ lab tours, to partners at our Expo, and to analysts under non-disclosure agreements. But we used these events as a way to get some early feedback about where we are focusing for the year ahead. Attendees gave us a positive response: @bkhan Saw some excellent UX innovations at the expo “@usableapps: Great job @mishavaughan and @vinoskey on #oow13 UX partner expo!” @WarnerTim @usableapps @mishavaughan @vinoskey @ultan Thanks for an interesting afternoon definitely liked the UX tool kits for partners. You can expect Oracle to continue pushing themes of simplicity, mobility, and extensibility even more aggressively in the next year.  If you are interested to find out what really goes on in the UX labs, such as what we are doing with smartphones, tablets, heads-up displays, and the AppsLab robots, feel free to reach out to me for more information: Misha Vaughan or on Twitter: @mishavaughan.

    Read the article

  • Getting UPK data into Excel

    - by maria.cozzolino(at)oracle.com
    Did you ever want someone to review your UPK outline outside of the Developer? You can send your outline to an Excel report, which can be distributed through email. Depending on how much additional data you want with your outline, there are two ways you can do this task. Basic data: • You can print a listing of all the items in the outline. • With your outline open, choose File/Print... • Choose the "Save document as" command on the right, and choose Excel (or xlsx). • HINT: If you have not expanded your entire outline, it's faster to use the commands in Developer to expand the entire outline. However, you can expand specific sections by clicking on them in the print preview. • NOTE: If you have the Details view displayed rather than the Player view, you can print all the data that appears in that view. Advanced data: If you desire a more detailed report, you can use the HP Quality Center publishing style, which also creates an Excel file. This style contains a default set of fields for use with Quality Center, but any of the metadata fields can be added to the report, and it can be used for more than just importing into HP Quality Center. To add additional columns to the HP Quality Center publishing style: 1. Make a copy of the publishing style. This process ensures that you have a good copy to revert to if something goes wrong with your customizations, and also allows you to keep your modifications when the software is upgraded. 2. Open the copy of the columnspec.xml file in your favorite XML editor - I use notepad. (This file is located in a language-specific folder in the HP Quality Center publishing style.) 3. Scroll down the columnspec file until you find the column to include. All the metadata fields that can be added to the report are listed in the columnspec file - you just need to tell the system to include the columns. 4. You will see a series of sections like this: 5. Change the value for "col export" to "yes". This will include the column in the Excel file. 6. If desired, change the value for "Play_ModesColHeader" to be whatever name you wish to appear in the Excel column heading. 7. Save the columnspec file. 8. Save the publishing style package. Now, when you publish for HP Quality Center, you will see your newly added columns. You can refer to the section on Customizing HP Quality Center Output in the Content Deployment Guide for additional customization details. Happy customization! I'd be interested in hearing what other uses you have for Excel reporting. Wishing you and yours a happy and healthy New Year! ~~Maria Cozzolino, Manager of Software Requirements and UI

    Read the article

  • concurrency::accelerator_view

    - by Daniel Moth
    Overview We saw previously that accelerator represents a target for our C++ AMP computation or memory allocation and that there is a notion of a default accelerator. We ended that post by introducing how one can obtain accelerator_view objects from an accelerator object through the accelerator class's default_view property and the create_view method. The accelerator_view objects can be thought of as handles to an accelerator. You can also construct an accelerator_view given another accelerator_view (through the copy constructor or the assignment operator overload). Speaking of operator overloading, you can also compare (for equality and inequality) two accelerator_view objects between them to determine if they refer to the same underlying accelerator. We'll see later that when we use concurrency::array objects, the allocation of data takes place on an accelerator at array construction time, so there is a constructor overload that accepts an accelerator_view object. We'll also see later that a new concurrency::parallel_for_each function overload can take an accelerator_view object, so it knows on what target to execute the computation (represented by a lambda that the parallel_for_each also accepts). Beyond normal usage, accelerator_view is a quality of service concept that offers isolation to multiple "consumers" of an accelerator. If in your code you are accessing the accelerator from multiple threads (or, in general, from different parts of your app), then you'll want to create separate accelerator_view objects for each thread. flush, wait, and queuing_mode When you create an accelerator_view via the create_view method of the accelerator, you pass in an option of immediate or deferred, which are the two members of the queuing_mode enum. At any point you can access this value from the queuing_mode property of the accelerator_view. When the queuing_mode value is immediate (which is the default), any commands sent to the device such as kernel invocations and data transfers (e.g. parallel_for_each and copy, as we'll see in future posts), will get submitted as soon as the runtime sees fit (that is the definition of immediate). When the value of queuing_mode is deferred, the commands will be batched up. To send all buffered commands to the device for execution, there is a non-blocking flush method that you can call. If you wish to block until all the commands have been sent, there is a wait method you can call. Deferring is a more advanced scenario aimed at performance gains when you are submitting many device commands and you want to avoid the tiny overhead of flushing/submitting each command separately. Querying information Just like accelerator, accelerator_view exposes the is_debug and version properties. In fact, you can always access the accelerator object from the accelerator property on the accelerator_view class to access the accelerator interface we looked at previously. Interop with D3D (aka DX) In a later post I'll show an example of an app that uses C++ AMP to compute data that is used in pixel shaders. In those scenarios, you can benefit by integrating C++ AMP into your graphics pipeline and one of the building blocks for that is being able to use the same device context from both the compute kernel and the other shaders. You can do that by going from accelerator_view to device context (and vice versa), through part of our interop API in amp.h: *get_device, create_accelerator_view. More on those in a later post. Comments about this post by Daniel Moth welcome at the original blog.

    Read the article

  • How John Got 15x Improvement Without Really Trying

    - by rchrd
    The following article was published on a Sun Microsystems website a number of years ago by John Feo. It is still useful and worth preserving. So I'm republishing it here.  How I Got 15x Improvement Without Really Trying John Feo, Sun Microsystems Taking ten "personal" program codes used in scientific and engineering research, the author was able to get from 2 to 15 times performance improvement easily by applying some simple general optimization techniques. Introduction Scientific research based on computer simulation depends on the simulation for advancement. The research can advance only as fast as the computational codes can execute. The codes' efficiency determines both the rate and quality of results. In the same amount of time, a faster program can generate more results and can carry out a more detailed simulation of physical phenomena than a slower program. Highly optimized programs help science advance quickly and insure that monies supporting scientific research are used as effectively as possible. Scientific computer codes divide into three broad categories: ISV, community, and personal. ISV codes are large, mature production codes developed and sold commercially. The codes improve slowly over time both in methods and capabilities, and they are well tuned for most vendor platforms. Since the codes are mature and complex, there are few opportunities to improve their performance solely through code optimization. Improvements of 10% to 15% are typical. Examples of ISV codes are DYNA3D, Gaussian, and Nastran. Community codes are non-commercial production codes used by a particular research field. Generally, they are developed and distributed by a single academic or research institution with assistance from the community. Most users just run the codes, but some develop new methods and extensions that feed back into the general release. The codes are available on most vendor platforms. Since these codes are younger than ISV codes, there are more opportunities to optimize the source code. Improvements of 50% are not unusual. Examples of community codes are AMBER, CHARM, BLAST, and FASTA. Personal codes are those written by single users or small research groups for their own use. These codes are not distributed, but may be passed from professor-to-student or student-to-student over several years. They form the primordial ocean of applications from which community and ISV codes emerge. Government research grants pay for the development of most personal codes. This paper reports on the nature and performance of this class of codes. Over the last year, I have looked at over two dozen personal codes from more than a dozen research institutions. The codes cover a variety of scientific fields, including astronomy, atmospheric sciences, bioinformatics, biology, chemistry, geology, and physics. The sources range from a few hundred lines to more than ten thousand lines, and are written in Fortran, Fortran 90, C, and C++. For the most part, the codes are modular, documented, and written in a clear, straightforward manner. They do not use complex language features, advanced data structures, programming tricks, or libraries. I had little trouble understanding what the codes did or how data structures were used. Most came with a makefile. Surprisingly, only one of the applications is parallel. All developers have access to parallel machines, so availability is not an issue. Several tried to parallelize their applications, but stopped after encountering difficulties. Lack of education and a perception that parallelism is difficult prevented most from trying. I parallelized several of the codes using OpenMP, and did not judge any of the codes as difficult to parallelize. Even more surprising than the lack of parallelism is the inefficiency of the codes. I was able to get large improvements in performance in a matter of a few days applying simple optimization techniques. Table 1 lists ten representative codes [names and affiliation are omitted to preserve anonymity]. Improvements on one processor range from 2x to 15.5x with a simple average of 4.75x. I did not use sophisticated performance tools or drill deep into the program's execution character as one would do when tuning ISV or community codes. Using only a profiler and source line timers, I identified inefficient sections of code and improved their performance by inspection. The changes were at a high level. I am sure there is another factor of 2 or 3 in each code, and more if the codes are parallelized. The study’s results show that personal scientific codes are running many times slower than they should and that the problem is pervasive. Computational scientists are not sloppy programmers; however, few are trained in the art of computer programming or code optimization. I found that most have a working knowledge of some programming language and standard software engineering practices; but they do not know, or think about, how to make their programs run faster. They simply do not know the standard techniques used to make codes run faster. In fact, they do not even perceive that such techniques exist. The case studies described in this paper show that applying simple, well known techniques can significantly increase the performance of personal codes. It is important that the scientific community and the Government agencies that support scientific research find ways to better educate academic scientific programmers. The inefficiency of their codes is so bad that it is retarding both the quality and progress of scientific research. # cacheperformance redundantoperations loopstructures performanceimprovement 1 x x 15.5 2 x 2.8 3 x x 2.5 4 x 2.1 5 x x 2.0 6 x 5.0 7 x 5.8 8 x 6.3 9 2.2 10 x x 3.3 Table 1 — Area of improvement and performance gains of 10 codes The remainder of the paper is organized as follows: sections 2, 3, and 4 discuss the three most common sources of inefficiencies in the codes studied. These are cache performance, redundant operations, and loop structures. Each section includes several examples. The last section summaries the work and suggests a possible solution to the issues raised. Optimizing cache performance Commodity microprocessor systems use caches to increase memory bandwidth and reduce memory latencies. Typical latencies from processor to L1, L2, local, and remote memory are 3, 10, 50, and 200 cycles, respectively. Moreover, bandwidth falls off dramatically as memory distances increase. Programs that do not use cache effectively run many times slower than programs that do. When optimizing for cache, the biggest performance gains are achieved by accessing data in cache order and reusing data to amortize the overhead of cache misses. Secondary considerations are prefetching, associativity, and replacement; however, the understanding and analysis required to optimize for the latter are probably beyond the capabilities of the non-expert. Much can be gained simply by accessing data in the correct order and maximizing data reuse. 6 out of the 10 codes studied here benefited from such high level optimizations. Array Accesses The most important cache optimization is the most basic: accessing Fortran array elements in column order and C array elements in row order. Four of the ten codes—1, 2, 4, and 10—got it wrong. Compilers will restructure nested loops to optimize cache performance, but may not do so if the loop structure is too complex, or the loop body includes conditionals, complex addressing, or function calls. In code 1, the compiler failed to invert a key loop because of complex addressing do I = 0, 1010, delta_x IM = I - delta_x IP = I + delta_x do J = 5, 995, delta_x JM = J - delta_x JP = J + delta_x T1 = CA1(IP, J) + CA1(I, JP) T2 = CA1(IM, J) + CA1(I, JM) S1 = T1 + T2 - 4 * CA1(I, J) CA(I, J) = CA1(I, J) + D * S1 end do end do In code 2, the culprit is conditionals do I = 1, N do J = 1, N If (IFLAG(I,J) .EQ. 0) then T1 = Value(I, J-1) T2 = Value(I-1, J) T3 = Value(I, J) T4 = Value(I+1, J) T5 = Value(I, J+1) Value(I,J) = 0.25 * (T1 + T2 + T5 + T4) Delta = ABS(T3 - Value(I,J)) If (Delta .GT. MaxDelta) MaxDelta = Delta endif enddo enddo I fixed both programs by inverting the loops by hand. Code 10 has three-dimensional arrays and triply nested loops. The structure of the most computationally intensive loops is too complex to invert automatically or by hand. The only practical solution is to transpose the arrays so that the dimension accessed by the innermost loop is in cache order. The arrays can be transposed at construction or prior to entering a computationally intensive section of code. The former requires all array references to be modified, while the latter is cost effective only if the cost of the transpose is amortized over many accesses. I used the second approach to optimize code 10. Code 5 has four-dimensional arrays and loops are nested four deep. For all of the reasons cited above the compiler is not able to restructure three key loops. Assume C arrays and let the four dimensions of the arrays be i, j, k, and l. In the original code, the index structure of the three loops is L1: for i L2: for i L3: for i for l for l for j for k for j for k for j for k for l So only L3 accesses array elements in cache order. L1 is a very complex loop—much too complex to invert. I brought the loop into cache alignment by transposing the second and fourth dimensions of the arrays. Since the code uses a macro to compute all array indexes, I effected the transpose at construction and changed the macro appropriately. The dimensions of the new arrays are now: i, l, k, and j. L3 is a simple loop and easily inverted. L2 has a loop-carried scalar dependence in k. By promoting the scalar name that carries the dependence to an array, I was able to invert the third and fourth subloops aligning the loop with cache. Code 5 is by far the most difficult of the four codes to optimize for array accesses; but the knowledge required to fix the problems is no more than that required for the other codes. I would judge this code at the limits of, but not beyond, the capabilities of appropriately trained computational scientists. Array Strides When a cache miss occurs, a line (64 bytes) rather than just one word is loaded into the cache. If data is accessed stride 1, than the cost of the miss is amortized over 8 words. Any stride other than one reduces the cost savings. Two of the ten codes studied suffered from non-unit strides. The codes represent two important classes of "strided" codes. Code 1 employs a multi-grid algorithm to reduce time to convergence. The grids are every tenth, fifth, second, and unit element. Since time to convergence is inversely proportional to the distance between elements, coarse grids converge quickly providing good starting values for finer grids. The better starting values further reduce the time to convergence. The downside is that grids of every nth element, n > 1, introduce non-unit strides into the computation. In the original code, much of the savings of the multi-grid algorithm were lost due to this problem. I eliminated the problem by compressing (copying) coarse grids into continuous memory, and rewriting the computation as a function of the compressed grid. On convergence, I copied the final values of the compressed grid back to the original grid. The savings gained from unit stride access of the compressed grid more than paid for the cost of copying. Using compressed grids, the loop from code 1 included in the previous section becomes do j = 1, GZ do i = 1, GZ T1 = CA(i+0, j-1) + CA(i-1, j+0) T4 = CA1(i+1, j+0) + CA1(i+0, j+1) S1 = T1 + T4 - 4 * CA1(i+0, j+0) CA(i+0, j+0) = CA1(i+0, j+0) + DD * S1 enddo enddo where CA and CA1 are compressed arrays of size GZ. Code 7 traverses a list of objects selecting objects for later processing. The labels of the selected objects are stored in an array. The selection step has unit stride, but the processing steps have irregular stride. A fix is to save the parameters of the selected objects in temporary arrays as they are selected, and pass the temporary arrays to the processing functions. The fix is practical if the same parameters are used in selection as in processing, or if processing comprises a series of distinct steps which use overlapping subsets of the parameters. Both conditions are true for code 7, so I achieved significant improvement by copying parameters to temporary arrays during selection. Data reuse In the previous sections, we optimized for spatial locality. It is also important to optimize for temporal locality. Once read, a datum should be used as much as possible before it is forced from cache. Loop fusion and loop unrolling are two techniques that increase temporal locality. Unfortunately, both techniques increase register pressure—as loop bodies become larger, the number of registers required to hold temporary values grows. Once register spilling occurs, any gains evaporate quickly. For multiprocessors with small register sets or small caches, the sweet spot can be very small. In the ten codes presented here, I found no opportunities for loop fusion and only two opportunities for loop unrolling (codes 1 and 3). In code 1, unrolling the outer and inner loop one iteration increases the number of result values computed by the loop body from 1 to 4, do J = 1, GZ-2, 2 do I = 1, GZ-2, 2 T1 = CA1(i+0, j-1) + CA1(i-1, j+0) T2 = CA1(i+1, j-1) + CA1(i+0, j+0) T3 = CA1(i+0, j+0) + CA1(i-1, j+1) T4 = CA1(i+1, j+0) + CA1(i+0, j+1) T5 = CA1(i+2, j+0) + CA1(i+1, j+1) T6 = CA1(i+1, j+1) + CA1(i+0, j+2) T7 = CA1(i+2, j+1) + CA1(i+1, j+2) S1 = T1 + T4 - 4 * CA1(i+0, j+0) S2 = T2 + T5 - 4 * CA1(i+1, j+0) S3 = T3 + T6 - 4 * CA1(i+0, j+1) S4 = T4 + T7 - 4 * CA1(i+1, j+1) CA(i+0, j+0) = CA1(i+0, j+0) + DD * S1 CA(i+1, j+0) = CA1(i+1, j+0) + DD * S2 CA(i+0, j+1) = CA1(i+0, j+1) + DD * S3 CA(i+1, j+1) = CA1(i+1, j+1) + DD * S4 enddo enddo The loop body executes 12 reads, whereas as the rolled loop shown in the previous section executes 20 reads to compute the same four values. In code 3, two loops are unrolled 8 times and one loop is unrolled 4 times. Here is the before for (k = 0; k < NK[u]; k++) { sum = 0.0; for (y = 0; y < NY; y++) { sum += W[y][u][k] * delta[y]; } backprop[i++]=sum; } and after code for (k = 0; k < KK - 8; k+=8) { sum0 = 0.0; sum1 = 0.0; sum2 = 0.0; sum3 = 0.0; sum4 = 0.0; sum5 = 0.0; sum6 = 0.0; sum7 = 0.0; for (y = 0; y < NY; y++) { sum0 += W[y][0][k+0] * delta[y]; sum1 += W[y][0][k+1] * delta[y]; sum2 += W[y][0][k+2] * delta[y]; sum3 += W[y][0][k+3] * delta[y]; sum4 += W[y][0][k+4] * delta[y]; sum5 += W[y][0][k+5] * delta[y]; sum6 += W[y][0][k+6] * delta[y]; sum7 += W[y][0][k+7] * delta[y]; } backprop[k+0] = sum0; backprop[k+1] = sum1; backprop[k+2] = sum2; backprop[k+3] = sum3; backprop[k+4] = sum4; backprop[k+5] = sum5; backprop[k+6] = sum6; backprop[k+7] = sum7; } for one of the loops unrolled 8 times. Optimizing for temporal locality is the most difficult optimization considered in this paper. The concepts are not difficult, but the sweet spot is small. Identifying where the program can benefit from loop unrolling or loop fusion is not trivial. Moreover, it takes some effort to get it right. Still, educating scientific programmers about temporal locality and teaching them how to optimize for it will pay dividends. Reducing instruction count Execution time is a function of instruction count. Reduce the count and you usually reduce the time. The best solution is to use a more efficient algorithm; that is, an algorithm whose order of complexity is smaller, that converges quicker, or is more accurate. Optimizing source code without changing the algorithm yields smaller, but still significant, gains. This paper considers only the latter because the intent is to study how much better codes can run if written by programmers schooled in basic code optimization techniques. The ten codes studied benefited from three types of "instruction reducing" optimizations. The two most prevalent were hoisting invariant memory and data operations out of inner loops. The third was eliminating unnecessary data copying. The nature of these inefficiencies is language dependent. Memory operations The semantics of C make it difficult for the compiler to determine all the invariant memory operations in a loop. The problem is particularly acute for loops in functions since the compiler may not know the values of the function's parameters at every call site when compiling the function. Most compilers support pragmas to help resolve ambiguities; however, these pragmas are not comprehensive and there is no standard syntax. To guarantee that invariant memory operations are not executed repetitively, the user has little choice but to hoist the operations by hand. The problem is not as severe in Fortran programs because in the absence of equivalence statements, it is a violation of the language's semantics for two names to share memory. Codes 3 and 5 are C programs. In both cases, the compiler did not hoist all invariant memory operations from inner loops. Consider the following loop from code 3 for (y = 0; y < NY; y++) { i = 0; for (u = 0; u < NU; u++) { for (k = 0; k < NK[u]; k++) { dW[y][u][k] += delta[y] * I1[i++]; } } } Since dW[y][u] can point to the same memory space as delta for one or more values of y and u, assignment to dW[y][u][k] may change the value of delta[y]. In reality, dW and delta do not overlap in memory, so I rewrote the loop as for (y = 0; y < NY; y++) { i = 0; Dy = delta[y]; for (u = 0; u < NU; u++) { for (k = 0; k < NK[u]; k++) { dW[y][u][k] += Dy * I1[i++]; } } } Failure to hoist invariant memory operations may be due to complex address calculations. If the compiler can not determine that the address calculation is invariant, then it can hoist neither the calculation nor the associated memory operations. As noted above, code 5 uses a macro to address four-dimensional arrays #define MAT4D(a,q,i,j,k) (double *)((a)->data + (q)*(a)->strides[0] + (i)*(a)->strides[3] + (j)*(a)->strides[2] + (k)*(a)->strides[1]) The macro is too complex for the compiler to understand and so, it does not identify any subexpressions as loop invariant. The simplest way to eliminate the address calculation from the innermost loop (over i) is to define a0 = MAT4D(a,q,0,j,k) before the loop and then replace all instances of *MAT4D(a,q,i,j,k) in the loop with a0[i] A similar problem appears in code 6, a Fortran program. The key loop in this program is do n1 = 1, nh nx1 = (n1 - 1) / nz + 1 nz1 = n1 - nz * (nx1 - 1) do n2 = 1, nh nx2 = (n2 - 1) / nz + 1 nz2 = n2 - nz * (nx2 - 1) ndx = nx2 - nx1 ndy = nz2 - nz1 gxx = grn(1,ndx,ndy) gyy = grn(2,ndx,ndy) gxy = grn(3,ndx,ndy) balance(n1,1) = balance(n1,1) + (force(n2,1) * gxx + force(n2,2) * gxy) * h1 balance(n1,2) = balance(n1,2) + (force(n2,1) * gxy + force(n2,2) * gyy)*h1 end do end do The programmer has written this loop well—there are no loop invariant operations with respect to n1 and n2. However, the loop resides within an iterative loop over time and the index calculations are independent with respect to time. Trading space for time, I precomputed the index values prior to the entering the time loop and stored the values in two arrays. I then replaced the index calculations with reads of the arrays. Data operations Ways to reduce data operations can appear in many forms. Implementing a more efficient algorithm produces the biggest gains. The closest I came to an algorithm change was in code 4. This code computes the inner product of K-vectors A(i) and B(j), 0 = i < N, 0 = j < M, for most values of i and j. Since the program computes most of the NM possible inner products, it is more efficient to compute all the inner products in one triply-nested loop rather than one at a time when needed. The savings accrue from reading A(i) once for all B(j) vectors and from loop unrolling. for (i = 0; i < N; i+=8) { for (j = 0; j < M; j++) { sum0 = 0.0; sum1 = 0.0; sum2 = 0.0; sum3 = 0.0; sum4 = 0.0; sum5 = 0.0; sum6 = 0.0; sum7 = 0.0; for (k = 0; k < K; k++) { sum0 += A[i+0][k] * B[j][k]; sum1 += A[i+1][k] * B[j][k]; sum2 += A[i+2][k] * B[j][k]; sum3 += A[i+3][k] * B[j][k]; sum4 += A[i+4][k] * B[j][k]; sum5 += A[i+5][k] * B[j][k]; sum6 += A[i+6][k] * B[j][k]; sum7 += A[i+7][k] * B[j][k]; } C[i+0][j] = sum0; C[i+1][j] = sum1; C[i+2][j] = sum2; C[i+3][j] = sum3; C[i+4][j] = sum4; C[i+5][j] = sum5; C[i+6][j] = sum6; C[i+7][j] = sum7; }} This change requires knowledge of a typical run; i.e., that most inner products are computed. The reasons for the change, however, derive from basic optimization concepts. It is the type of change easily made at development time by a knowledgeable programmer. In code 5, we have the data version of the index optimization in code 6. Here a very expensive computation is a function of the loop indices and so cannot be hoisted out of the loop; however, the computation is invariant with respect to an outer iterative loop over time. We can compute its value for each iteration of the computation loop prior to entering the time loop and save the values in an array. The increase in memory required to store the values is small in comparison to the large savings in time. The main loop in Code 8 is doubly nested. The inner loop includes a series of guarded computations; some are a function of the inner loop index but not the outer loop index while others are a function of the outer loop index but not the inner loop index for (j = 0; j < N; j++) { for (i = 0; i < M; i++) { r = i * hrmax; R = A[j]; temp = (PRM[3] == 0.0) ? 1.0 : pow(r, PRM[3]); high = temp * kcoeff * B[j] * PRM[2] * PRM[4]; low = high * PRM[6] * PRM[6] / (1.0 + pow(PRM[4] * PRM[6], 2.0)); kap = (R > PRM[6]) ? high * R * R / (1.0 + pow(PRM[4]*r, 2.0) : low * pow(R/PRM[6], PRM[5]); < rest of loop omitted > }} Note that the value of temp is invariant to j. Thus, we can hoist the computation for temp out of the loop and save its values in an array. for (i = 0; i < M; i++) { r = i * hrmax; TEMP[i] = pow(r, PRM[3]); } [N.B. – the case for PRM[3] = 0 is omitted and will be reintroduced later.] We now hoist out of the inner loop the computations invariant to i. Since the conditional guarding the value of kap is invariant to i, it behooves us to hoist the computation out of the inner loop, thereby executing the guard once rather than M times. The final version of the code is for (j = 0; j < N; j++) { R = rig[j] / 1000.; tmp1 = kcoeff * par[2] * beta[j] * par[4]; tmp2 = 1.0 + (par[4] * par[4] * par[6] * par[6]); tmp3 = 1.0 + (par[4] * par[4] * R * R); tmp4 = par[6] * par[6] / tmp2; tmp5 = R * R / tmp3; tmp6 = pow(R / par[6], par[5]); if ((par[3] == 0.0) && (R > par[6])) { for (i = 1; i <= imax1; i++) KAP[i] = tmp1 * tmp5; } else if ((par[3] == 0.0) && (R <= par[6])) { for (i = 1; i <= imax1; i++) KAP[i] = tmp1 * tmp4 * tmp6; } else if ((par[3] != 0.0) && (R > par[6])) { for (i = 1; i <= imax1; i++) KAP[i] = tmp1 * TEMP[i] * tmp5; } else if ((par[3] != 0.0) && (R <= par[6])) { for (i = 1; i <= imax1; i++) KAP[i] = tmp1 * TEMP[i] * tmp4 * tmp6; } for (i = 0; i < M; i++) { kap = KAP[i]; r = i * hrmax; < rest of loop omitted > } } Maybe not the prettiest piece of code, but certainly much more efficient than the original loop, Copy operations Several programs unnecessarily copy data from one data structure to another. This problem occurs in both Fortran and C programs, although it manifests itself differently in the two languages. Code 1 declares two arrays—one for old values and one for new values. At the end of each iteration, the array of new values is copied to the array of old values to reset the data structures for the next iteration. This problem occurs in Fortran programs not included in this study and in both Fortran 77 and Fortran 90 code. Introducing pointers to the arrays and swapping pointer values is an obvious way to eliminate the copying; but pointers is not a feature that many Fortran programmers know well or are comfortable using. An easy solution not involving pointers is to extend the dimension of the value array by 1 and use the last dimension to differentiate between arrays at different times. For example, if the data space is N x N, declare the array (N, N, 2). Then store the problem’s initial values in (_, _, 2) and define the scalar names new = 2 and old = 1. At the start of each iteration, swap old and new to reset the arrays. The old–new copy problem did not appear in any C program. In programs that had new and old values, the code swapped pointers to reset data structures. Where unnecessary coping did occur is in structure assignment and parameter passing. Structures in C are handled much like scalars. Assignment causes the data space of the right-hand name to be copied to the data space of the left-hand name. Similarly, when a structure is passed to a function, the data space of the actual parameter is copied to the data space of the formal parameter. If the structure is large and the assignment or function call is in an inner loop, then copying costs can grow quite large. While none of the ten programs considered here manifested this problem, it did occur in programs not included in the study. A simple fix is always to refer to structures via pointers. Optimizing loop structures Since scientific programs spend almost all their time in loops, efficient loops are the key to good performance. Conditionals, function calls, little instruction level parallelism, and large numbers of temporary values make it difficult for the compiler to generate tightly packed, highly efficient code. Conditionals and function calls introduce jumps that disrupt code flow. Users should eliminate or isolate conditionls to their own loops as much as possible. Often logical expressions can be substituted for if-then-else statements. For example, code 2 includes the following snippet MaxDelta = 0.0 do J = 1, N do I = 1, M < code omitted > Delta = abs(OldValue ? NewValue) if (Delta > MaxDelta) MaxDelta = Delta enddo enddo if (MaxDelta .gt. 0.001) goto 200 Since the only use of MaxDelta is to control the jump to 200 and all that matters is whether or not it is greater than 0.001, I made MaxDelta a boolean and rewrote the snippet as MaxDelta = .false. do J = 1, N do I = 1, M < code omitted > Delta = abs(OldValue ? NewValue) MaxDelta = MaxDelta .or. (Delta .gt. 0.001) enddo enddo if (MaxDelta) goto 200 thereby, eliminating the conditional expression from the inner loop. A microprocessor can execute many instructions per instruction cycle. Typically, it can execute one or more memory, floating point, integer, and jump operations. To be executed simultaneously, the operations must be independent. Thick loops tend to have more instruction level parallelism than thin loops. Moreover, they reduce memory traffice by maximizing data reuse. Loop unrolling and loop fusion are two techniques to increase the size of loop bodies. Several of the codes studied benefitted from loop unrolling, but none benefitted from loop fusion. This observation is not too surpising since it is the general tendency of programmers to write thick loops. As loops become thicker, the number of temporary values grows, increasing register pressure. If registers spill, then memory traffic increases and code flow is disrupted. A thick loop with many temporary values may execute slower than an equivalent series of thin loops. The biggest gain will be achieved if the thick loop can be split into a series of independent loops eliminating the need to write and read temporary arrays. I found such an occasion in code 10 where I split the loop do i = 1, n do j = 1, m A24(j,i)= S24(j,i) * T24(j,i) + S25(j,i) * U25(j,i) B24(j,i)= S24(j,i) * T25(j,i) + S25(j,i) * U24(j,i) A25(j,i)= S24(j,i) * C24(j,i) + S25(j,i) * V24(j,i) B25(j,i)= S24(j,i) * U25(j,i) + S25(j,i) * V25(j,i) C24(j,i)= S26(j,i) * T26(j,i) + S27(j,i) * U26(j,i) D24(j,i)= S26(j,i) * T27(j,i) + S27(j,i) * V26(j,i) C25(j,i)= S27(j,i) * S28(j,i) + S26(j,i) * U28(j,i) D25(j,i)= S27(j,i) * T28(j,i) + S26(j,i) * V28(j,i) end do end do into two disjoint loops do i = 1, n do j = 1, m A24(j,i)= S24(j,i) * T24(j,i) + S25(j,i) * U25(j,i) B24(j,i)= S24(j,i) * T25(j,i) + S25(j,i) * U24(j,i) A25(j,i)= S24(j,i) * C24(j,i) + S25(j,i) * V24(j,i) B25(j,i)= S24(j,i) * U25(j,i) + S25(j,i) * V25(j,i) end do end do do i = 1, n do j = 1, m C24(j,i)= S26(j,i) * T26(j,i) + S27(j,i) * U26(j,i) D24(j,i)= S26(j,i) * T27(j,i) + S27(j,i) * V26(j,i) C25(j,i)= S27(j,i) * S28(j,i) + S26(j,i) * U28(j,i) D25(j,i)= S27(j,i) * T28(j,i) + S26(j,i) * V28(j,i) end do end do Conclusions Over the course of the last year, I have had the opportunity to work with over two dozen academic scientific programmers at leading research universities. Their research interests span a broad range of scientific fields. Except for two programs that relied almost exclusively on library routines (matrix multiply and fast Fourier transform), I was able to improve significantly the single processor performance of all codes. Improvements range from 2x to 15.5x with a simple average of 4.75x. Changes to the source code were at a very high level. I did not use sophisticated techniques or programming tools to discover inefficiencies or effect the changes. Only one code was parallel despite the availability of parallel systems to all developers. Clearly, we have a problem—personal scientific research codes are highly inefficient and not running parallel. The developers are unaware of simple optimization techniques to make programs run faster. They lack education in the art of code optimization and parallel programming. I do not believe we can fix the problem by publishing additional books or training manuals. To date, the developers in questions have not studied the books or manual available, and are unlikely to do so in the future. Short courses are a possible solution, but I believe they are too concentrated to be much use. The general concepts can be taught in a three or four day course, but that is not enough time for students to practice what they learn and acquire the experience to apply and extend the concepts to their codes. Practice is the key to becoming proficient at optimization. I recommend that graduate students be required to take a semester length course in optimization and parallel programming. We would never give someone access to state-of-the-art scientific equipment costing hundreds of thousands of dollars without first requiring them to demonstrate that they know how to use the equipment. Yet the criterion for time on state-of-the-art supercomputers is at most an interesting project. Requestors are never asked to demonstrate that they know how to use the system, or can use the system effectively. A semester course would teach them the required skills. Government agencies that fund academic scientific research pay for most of the computer systems supporting scientific research as well as the development of most personal scientific codes. These agencies should require graduate schools to offer a course in optimization and parallel programming as a requirement for funding. About the Author John Feo received his Ph.D. in Computer Science from The University of Texas at Austin in 1986. After graduate school, Dr. Feo worked at Lawrence Livermore National Laboratory where he was the Group Leader of the Computer Research Group and principal investigator of the Sisal Language Project. In 1997, Dr. Feo joined Tera Computer Company where he was project manager for the MTA, and oversaw the programming and evaluation of the MTA at the San Diego Supercomputer Center. In 2000, Dr. Feo joined Sun Microsystems as an HPC application specialist. He works with university research groups to optimize and parallelize scientific codes. Dr. Feo has published over two dozen research articles in the areas of parallel parallel programming, parallel programming languages, and application performance.

    Read the article

  • Java Spotlight Episode 103: 2012 Duke Choice Award Winners

    - by Roger Brinkley
    Our annual interview with the 2012 Duke Choice Award Winners recorded live at the JavaOne 2012. Right-click or Control-click to download this MP3 file. You can also subscribe to the Java Spotlight Podcast Feed to get the latest podcast automatically. If you use iTunes you can open iTunes and subscribe with this link:  Java Spotlight Podcast in iTunes. Show Notes Events Oct 13, Devoxx 4 Kids Nederlands Oct 15-17, JAX London Oct 20, Devoxx 4 Kids Français Oct 22-23, Freescale Technology Forum - Japan, Tokyo Oct 30-Nov 1, Arm TechCon, Santa Clara Oct 31, JFall, Netherlands Nov 2-3, JMagreb, Morocco Nov 13-17, Devoxx, Belgium Feature Interview Duke Choice Award Winners 2012 - Show Presentation London Java CommunityThe second user group receiving a Duke’s Choice Award this year, the London Java Community (LJC) and its users have been active in the OpenJDK, the Java Community Process (JCP) and other efforts within the global Java community. Student Nokia Developer GroupThis year’s student winner, Ram Kashyap, is the founder and president of the Nokia Student Network, and was profiled in the “The New Java Developers” feature in the March/April 2012 issue of Java Magazine. Since then, Ram has maintained a hectic pace, graduating from the People’s Education Society Institute of Technology in Bangalore, India, while working on a Java mobile startup and training students on Java ME. Jelastic, Inc.Moving existing Java applications to the cloud can be a daunting task, but startup Jelastic, Inc. offers the first all-Java platform-as-a-service (PaaS) that enables existing Java applications to be deployed in the cloud without code changes or lock-in. NATOThe first-ever Community Choice Award goes to the MASE Integrated Console Environment (MICE) in use at NATO. Built in Java on the NetBeans platform, MICE provides a high-performance visualization environment for conducting air defense and battle-space operations. DuchessRather than focus on a specific geographic area like most Java User Groups (JUGs), Duchess fosters the participation of women in the Java community worldwide. The group has more than 500 members in 60 countries, and provides a platform through which women can connect with each other and get involved in all aspects of the Java community. AgroSense ProjectImproving farming methods to feed a hungry world is the goal of AgroSense, an open source farm information management system built in Java and the NetBeans platform. AgroSense enables farmers, agribusinesses, suppliers and others to develop modular applications that will easily exchange information through a common underlying NetBeans framework. Apache Software Foundation Hadoop ProjectThe Apache Software Foundation’s Hadoop project, written in Java, provides a framework for distributed processing of big data sets across clusters of computers, ranging from a few servers to thousands of machines. This harnessing of large data pools allows organizations to better understand and improve their business. Parleys.comE-learning specialist Parleys.com, based in Brussels, Belgium, uses Java technologies to bring online classes and full IT conferences to desktops, laptops, tablets and mobile devices. Parleys.com has hosted more than 1,700 conferences—including Devoxx and JavaOne—for more than 800,000 unique visitors. Winners not presenting at JavaOne 2012 Duke Choice Awards BOF Liquid RoboticsRobotics – Liquid Robotics is an ocean data services provider whose Wave Glider technology collects information from the world’s oceans for application in government, science and commercial applications. The organization features the “father of Java” James Gosling as its chief software architect.United Nations High Commissioner for RefugeesThe United Nations High Commissioner for Refugees (UNHCR) is on the front lines of crises around the world, from civil wars to natural disasters. To help facilitate its mission of humanitarian relief, the UNHCR has developed a light-client Java application on the NetBeans platform. The Level One registration tool enables the UNHCR to collect information on the number of refugees and their water, food, housing, health, and other needs in the field, and combines that with geocoding information from various sources. This enables the UNHCR to deliver the appropriate kind and amount of assistance where it is needed.

    Read the article

  • BIP BIServer Query Debug

    - by Tim Dexter
    With some help from Bryan, I have uncovered a way of being able to debug or at least log what BIServer is doing when BIP sends it a query request. This is not for those of you querying the database directly but if you are using the BIServer and its datamodel to fetch data for a BIP report. If you have written or used the query builder against BIServer and when you run the report it chokes with a cryptic message, that you have no clue about, read on. When BIP runs a piece of BIServer logical SQL to fetch data. It does not appear to validate it, it just passes it through, so what is BIServer doing on its end? As you may know, you are not writing regular physical sql its actually logical sql e.g. select Jobs."Job Title" as "Job Title", Employees."Last Name" as "Last Name", Employees.Salary as Salary, Locations."Department Name" as "Department Name", Locations."Country Name" as "Country Name", Locations."Region Name" as "Region Name" from HR.Locations Locations, HR.Employees Employees, HR.Jobs Jobs The tables might not even be a physical tables, we don't care, that's what the BIServer and its model are for. You have put all the effort into building the model, just go get me the data from where ever it might be. The BIServer takes the logical sql and uses its vast brain to work out what the physical SQL is, executes it and passes the result back to BIP. select distinct T32556.JOB_TITLE as c1, T32543.LAST_NAME as c2, T32543.SALARY as c3, T32537.DEPARTMENT_NAME as c4, T32532.COUNTRY_NAME as c5, T32577.REGION_NAME as c6 from JOBS T32556, REGIONS T32577, COUNTRIES T32532, LOCATIONS T32569, DEPARTMENTS T32537, EMPLOYEES T32543 where ( T32532.COUNTRY_ID = T32569.COUNTRY_ID and T32532.REGION_ID = T32577.REGION_ID and T32537.DEPARTMENT_ID = T32543.DEPARTMENT_ID and T32537.LOCATION_ID = T32569.LOCATION_ID and T32543.JOB_ID = T32556.JOB_ID ) Not a very tough example I know but you get the idea. How do I know what the BIServer is up to? How can I find out what the issue might be if BIServer chokes on my query? There are a couple of steps: In the Administrator tool you need to set the logging level for the Administrator user to something greater than the default '0'. '7' is going to give you the max. Just remember to take it back down after you have finished the debug. I needed to bounce my BIServer service Now here's the secret sauce. Prefix the following to your BIP query set variable LOGLEVEL = 7; Set the log level to that you have in the admin tool Now run your BIP report. With the prefix in place; BIServer will write to the NQQuery.log file. This is located in the ./OracleBI/server/Log directory. In there you are going to find the complete process the BIServer has gone through to try and get the data back for you A quick note, if the BIServer can, its going to hit that great BIEE cache to get your data and you may not see the full log. IF this is the case. Get inot hte Administration page (via the browser login) and clear out your BIP report cursor. Then re-run. This will hopefully help out if you are trying to debug that annoying BIP report that will not run or is getting some strange data. Don't forget to turn that logging level back down once you are done. This will avoid the DBA screaming at you for sucking up all the disk space on the system.

    Read the article

  • Archiving SQLHelp tweets

    - by jamiet
    #SQLHelp is a Twitter hashtag that can be used by any Twitter user to get help from the SQL Server community. I think its fair to say that in its first year of being it has proved to be a very useful resource however Kendra Little (@kendra_little) made a very salient point yesterday when she tweeted: Is there a way to search the archives of #sqlhelp Trying to remember answer to a question I know I saw a couple months ago http://twitter.com/#!/Kendra_Little/status/15538234184441856 This highlights an inherent problem with Twitter’s search capability – it simply does not reach far enough back in time. I have made steps to remedy that situation by putting into place two initiatives to archive Tweets that contain the #sqlhelp hashtag. The Archivist http://archivist.visitmix.com/ is a free service that, quite simply, archives a history of tweets that contain a given search term by periodically polling Twitter’s search service with that search term and subsequently displaying a dashboard providing an aggregate view of those tweets for things like tweet volume over time, top users and top words (Archivist FAQ). I have set up an archive on The Archivist for “sqlhelp” which you can view at http://archivist.visitmix.com/jamiet/7. Here is a screenshot of the SQLHelp dashboard 36 minutes after I set it up: There is lots of good information in there, including the fact that Jonathan Kehayias (@SQLSarg) is the most active SQLHelp tweeter (I suspect as an answerer rather than a questioner ) and that SSIS has proven to be a rather (ahem) popular subject!! Datasift The Archivist has its uses though for our purposes it has a couple of downsides. For starters you cannot search through an archive (which is what Kendra was after) and nor can you export the contents of the archive for offline analysis. For those functions we need something a bit more heavyweight and for that I present to you Datasift. Datasift is a tool (currently an alpha release) that allows you to search for tweets and provide them through an object called a Datasift stream. That sounds very similar to normal Twitter search though it has one distinct advantage that other Twitter search tools do not – Datasift has access to Twitter’s Streaming API (aka the Twitter Firehose). In addition it has access to a lot of other rather nice features: It provides the Datasift API that allows you to consume the output of a Datasift stream in your tool of choice (bring on my favourite ultimate mashup tool J ) It has a query language (called Filtered Stream Definition Language – FSDL for short) A Datasift stream can consume (and filter) other Datasift streams Datasift can (and does) consume services other than Twitter If I refer to Datasift as “ETL for tweets” then you may get some sort of idea what it is all about. Just as I did with The Archivist I have set up a publicly available Datasift stream for “sqlhelp” at http://datasift.net/stream/1581/sqlhelp. Here is the FSDL query that provides the data: twitter.text contains "sqlhelp" Pretty simple eh? At the current time it provides little more than a rudimentary dashboard but as Datasift is currently an alpha release I think this may be worth keeping an eye on. The real value though is the ability to consume the output of a stream via Datasift’s RESTful API, observe: http://api.datasift.net/stream.xml?stream_identifier=c7015255f07e982afdeebdf1ae6e3c0d&username=jamiet&api_key=XXXXXXX (Note that an api_key is required during the alpha period so, given that I’m not supplying my api_key, this URI will not work for you) Just to prove that a Datasift stream can indeed consume data from another stream I have set up a second stream that further filters the first one for tweets containing “SSIS”. That one is at http://datasift.net/stream/1586/ssis-sqlhelp and here is the FSDL query: rule "414c9845685ff8d2548999cf3162e897" and (interaction.content contains "ssis") When Datasift moves beyond alpha I’ll re-assess how useful this is going to be and post a follow-up blog. @Jamiet

    Read the article

  • Silverlight Cream for December 13, 2010 -- #1010

    - by Dave Campbell
    In this Issue: Rénald Nollet, Benjamin Gavin, Dennis Doomen, Tim Greenfield, Mike Taulty, Jeff Blankenburg, Michael Crump, Laurent Duveau, Dragos Manolescu, KeyboardP, Yochay Kiriaty. Above the Fold: Silverlight: "Silverlight RIA Services and Basic, Anonymous Authentication" Benjamin Gavin WP7: "lving Circular Navigation in Windows Phone Silverlight Applications" Yochay Kiriaty SQL Azure: "SQL Azure Database Manager – Part 1 : How to connect to your SQL Azure DB" Rénald Nollet Shoutouts: Yochay Kiriaty has a post up on the Windows Phone Devloper Blog about open source (MSPL) projects helping WP7 devs: Windows Phone Recipes – Helping the Community Jesse Liberty's latest Yet Another Podcast is up and thie time it's Joe Stagner: Yet Another Podcast #18 – Joe Stagner Josh Schwartzberg sent me this link to what is apparently his yearly web-only rock Christmas album: MetalXmas... done in Silverlight and RIA Services From SilverlightCream.com: SQL Azure Database Manager – Part 1 : How to connect to your SQL Azure DB Rénald Nollet posted Part 1 of a series on a SQL Azure database manager all in Silverlight... has a live demo running, some description, and is making us wait for the next part! Silverlight RIA Services and Basic, Anonymous Authentication Benjamin Gavin has a quick post up resolving a basic RIA Services problem that I bet a lot of folks are looking for the answer on... like 500 series errors... cool little find he ferreted out... A night of Silverlight, WPF, unit testing and Caliburn Micro Dennis Doomen in concert with his employer gave a couple talks at the local DotNED user group, and covered literally a cornucopia of topics... slides, and example code for both talks... lotsa material here... Tim Greenfield on PuzzleTouch WP7 Application Tim Greenfield is the latest WP7 app developer to be interviewed by the SilverlightShow crew... lots of interesting comments and insight from Tim. Rebuilding the PDC 2010 Silverlight Application (Part 4) Mike Taulty has part 4 of his PDC 2010 Silverlight app construction project up and is taking the app into Blend, and the considerations that brought to the table. What I Learned In WP7 – Issue #2 Jeff Blankenburg continues his "What I Learned" series with this discussion about fonts, the Non-Linear Navigation service I mention below, and possible WP7 jobs. Part 3 of 4 : Tips/Tricks for Silverlight Developers Michael Crump has Part 3 of his Tips/Tricks up today. Lots of goodies this time: underlining in a TextBlock, getting browser info, startup params, VisualTreeHelper, and child windows. My Windows Phone 7 presentation in Montreal Laurent Duveau gave a WP7 presentation in Montreal as part of the Microsoft Windows Phone 7 Developer's Briefing, and has posted his materials and slide deck WP7 Code: Mocking Event Streams with IEnumerable Dragos Manolescu has a very cool post up on using IEnumerable to Mock event streams by leveraging the IObservable/IEnumerable duality, and uses the 2D bubble app that you can run and test in the emulator without needing an accelerometer Transparent Wallpapers – Video Tutorial KeyboardP has had so many queries about his Transparent wallpaper for WP7 that he produced a video tutorial for it... Solving Circular Navigation in Windows Phone Silverlight Applications Yochay Kiriaty discusses the first recipe they are releasing ... see the shoutout above, a Nonlinear Navigation Service ... to help with apps that have loops in navigation. Stay in the 'Light! Twitter SilverlightNews | Twitter WynApse | WynApse.com | Tagged Posts | SilverlightCream Join me @ SilverlightCream | Phoenix Silverlight User Group Technorati Tags: Silverlight    Silverlight 3    Silverlight 4    Windows Phone MIX10

    Read the article

< Previous Page | 525 526 527 528 529 530 531 532 533 534 535 536  | Next Page >