Search Results

Search found 6605 results on 265 pages for 'complex networks'.

Page 50/265 | < Previous Page | 46 47 48 49 50 51 52 53 54 55 56 57  | Next Page >

  • XNA and C# vs the 360's in order processor

    - by Richard Fabian
    Having read this rant, I feel that they're probably right (seeing as the Xenon and the CELL BE are both highly sensitive to branchy and cache ignorant code), but I've heard that people are making games fine with C# too. So, is it true that it's impossible to do anything professional with XNA, or was the poster just missing what makes XNA the killer API for game development? By professional, I don't mean in the sense that you can make money from it, but more in the sense that the more professional games have phsyics models and animation systems that seem outside the reach of a language with such definite barriers to the intrinsics. If I wanted to write a collision system or fluid dynamics engine for my game, I don't feel like C# offers me the chance to do that as the runtime code generator and the lack of intrinsics gets in the way of performance. However, many people seem to be fine working within these constraints, making their successful games but seemingly avoiding any of the problems by omission. I've not noticed any XNA games do anything complex other than what's already provided by the libraries, or handled by shaders. Is this avoidance of the more complex game dynamics because of teh limitations of C#, or just people concentrating on getting it done? In all honesty, I can't believe that AI war can maintain that many units of AI without some data oriented approach, or some dirty C# hacks to make it run better than the standard approach, and I guess that's partly my question too, how have people hacked C# so it's able to do what games coders do naturally with C++?

    Read the article

  • Oracle's SPARC T4, 007 Style

    - by Kristin Rose
    The names 4, T4, and this power house travels hand in hand with its good friend SPARC. About 6 years ago on-chip encryption acceleration was first shipped in a commercial system, the SPARC T1. Today, thanks to Oracle SPARC innovative leadership in on-chip encryption acceleration, complex cryptographic computations was born and has since rapidly evolved. Customers can now have security with performance because we my friend, are in the Age of Big Data.If you need some high speed action in your life, listen here. The SPARC T4 systems offer customers much more value for applications than just increased performance through its cross sell opportunity. This is done by enabling partners to integrate your own applications to Oracle’s SPARC T4 Servers for Cloud deployments, and providing direct business benefits that supersedes the commodity approach to data center computing such as security, performance and optimization.As companies continue down this complex path of big data, eCommerce, and mobility, the need to provide better and more in-depth security is more prominent than ever. Oracle’s SPARC T4 processor allows customers to deliver the highest levels of application security, as well as deliver the necessary level performance without added cost, and complexity.To learn more behind the value of SPARC T4, check out a more in-depth blog here. For more on the SPARC T4 family of products, click here.Encryption Lives Another Day,The OPN Communications Team 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:"Times New Roman","serif";}

    Read the article

  • What is a simple deformer in which vertices deform linearly with control points?

    - by sebf
    In my project I want to deform a complex mesh, using a simpler 'proxy' mesh. In effect, each vertex of the proxy/collision mesh will be a control point/bone, which should deform the vertices of the main mesh attached to it depending on weight, but where the weight is not dependant on the absolute distance from the control point but rather distance relative to the other affecting control points. The point of this is to preserve complex three dimensional features of the main mesh while using physics implementations which expect something far simpler, low resolution, single surface, etc. Therefore, the vertices must deform linearly with their respective weighted control points (i.e. no falloff fields or all the mesh features will end up collapsed) - as if each vertex was linked to a point on the plane created by the attached control points and deformed with it. I have tried implementing the weight computation algorithm in this paper (page 4) but it is not working as expected and I am wondering if it is really the best way to do what I want. What is the simplest way to 'skin'* an arbitrary mesh, to another arbitrary mesh? *By skin I mean I need an algorithm to determine the best control points for a vertex, and their weights.

    Read the article

  • Simultaneous AI in turn based games

    - by Eduard Strehlau
    I want to hack together a roguelike. Now I thought about entity and world representation and got to a quite big problem. If you want all the AI to act simultaneously you would normally(in cellular automa for examble) just copy the cell buffer and let all action of indiviual cells depend on the copy. Actions which are not valid anymore after some cell before the cell you are currently operating on changed the original enviourment(blocking the path) are just ignored or reapplied with the "current"(between turns) environment. After all cells have acted you copy the current map to the buffer again. Now for an environment with complex AI and big(datawise) entities the copying would take too long. So I thought you could put every action and entity makes into a que(make no changes to the environment) and execute the whole que after everyone took their move. Every interaction on this que are realy interacting entities, so if a entity tries to attack another entity it sends a message to it, the consequences of the attack would be visible next turn, either by just examining the entity or asking the entity for data. This would remove problems like what happens if an entity dies middle in the cue but got actions or is messaged later on(all messages would go to null, and the messages from the entity would either just be sent or deleted(haven't decided yet) But what would happen if a monster spawns a fireball which by itself tracks the player(in the same turn). Should I add the fireball to the enviourment beforehand, so make a change to the environment before executing the action list or just add the ball to the "need updated" list as a special case so it doesn't exist in the environment and still operates on it, spawing after evaluating the action list? Are there any solutions or papers on this subject which I can take a look at? EDIT: I don't need information on writing a roguelike I need information on turn based ai in respective to a complex enviourment.

    Read the article

  • Internship in License Contract Management

    - by cristian.condurache(at)oracle.com
    Hi Everyone, My name is Luca. I am an intern in the License Contract Management team in Italy. I have studied Economics and Business in Pescara and finished my Master’s Degree in July 2009. After a short work experience near my home town I decided to look for a job in an International Company. I got in touch with Oracle in January 2010. I had a telephone interview and then a face-to-face interview. On a cold and grey morning, I arrived in Milan....my first impression was fantastic....a big modern building with wide TVs everywhere. I was a little nervous but very excited. I understood this could be a great opportunity... The interview went well and I started to work in March. After a training period I was quickly involved in the closing of the last quarter of the fiscal year - of which May is the last month at Oracle. Working as a License Contract Manager is a real challenge for a fresh graduate. It involves thoroughly understanding the Oracle Policies and Practices with regards to License Contracts. In my experience, especially in May, I learnt to work under high pressure, within time constrains, and to keep up with constant changes. In this period I also had the opportunity to be involved in different negotiations, being directly in contact with the customers. This helped me to develop my relational skills during complex transactions. Looking back at the nine months at Oracle I can say I have a better understanding of the IT world. It is a complex environment that changes continously, offering new challenges to learn from everytime. If you have any questions related to this article feel free to contact [email protected]. You can find our job opportunities via http://campus.oracle.com. Technorati Tags: License Contract Management,oppotunity,Oracle Policies,internship

    Read the article

  • What is the best way to "carve" a terrain created from a heightmap?

    - by tigrou
    I have a 3d landscape created from a heightmap. I'd like to "carve" some holes in that terrain. That will allow me to create bridges, caverns and tunnels inside it. That operation will be done in the game editor so it doesn't need to be realtime. In the end, rendering is done using traditional polygons. What would be the best/easiest way to do that ? I already think about several solutions : Solution 1 1) Create voxels from the heightmap (very easy). In other words, fill a 3D array like this : voxels[32][32][32] from the heightmap values. 2) Carve holes in the voxels as i want (easy too). 3) Convert voxels to polygons using some iso-surface extraction technique (like marching cubes). 4) Reduce (decimate) polygons created in 3). This technique seems to be the most promising for giving good results (untested). However the problem with marching cubes is that they tends to produce lots of polygons thus reducing them is mandatory. Implementing 4) also seems not trivial, i have read several papers on the web and it seems pretty complex. I was also unable to find an example, code snippet or something to start writing an algorithm for triangle mesh decimation. Maybe there is a special decimation algorithm (simpler) for meshes created from marching cubes ? Solution 2 1) Create some triangle mesh from the heighmap (easy). 2) Apply severals 3D boolean operation (eg: subtraction with a sphere) to carve the mesh. 3) apply some procedure to reduce polygons (optional). Operation 2) seems to be very complex and to be honest i have no idea how to do that. Also applying many boolean operation seems to be slow and will maybe degrade the triangle mesh every time a boolean operation is applied.

    Read the article

  • Single Responsibility Principle - How Can I Avoid Code Fragmentation?

    - by Dean Chalk
    I'm working on a team where the team leader is a virulent advocate of SOLID development principles. However, he lacks a lot of experience in getting complex software out of the door. We have a situation where he has applied SRP to what was already quite a complex code base, which has now become very highly fragmented and difficult to understand and debug. We now have a problem not only with code fragmentation, but also encapsulation, as methods within a class that may have been private or protected have been judged to represent a 'reason to change' and have been extracted to public or internal classes and interfaces which is not in keeping with the encapsulation goals of the application. We have some class constructors which take over 20 interface parameters, so our IoC registration and resolution is becoming a monster in its own right. I want to know if there is any 'refactor away from SRP' approach we could use to help fix some of these issues. I have read that it doesn't violate SOLID if I create a number of empty courser-grained classes that 'wrap' a number of closely related classes to provide a single-point of access to the sum of their functionality (i.e. mimicking a less overly SRP'd class implementation). Apart from that, I cannot think of a solution which will allow us to pragmatically continue with our development efforts, while keeping everyone happy. Any suggestions ?

    Read the article

  • Generating Deep Arrays: Shallow to Deep, Deep to Shallow or Bad idea?

    - by MobyD
    I'm working on an array structure that will be used as the data source for a report template in a web app. The data comes from relatively complex SQL queries that return one or many rows as one dimensional associative arrays. In the case of many, they are turned into two dimensional indexed array. The data is complex and in some cases there is a lot of it. To save trips to the database (which are extremely expensive in this scenario) I'm attempting to get all of the basic arrays (1 and 2 dimension raw database data) and put them, conditionally, into a single, five level deep array. Organizing the data in PHP seems like a better idea than by using where statements in the SQL. Array Structure Array of years( year => array of types( types => array of information( total => value, table => array of data( index => db array ) ) ) ) My first question is, is this a bad idea. Are arrays like this appropriate for this situation? If this would work, how should I go about populating it? My initial thought was shallow to deep, but the more I work on this, the more I realize that it'd be very difficult to abstract out the conditionals that determine where each item goes in the array. So it seems that starting from the most deeply nested data may be the approach I should take. If this is array abuse, what alternatives exist?

    Read the article

  • ASP.NET Session Management

    - by geekrutherford
    Great article (a little old but still relevant) about the inner workings of session management in ASP.NET: Underpinnings of the Session State Management Implementation in ASP.NET.   Using StateServer and the BinaryFormatter serialization occuring caused me quite the headache over the last few days. Curiously, it appears the w3wp.exe process actually consumes more memory when utilizing StateServer and storing somewhat large and complex data types in session.   Users began experiencing Out Of Memory exceptions in the production environment. Looking at the stack trace it related to serialization using the BinaryFormatter. Using remote debugging against our QA server I noted that the code in the application functioned without issue. The exception occured outside the context of the application itself when the request had completed and the web server was trying to serialize session state into the StateServer.   The short term solution is switching back to the InProc method. Thus far this has proven to consume considerably less memory and has caused no issues. Long term the complex object stored in session will be off-loaded into a web service used to access the information directly from the database outside the context of the object used to encapsulate it.

    Read the article

  • With Choice Comes Complexity

    - by BuckWoody
    "Complex" may be defined as "Having many steps, details or parts." Many of Microsoft's products, including SQL Server, can be complex. I'm stating what most data professionals already know - there's usually multiple ways to do things in SQL Server. For instance, to import some data into a table you can use graphical tools, SQLCMD, bcp, SQL Server Integration Services, BULK INSERT, even PowerShell, just to name a few tools at your disposal. That's really not the issue, though. The bigger issue is that there are normally multiple thought-processes, or methods, that you have available for a task. That's both a strength and a weakness. If things were more simple, you would have fewer choices. Sometimes that's a good thing. Just tell me what I need to do and I'll do it. However, your particular situation may not fit that tool or process, so having more options increases your ability to get your job done the way you need to do it. On the other hand, that's more for you to learn, which is harder. There's another side of this benefit/difficulty that you need to be aware of. Even if you're quite good at what you do, keep in mind that the way you know how to do something may not be the only way to do it. Keep your mind open to new possibilities, and most importantly - to new knowledge. SQL Server professionals teach me something new every day. So embrace the complexity - on balance, it's a good thing! Share this post: email it! | bookmark it! | digg it! | reddit! | kick it! | live it!

    Read the article

  • Evolution of mainstream programming languages: simplicity versus complexity.

    - by Giorgio
    I had posted this question on http://stackoverflow.com but I was suggested that it may be more appropriate to post it on this forum. I did a quick search on this site and it seems to me that this question has not been asked yet. Please give me a hint if the topic has been raised already by someone else. Update I have rephrased this question, removed personal opinions and made it shorter. I hope in this way it is better suited for this forum. By looking at the recent development of Java (Java 7) and C++ (C++0x) I see that new features are added to these languages. For sure this makes it easier to use certain programming idioms, adding to the productivity of developers. On the other hand, there might be the following risks A language becomes too big, complex, and difficult to understand. It lacks coherence in the design, e.g. if it mixes different paradigms like object-orientation and functional programming, which might not fit well together. Questions: what is more important to you as a developer: to have a rich language that captures a large collection of programming idioms or to have a small language that aims at coherence and simplicity (of course, with a good deal of libraries and tools accompanying it)? Or is it possible to have both? With respect to these issues: How do you judge the current evolutions of main-stream programming languages like Java or C++? Are they becoming too complex, less intuitive? Do they have enough features? Do they need more? Are they still easy enough to understand and use?

    Read the article

  • Why doesn't my IDE do background compiling/building?

    - by MKO
    Today I develop on a fairly complex computer, it has multiple cores, SSD drives and what not. Still, most of the time I'm programming the computer is leasurely doing nothing. When I need to compile and run/deploy a somewhat complex project at best it still takes a couple of seconds. Why? Now that we're living more and more in the "age of instant" why can't I press F5 in Visual studio and launch/deploy my application instantly? A couple of seconds might not sound so bad but it's still cognitive friction and time that adds up, and frankly it makes programming less fun. So how could compilation be instant? Well, People tend to edit files in different assemblies, what if Visual Studio/The IDE constantly did compilation/and building of everything that I modified anytime that it might be appropriate. Heck if they wanted to go really advanced they could do per-class compilation. The compilation might not work but then it could just silently do nothing (except adding error messages to the error window). Surely todays computer could dedicate a core or two to this task, and if someone found it annoying it could be disabled by option. I know there's probably a thousand technical issues and some fancy shadow copying that would need to be resolved for this to be seamless and practical but it sure would make programming more seamless. Is there any practical reason why this scenario isn't possible? Would the wear and tear of continually writing binaries be too much? Couldn't assemblies be held in memory until deployed/run?

    Read the article

  • How to go from Mainframe to the Cloud?

    - by Ruma Sanyal
    Running applications on IBM mainframes is expensive, complex, and hinders IT responsiveness. The high costs from frequent forced upgrades, long integration cycles, and complex operations infrastructures can only be alleviated by migrating away from a mainframe environment.  Further, data centers are planning for cloud enablement pinned on principles of operating at significantly lower cost, very low upfront investment, operating on commodity hardware and open, standards based systems, and decoupling of hardware, infrastructure software, and business applications. These operating principles are in direct contrast with the principles of operating businesses on mainframes. By utilizing technologies such as Oracle Tuxedo, Oracle Coherence, and Oracle GoldenGate, businesses are able to quickly and safely migrate away from their IBM mainframe environments. Further, running Oracle Tuxedo and Oracle Coherence on Oracle Exalogic, the first and only integrated cloud machine on the market, Oracle customers can not only run their applications on standards-based open systems, significantly cutting their time to market and costs, they can start their journey of cloud enabling their mainframe applications. Oracle Tuxedo re-hosting tools and techniques can provide automated migration coverage for more than 95% of mainframe application assets, at a fraction of the cost Oracle GoldenGate can migrate data from mainframe systems to open systems, eliminating risks associated with the data migration Oracle Coherence hosts transactional data in memory providing mainframe-like data performance and linear scalability Running Oracle software on top of Oracle Exalogic empowers customers to start their journey of cloud enabling their mainframe applications Join us in a series of events across the globe where you you'll learn how you can build your enterprise cloud and add tremendous value to your business. In addition, meet with Oracle experts and your peers to discuss best practices and see how successful organizations are lowering total cost of ownership and achieving rapid returns by moving to the cloud. Register for the Oracle Fusion Middleware Forum event in a city new you!

    Read the article

  • What are the most commonly used and basic Apache htaccess redirects?

    - by bybe
    This question is here so we can offer users who are looking for information on how to make one or more common or basic redirects in Apache using the htaccess file. All future questions pertaining to finding information that is that is covered by the question should be closed as a duplication of this question. As per this Meta question. Whats the point in this question? The idea while not perfect is catch the most commonly asked questions regarding redirects using the htaccess on the Apache platform either on some type of lamp or a live server. The type of answers should be generally those that you could imagine are used by 100,000’s of sites world-wide and are constantly asked here at Pro Webmasters repeatedly over and over in various forms. A few examples of the type of answers we are looking for: How can I redirect non-www to www? How can I redirect a sub domain to the main domain? How can I redirect a sub folder from domain to a root or a subdomain? How can I redirect an old URL to a new URL? A few examples of the types of answers that we are not looking for: Answers that do not involve a redirect. Any answers relating to NGinx, IIS or any other non-Apache platform. Answers that involve custom and complex string or query removals. Resources for Advanced to Complex Mod_Rewrite Rules: Everything you ever wanted to know about mod rewrite rules but were afraid to ask Please note: that this question is still in construction and may need some refining either by myself or a real moderator of Pro Webmasters, if you have any concerns or questions please use the meta page I made a few days back here.

    Read the article

  • Software Licencing [closed]

    - by Craig
    A colleague of mine wanted a means to do something, so it was suggested that I write some software to do this. The software has turned into more than the original specification and is now something rather complex, however it is not fully functional still. My colleague has not paid me anything so far and I am unwilling to continue writing the software until some faith has been reciprocated in my direction, as I have put a lot of effort into writing the software. I am also unwilling to finish the software as I do not want to give away a huge chunk of my time and effort away as free, neither do I want to be under compensated for my efforts. Some concerns I have. If I finish the software, what if the client doesn't pay me anything or pays too little, or what if I write the software to a usable level, but not complete and the client pays me a too little. I have lost my motivation to finish the software, as more and more specifications have been added to the software and I have developed a substantially complex piece of software and been effectively paid nothing. To finish the software, I need motivation, money would do this, however the client doesn't want to pay for something that isn't complete, yet keeps adding more requirements. I seem to be in a catch 22 with this, as I have developed some software on faith and have had no faith reciprocated in my direction. I'm really not sure how to get some payment from the client or on how to develop a licencing model so that I get some money from the client and development resumes.

    Read the article

  • TDD with SQL and data manipulation functions

    - by Xophmeister
    While I'm a professional programmer, I've never been formally trained in software engineering. As I'm frequently visiting here and SO, I've noticed a trend for writing unit tests whenever possible and, as my software gets more complex and sophisticated, I see automated testing as a good idea in aiding debugging. However, most of my work involves writing complex SQL and then processing the output in some way. How would you write a test to ensure your SQL was returning the correct data, for example? Then, say if the data wasn't under your control (e.g., that of a 3rd party system), how can you efficiently test your processing routines without having to hand write reams of dummy data? The best solution I can think of is making views of the data that, together, cover most cases. I can then join those views with my SQL to see if it's returning the correct records and manually process the views to see if my functions, etc. are doing what they're supposed to. Still, it seems excessive and flakey; particularly finding data to test against...

    Read the article

  • What is the value to checking in broken unit tests?

    - by Adam W.
    While there are ways of keeping unit tests from being executed, what is the value of checking in broken unit tests? I will use a simple example. Case sensitivity. The current code is Case Sensitive. A valid input into the method is "Cat" and it would return an enum of Animal.Cat. However, the desired functionality of the method should not be case sensitive. So if the method described was passed "cat" it could possibly return something like Animal.Null instead of Animal.Cat and the unit test would fail. Though a simple code change would make this work, a more complex issue may take weeks to fix, but identifying the bug with a unit test could be a less complex task. The application currently being analyzed has 4 years of code that "works". However, recent discussions regarding unit tests has found flaws in the code. Some just need explicit implementation documentation (ex. case sensitive or not), or code that does not execute the bug based on how it is currently called. But unit tests can be created executing specific scenarios that will cause the bug to be seen and are valid inputs. What is the value of checking in unit tests that exercise the bug until someone can get around to fixing the code? Should this unit test be flagged with ignore, priority, category etc, to determine whether a build was successful based on tests executed? Eventually the unit test should be created to execute the code once someone fixes it. On one hand it shows that identified bugs have not been fixed. On the other, there could be hundreds of failed unit tests showing up in the logs and weeding through the ones that should fail vs. failures due to a code check-in would be difficult to find.

    Read the article

  • for an ajax heavy web application which would be better SOAP or REST?

    - by coder
    I'm building an ajax heavy application (client-side strictly html/css/js) which will be getting all the data and using server business logic via webservices. I know REST seems to be the hot topic but I can't find any good arguments. The main argument seems to be its "light-weight". My impression so far is that wsdl/soap based services are more expressive and allow for more a more complex transfer of data. It appears that soap would be more useful in the application I'm building where the only code consuming the services will be the js downloaded in the client browser. REST on the other hand seems to have a smaller entry barrier and so can be more useful for services like twitter in allowing other developers to consume these services easily. Also, REST seems to Te better suited for simple data transfers. So in summary SOAP is useful for complex data transfer and REST is useful in simple data transfer. I'm currently under the impression that using SOAP would be best due to the complexity of the messages but perhaps there's other factors. What are your thoughts on the pros/cons of soap/rest for a heavy ajax web app?

    Read the article

  • Are long methods always bad?

    - by wobbily_col
    So looking around earlier I noticed some comments about long methods being bad practice. I am not sure I always agree that long methods are bad (and would like opinions from others). For example I have some Django views that do a bit of processing of the objects before sending them to the view, a long method being 350 lines of code. I have my code written so that it deals with the paramaters - sorting / filtering the queryset, then bit by bit does some processing on the objects my query has returned. So the processing is mainly conditional aggregation, that has complex enough rules it can't easily be done in the database, so I have some variables declared outside the main loop then get altered during the loop. varaible_1 = 0 variable_2 = 0 for object in queryset : if object.condition_condition_a and variable_2 > 0 : variable 1+= 1 ..... ... . more conditions to alter the variables return queryset, and context So according to the theory I should factor out all the code into smaller methods, so That I have the view method as being maximum one page long. However having worked on various code bases in the past, I sometimes find it makes the code less readable, when you need to constantly jump from one method to the next figuring out all the parts of it, while keeping the outermost method in your head. I find that having a long method that is well formatted, you can see the logic more easily, as it isn't getting hidden away in inner methods. I could factor out the code into smaller methods, but often there is is an inner loop being used for two or three things, so it would result in more complex code, or methods that don't do one thing but two or three (alternatively I could repeat inner loops for each task, but then there will be a performance hit). So is there a case that long methods are not always bad? Is there always a case for writing methods, when they will only be used in one place?

    Read the article

  • Am I programming too slow?

    - by Jonn
    I've only been a year in the industry and I've had some problems making estimates for specific tasks. Before you close this, yes, I've already read this: http://programmers.stackexchange.com/questions/648/how-to-respond-when-you-are-asked-for-an-estimate and that's about the same problem I'm having. But I'm looking for a more specific gauge of experiences, something quantifiable or probably other programmer's average performances which I should aim for and base my estimates. The answers range from weeks, and I was looking more for an answer on the level of a task assigned for a day or so. (Note that this doesn't include submitting for QA or documentations, just the actual development time from writing tests if I used TDD, to making the page, before having it submitted to testing) My current rate right now is as follows (on ASP.NET webforms): Right now, I'm able to develop a simple data entry page with a grid listing (no complex logic, just Creating and Reading) on an already built architecture, given one full day's (8 hours) time. Adding complex functionality, and Update and Delete pages add another full day to the task. If I have to start the page from scratch (no solution, no existing website) it takes me another full day. (Not always) but if I encounter something new or haven't done yet it takes me another full day. Whenever I make an estimate that's longer than the expected I feel that others think that I'm lagging a lot behind everyone else. I'm just concerned as there have been expectations that when it's just one page it should take me no more than a full day. Yes, there definitely is more room for improvement. There always is. I have a lot to learn. But I would like to know if my current rate is way too slow, just average, or average for someone no longer than a year in the industry.

    Read the article

  • No sound after upgrading to Ubuntu 11.10 from win7

    - by Tilman
    just as a prefix to my question, i'd like to note that i'm just now entering the world of Linux (unless you count my android, but that's a very different experience...) i have two computers now that run Ubuntu 11.10, the first of which i've had very little problems with, aside from figuring out the basics. the second, from which i'm writing this question, has (up to this point) only had one problem.... no sound. i've read a couple questions similiar and found little help as the component catalog doesn't have my computer listed. (in fact i'm not suprised this is a pos i had my mom grab from her work before they officially closed the doors behind them) had perfect sound before hand, and no sound now. sudo lspci -v brings up 00:1b.0 Audio device: Intel Corporation N10/ICH 7 Family High Definition Audio Controller (rev 01) Subsystem: Intel Corporation Device d608 Flags: bus master, fast devsel, latency 0, IRQ 45 Memory at ff980000 (64-bit, non-prefetchable) [size=16K] Capabilities: [50] Power Management version 2 Capabilities: [60] MSI: Enable+ Count=1/1 Maskable- 64bit+ Capabilities: [70] Express Root Complex Integrated Endpoint, MSI 00 Capabilities: [100] Virtual Channel Capabilities: [130] Root Complex Link Kernel driver in use: HDA Intel Kernel modules: snd-hda-intel any help would be greatly appreciated, me and my gf just wanna watch a damn movie lol

    Read the article

  • What is the value of checking in failing unit tests?

    - by Adam W.
    While there are ways of keeping unit tests from being executed, what is the value of checking in failing unit tests? I will use a simple example: Case Sensitivity. The current code is case sensitive. A valid input into the method is "Cat" and it would return an enum of Animal.Cat. However, the desired functionality of the method should not be case sensitive. So if the method described was passed "cat" it could possibly return something like Animal.Null instead of Animal.Cat and the unit test would fail. Though a simple code change would make this work, a more complex issue may take weeks to fix, but identifying the bug with a unit test could be a less complex task. The application currently being analyzed has 4 years of code that "works". However, recent discussions regarding unit tests have found flaws in the code. Some just need explicit implementation documentation (ex. case sensitive or not), or code that does not execute the bug based on how it is currently called. But unit tests can be created executing specific scenarios that will cause the bug to be seen and are valid inputs. What is the value of checking in unit tests that exercise the bug until someone can get around to fixing the code? Should this unit test be flagged with ignore, priority, category etc, to determine whether a build was successful based on tests executed? Eventually the unit test should be created to execute the code once someone fixes it. On one hand it shows that identified bugs have not been fixed. On the other, there could be hundreds of failed unit tests showing up in the logs and weeding through the ones that should fail vs. failures due to a code check-in would be difficult to find.

    Read the article

  • Is it okay to showcase templates/layouts recreated in different codes in a portfolio?

    - by Souta
    I have several different templates/layouts, both simple and complex. I recreated these templates multiple times, just using different codes. (Say, a complex one was originally made in only HTML and CSS, I recreated it using HTML, Javascript, CSS, then again with a HTML and PHP concoction, and etc.) I wanted to showcase my work and skills by doing this, but I don't know if it would be okay for that all to go into a resumé/portfolio. This is why: Freelancing Does potential business really care about how their site is made, as long as it looks and functions to their liking? (As in, should I just only show the one example of each template/layout and not the multiple recreations?) Potential Hire However, if a potential employer were to stumble across my resumé/portfolio, would having the multiple recreations do any good for a career outlook? (As in, this potential employer is a company where I could be working on a team to create/develop sites and not be freelancing; would a lack of skill-shining turn this employer away because I didn't set myself apart and show that I'm not just like every other budding web designer?) Those two issues have me wondering if it is okay to have a resumé/portfolio combined for this specific reason. Or does something like this not matter to potential business (as a freelancer) because they wouldn't care either way as long as it looks and functions to their liking and therefore it is okay to showcase the recreations with the originals?

    Read the article

  • How to optimize calls to multiple APIs at once and return as one set?

    - by Martin
    I have a web app that searches across 2 APIs right now. I have my own Restful web service that I call, and it does all the work on the backend to asynchronously call the 2 APIs and concatenate them into one result set for my web app to use. I want to scale this out and add as many other APIs as I can (currently looking at about 10 more). But as I add APIs, the call to my service gets (potentially) slower and more complex. How do I handle one API not responding ... and other issues that arise? What would be the best way to approach this? Should I create a service call for each API, that way each one is independent and not coupled to all the other calls? Is there a way on the backend to handle the multiple API calls without all the extra complexity it adds? If I go the route of a service call per API, now my client code gets more complex (and I have a lot of clients)? And it's more work for the client, and since I have mobile apps, it will cost the client more data usage. If I go one service call, is there a way to set up some sort of connection so I can return data as I get it, in case one service call hangs?

    Read the article

  • Upgrading to 12.10 on an external hard drive

    - by Tom Childers
    I did some googling on this and didn't find anything specific for my situation. I currently have 12.04 installed on an external USB hard drive. It's working great. I want to upgrade it to 12.10. My bandwidth is very limited so I have a friend who will download 12.10 for me and put it on a flash stik. Then I can upgrade without having to do the download myself. Which particular version of the 12.10 download file(s) should I get? Are there alternate 12.10 downloads that have all the packages? How do I set it up so when I upgrade 12.04 I can specify that it look in some local repository for the 12.10 files? Can I just dump the 12.10 files in some local directory? Or do I have do go thru some complex commands to create a local repository? I'm pretty new to Linux so a long process of complex terminal commands will probably be a show stopper for me. Remember that my 12.04 install resides on an external hard drive. And I have a laptop with multiple USB ports. Thanks! Advait

    Read the article

< Previous Page | 46 47 48 49 50 51 52 53 54 55 56 57  | Next Page >