Search Results

Search found 1352 results on 55 pages for 'perspective'.

Page 49/55 | < Previous Page | 45 46 47 48 49 50 51 52 53 54 55  | Next Page >

  • JNI - GetObjectField returns NULL

    - by Daniel
    I'm currently working on Mangler's Android implementation. I have a java class that looks like so: public class VentriloEventData { public short type; public class _pcm { public int length; public short send_type; public int rate; public byte channels; }; _pcm pcm; } The signature for my pcm object: $ javap -s -p VentriloEventData ... org.mangler.VentriloEventData$_pcm pcm; Signature: Lorg/mangler/VentriloEventData$_pcm; I am implementing a native JNI function called getevent, which will write to the fields in an instance of the VentriloEventData class. For what it's worth, it's defined and called in Java like so: public static native int getevent(VentriloEventData data); VentriloEventData data = new VentriloEventData(); getevent(data); And my JNI implementation of getevent: JNIEXPORT jint JNICALL Java_org_mangler_VentriloInterface_getevent(JNIEnv* env, jobject obj, jobject eventdata) { v3_event *ev = v3_get_event(V3_BLOCK); if(ev != NULL) { jclass event_class = (*env)->GetObjectClass(env, eventdata); // Event type. jfieldID type_field = (*env)->GetFieldID(env, event_class, "type", "S"); (*env)->SetShortField( env, eventdata, type_field, 1234 ); // Get PCM class. jfieldID pcm_field = (*env)->GetFieldID(env, event_class, "pcm", "Lorg/mangler/VentriloEventData$_pcm;"); jobject pcm = (*env)->GetObjectField( env, eventdata, pcm_field ); jclass pcm_class = (*env)->GetObjectClass(env, pcm); // Set PCM fields. jfieldID pcm_length_field = (*env)->GetFieldID(env, pcm_class, "length", "I"); (*env)->SetIntField( env, pcm, pcm_length_field, 1337 ); free(ev); } return 0; } The code above works fine for writing into the type field (that is not wrapped by the _pcm class). Once getevent is called, data.type is verified to be 1234 at the java side :) My problem is that the assertion "pcm != NULL" will fail. Note that pcm_field != NULL, which probably indicates that the signature to that field is correct... so there must be something wrong with my call to GetObjectField. It looks fine though if I compare it to the official JNI docs. Been bashing my head on this problem for the past 2 hours and I'm getting a little desperate.. hoping a different perspective will help me out on this one.

    Read the article

  • Constructor or Explicit cast

    - by Felan
    In working with Linq to Sql I create a seperate class to ferry data to a web page. To simplify creating these ferry objects I either use a specialized constructor or an explicit conversion operator. I have two questions. First which approach is better from a readibility perspective? Second while the clr code that is generated appeared to be the same to me, are there situations where one would be treated different than the other by the compiler (in lambda's or such). Example code (DatabaseFoo uses specialized constructor and BusinessFoo uses explicit operator): public class DatabaseFoo { private static int idCounter; // just to help with generating data public int Id { get; set; } public string Name { get; set; } public DatabaseFoo() { Id = idCounter++; Name = string.Format("Test{0}", Id); } public DatabaseFoo(BusinessFoo foo) { this.Id = foo.Id; this.Name = foo.Name; } } public class BusinessFoo { public int Id { get; set; } public string Name { get; set; } public static explicit operator BusinessFoo(DatabaseFoo foo) { return FromDatabaseFoo(foo); } public static BusinessFoo FromDatabaseFoo(DatabaseFoo foo) { return new BusinessFoo {Id = foo.Id, Name = foo.Name}; } } public class Program { static void Main(string[] args) { Console.WriteLine("Creating the initial list of DatabaseFoo"); IEnumerable<DatabaseFoo> dafoos = new List<DatabaseFoo>() { new DatabaseFoo(), new DatabaseFoo(), new DatabaseFoo(), new DatabaseFoo(), new DatabaseFoo(), new DatabaseFoo()}; foreach(DatabaseFoo dafoo in dafoos) Console.WriteLine(string.Format("{0}\t{1}", dafoo.Id, dafoo.Name)); Console.WriteLine("Casting the list of DatabaseFoo to a list of BusinessFoo"); IEnumerable<BusinessFoo> bufoos = from x in dafoos select (BusinessFoo) x; foreach (BusinessFoo bufoo in bufoos) Console.WriteLine(string.Format("{0}\t{1}", bufoo.Id, bufoo.Name)); Console.WriteLine("Creating a new list of DatabaseFoo by calling the constructor taking BusinessFoo"); IEnumerable<DatabaseFoo> fufoos = from x in bufoos select new DatabaseFoo(x); foreach(DatabaseFoo fufoo in fufoos) Console.WriteLine(string.Format("{0}\t{1}", fufoo.Id, fufoo.Name)); } }

    Read the article

  • Spring.NET & Immediacy CMS (or how to inject to server side controls without using PageHandlerFactor

    - by Simon Rice
    Is there any way to inject dependencies into an Immediacy CMS control using Spring.NET, ideally without having to use to ContextRegistry when initialising the control? Update, with my own answer The issue here is that Immediacy already has a handler defined in web.config that deals with all aspx pages, & so it's not possible add an entry for Spring.NET's PageHandlerFactory in web.config as per a normal webforms app. That rules out making the control implement ISupportsWebDependencyInjection. Furthermore, most of Immediacy's generated pages are aspx pages that don't physically exist on the drive. I have changed the title of the question to reflect this. What I have done to get Dependency Injection working is: Add the usual entries to web.config for Spring.NET as outlined in the documentation, except for the adding the entry to the <httpHandlers> section. In this case I've got my object definitions in Spring.config. Create the following abstract base class that will deal with all of the Dependency Injection work: DIControl.cs public abstract class DIControl : ImmediacyControl { protected virtual string DIName { get { return this.GetType().Name; } } protected override void OnInit(EventArgs e) { if (ContextRegistry.GetContext().GetObject(DIName, this.GetType()) != null) ContextRegistry.GetContext().ConfigureObject(this, DIName); base.OnInit(e); } } For non-immediacy controls, you can make this server side control inherit from Control or whatever subclass of that you like. For any control with which you wish to use with Spring.NET's Inversion of Control container, define it to inherit from DIControl & add the relelvant entry to Spring.config, for example: SampleControl.cs public class SampleControl : DIControl, INamingContainer { public string Text { get; set; } protected string InjectedText { get; set; } public SampleControl() : base() { Text = "Hello world"; } protected override void RenderContents(HtmlTextWriter output) { output.Write(string.Format("{0} {1}", Text, InjectedText)); } } Spring.config <objects xmlns="http://www.springframework.net"> <object id="SampleControl" type="MyProject.SampleControl, MyAssembly"> <property name="InjectedText" value="from Spring.NET" /> </object> </objects> You can optionally override DIName if you wish to name your entry in Spring.config differently from the name of your class. Provided everything's done correctly, you will have the control writing out "Hello world from Spring.NET!" when used in a page. This solution uses Spring.NET's ContextRegistry from within the control, but I would be surprised if there's no way around that for Immediacy at least since the page objects themselves aren't accessible. However, can this be improved at all from a Spring.NET perspective? Is there maybe an Immediacy plugin that already does this that I'm completely unaware of? Or is there an approach that does this in a more elegant way? I'm open to suggestions.

    Read the article

  • Recommended textbook for machine-level programming?

    - by Norman Ramsey
    I'm looking at textbooks for an undergraduate course in machine-level programming. If the perfect book existed, this is what it would look like: Uses examples written in C or assembly language, or both. Covers machine-level operations such as two's-complement integer arithmetic, bitwise operations, and floating-point arithmetic. Explains how caches work and how they affect performance. Explains machine instructions or assembly instructions. Bonus if the example assembly language includes x86; triple bonus if it includes x86-64 (aka AMD64). Explains how C values and data structures are represented using hardware registers and memory. Explains how C control structures are translated into assembly language using conditional and unconditional branch instructions. Explains something about procedure calling conventions and how procedure calls are implemented at the machine level. Books I might be interested in would probably have the words "machine organization" or "computer architecture" in the title. Here are some books I'm considering but am not quite happy with: Computer Systems: A Programmer's Perspective by Randy Bryant and Dave O'Hallaron. This is quite a nice book, but it's a book for a broad, shallow course in systems programming, and it contains a great deal of material my students don't need. Also, it is just out in a second edition, which will make it expensive. Computer Organization and Design: The Hardware/Software Interface by Dave Patterson and John Hennessy. This is also a very nice book, but it contains way more information about how the hardware works than my students need. Also, the exercises look boring. Finally, it has a show-stopping bug: it is based very heavily on MIPS hardware and the use of a MIPS simulator. My students need to learn how to use DDD, and I can't see getting this to work on a simulator. Not to mention that I can't see them cross-compiling their code for the simulator, and so on and so forth. Another flaw is that the book mentions the x86 architecture only to sneer at it. I am entirely sympathetic to this point of view, but news flash! You guys lost! Write Great Code Vol I: Understanding the Machine by Randall Hyde. I haven't evaluated this book as thoroughly as the other two. It has a lot of what I need, but the translation from high-level language to assembler is deferred to Volume Two, which has mixed reviews. My students will be annoyed if I make them buy a two-volume series, even if the price of those two volumes is smaller than the price of other books. I would really welcome other suggestions of books that would help students in a class where they are to learn how C-language data structures and code are translated to machine-level data structures and code and where they learn how to think about performance, with an emphasis on the cache.

    Read the article

  • OpenGL Projection matrix won't allow displaying anything

    - by user272973
    I'm trying to get some basic OpenGL-ES with Shaders to run on the iPhone, based on some examples. For some reason my projection matrix refuses to result in something on the screen. It feels like a clipping plane is set very near but that contradicts with the values I supply. If I render the same scene with an Orthogonal projection matrix I see my object just no perspective obviously. Here's the code that generates the projection matrix: esPerspective(&proj, 45.f, 768.0/1024.0, 1.f, 10000.f); void esPerspective(ESMatrix *result, float fovy, float aspect, float nearZ, float farZ) { float frustumW, frustumH; frustumH = tanf( fovy / 360.0f * PI ) * nearZ; frustumW = frustumH * aspect; esFrustum( result, -frustumW, frustumW, -frustumH, frustumH, nearZ, farZ ); } void esFrustum(ESMatrix *result, float left, float right, float bottom, float top, float nearZ, float farZ) { float deltaX = right - left; float deltaY = top - bottom; float deltaZ = farZ - nearZ; ESMatrix frust; if ( (nearZ <= 0.0f) || (farZ <= 0.0f) || (deltaX <= 0.0f) || (deltaY <= 0.0f) || (deltaZ <= 0.0f) ) return; frust.m[0][0] = 2.0f * nearZ / deltaX; frust.m[0][1] = frust.m[0][2] = frust.m[0][3] = 0.0f; frust.m[1][1] = 2.0f * nearZ / deltaY; frust.m[1][0] = frust.m[1][2] = frust.m[1][3] = 0.0f; frust.m[2][0] = (right + left) / deltaX; frust.m[2][1] = (top + bottom) / deltaY; frust.m[2][2] = -(nearZ + farZ) / deltaZ; frust.m[2][3] = -1.0f; frust.m[3][2] = -2.0f * nearZ * farZ / deltaZ; frust.m[3][0] = frust.m[3][1] = frust.m[3][3] = 0.0f; esMatrixMultiply(result, &frust, result); } My projection matrix comes out as: [3.21, 0, 0, 0] [0, 2.41, 0, 0] [0, 0, -1, -1] [0, 0, -2, 0] Even if I manually set the [3][3] cell to 1 I still don't see anything. Any ideas?

    Read the article

  • In a PHP project, how do you organize and access your helper objects?

    - by Pekka
    How do you organize and manage your helper objects like the database engine, user notification, error handling and so on in a PHP based, object oriented project? Say I have a large PHP CMS. The CMS is organized in various classes. A few examples: the database object user management an API to create/modify/delete items a messaging object to display messages to the end user a context handler that takes you to the right page a navigation bar class that shows buttons a logging object possibly, custom error handling etc. I am dealing with the eternal question, how to best make these objects accessible to each part of the system that needs it. my first apporach, many years ago was to have a $application global that contained initialized instances of these classes. global $application; $application->messageHandler->addMessage("Item successfully inserted"); I then changed over to the Singleton pattern and a factory function: $mh =&factory("messageHandler"); $mh->addMessage("Item successfully inserted"); but I'm not happy with that either. Unit tests and encapsulation become more and more important to me, and in my understanding the logic behind globals/singletons destroys the basic idea of OOP. Then there is of course the possibility of giving each object a number of pointers to the helper objects it needs, probably the very cleanest, resource-saving and testing-friendly way but I have doubts about the maintainability of this in the long run. Most PHP frameworks I have looked into use either the singleton pattern, or functions that access the initialized objects. Both fine approaches, but as I said I'm happy with neither. I would like to broaden my horizon on what is possible here and what others have done. I am looking for examples, additional ideas and pointers towards resources that discuss this from a long-term, real-world perspective. Also, I'm interested to hear about specialized, niche or plain weird approaches to the issue. Bounty I am following the popular vote in awarding the bounty, the answer which is probably also going to give me the most. Thank you for all your answers!

    Read the article

  • Some clarification on rvalue references

    - by Dennis Zickefoose
    First: where are std::move and std::forward defined? I know what they do, but I can't find proof that any standard header is required to include them. In gcc44 sometimes std::move is available, and sometimes its not, so a definitive include directive would be useful. When implementing move semantics, the source is presumably left in an undefined state. Should this state necessarily be a valid state for the object? Obviously, you need to be able to call the object's destructor, and be able to assign to it by whatever means the class exposes. But should other operations be valid? I suppose what I'm asking is, if your class guarantees certain invariants, should you strive to enforce those invariants when the user has said they don't care about them anymore? Next: when you don't care about move semantics, are there any limitations that would cause a non-const reference to be preferred over an rvalue reference when dealing with function parameters? void function(T&); over void function(T&&); From a caller's perspective, being able to pass functions temporary values is occasionally useful, so it seems as though one should grant that option whenever it is feasible to do so. And rvalue references are themselves lvalues, so you can't inadvertently call a move-constructor instead of a copy-constructor, or something like that. I don't see a downside, but I'm sure there is one. Which brings me to my final question. You still can not bind temporaries to non-const references. But you can bind them to non-const rvalue references. And you can then pass along that reference as a non-const reference in another function. void function1(int& r) { r++; } void function2(int&& r) { function1(r); } int main() { function1(5); //bad function2(5); //good } Besides the fact that it doesn't do anything, is there anything wrong with that code? My gut says of course not, since changing rvalue references is kind of the whole point to their existence. And if the passed value is legitimately const, the compiler will catch it and yell at you. But by all appearances, this is a runaround of a mechanism that was presumably put in place for a reason, so I'd just like confirmation that I'm not doing anything foolish.

    Read the article

  • Grid sorting with persistent master sort

    - by MikeWyatt
    I have a UI with a grid. Each record in the grid is sorted by a "master" sort column, let's call it a page number. Each record is a story in a magazine. I want the user to be able to drag and drop a record to a new position in the grid and automatically update the page number field to reflect the updated position. Easy enough, right? Now imagine that I also want to have the grid sortable by any other column (story title, section, author name, etc.). How does the drag and drop operation work now? Revert to page number sort during or after the drag and drop operation? This could confuse the user (why did my sort just change?). It would also result in arbitrary row positioning. Would the story now be before the row that was after it when the user dropped it? Or, would it be after the row that was before it? Those rows may now be widely separated after the master order sort. Disable the drag and drop feature if the grid isn't currently sorted by the page number? This would be easy, but the user might wonder why he can't drag and drop at certain times. Knowing to first sort by page number may not be very intuitive. Let the user rearrange his rows, but not make any changes to the page number? Require the user to enter a "Arrange Stories" mode, in which the grid sort is temporarily switched to page number and drag and drop is enabled? They would then exit the mode, and the previous sort would be reapplied. The big difference between this and the second option is that it would be more explicit than simply clicking on a column header. Any other ideas, or reasons why one of the above is the way to go? EDIT I'd like to point out that any of the above is technically possible, and easy to implement. My question is design-related. What is the most intuitive way to solve this problem, from the user's perspective?

    Read the article

  • gcc optimization? bug? and its practial implication to project

    - by kumar_m_kiran
    Hi All, My questions are divided into three parts Question 1 Consider the below code, #include <iostream> using namespace std; int main( int argc, char *argv[]) { const int v = 50; int i = 0X7FFFFFFF; cout<<(i + v)<<endl; if ( i + v < i ) { cout<<"Number is negative"<<endl; } else { cout<<"Number is positive"<<endl; } return 0; } No specific compiler optimisation options are used or the O's flag is used. It is basic compilation command g++ -o test main.cpp is used to form the executable. The seemingly very simple code, has odd behaviour in SUSE 64 bit OS, gcc version 4.1.2. The expected output is "Number is negative", instead only in SUSE 64 bit OS, the output would be "Number is positive". After some amount of analysis and doing a 'disass' of the code, I find that the compiler optimises in the below format - Since i is same on both sides of comparison, it cannot be changed in the same expression, remove 'i' from the equation. Now, the comparison leads to if ( v < 0 ), where v is a constant positive, So during compilation itself, the else part cout function address is added to the register. No cmp/jmp instructions can be found. I see that the behaviour is only in gcc 4.1.2 SUSE 10. When tried in AIX 5.1/5.3 and HP IA64, the result is as expected. Is the above optimisation valid? Or, is using the overflow mechanism for int not a valid use case? Question 2 Now when I change the conditional statement from if (i + v < i) to if ( (i + v) < i ) even then, the behaviour is same, this atleast I would personally disagree, since additional braces are provided, I expect the compiler to create a temporary built-in type variable and them compare, thus nullify the optimisation. Question 3 Suppose I have a huge code base, an I migrate my compiler version, such bug/optimisation can cause havoc in my system behaviour. Ofcourse from business perspective, it is very ineffective to test all lines of code again just because of compiler upgradation. I think for all practical purpose, these kinds of error are very difficult to catch (during upgradation) and invariably will be leaked to production site. Can anyone suggest any possible way to ensure to ensure that these kind of bug/optimization does not have any impact on my existing system/code base? PS : When the const for v is removed from the code, then optimization is not done by the compiler. I believe, it is perfectly fine to use overflow mechanism to find if the variable is from MAX - 50 value (in my case).

    Read the article

  • Primary language - C++/Qt, C#, Java?

    - by Airjoe
    I'm looking for some input, but let me start with a bit of background (for tl;dr skip to end). I'm an IT major with a concentration in networking. While I'm not a CS major nor do I want to program as a vocation, I do consider myself a programmer and do pretty well with the concepts involved. I've been programming since about 6th grade, started out with a proprietary game creation language that made my transition into C++ at college pretty easy. I like to make programs for myself and friends, and have been paid to program for local businesses. A bit about that- I wrote some programs for a couple local businesses in my senior year in high school. I wrote management systems for local shops (inventory, phone/pos orders, timeclock, customer info, and more stuff I can't remember). It definitely turned out to be over my head, as I had never had any formal programming education. It was a great learning experience, but damn was it crappy code. Oh yeah, by the way, it was all vb6. So, I've used vb6 pretty extensively, I've used c++ in my classes (intro to programming up to algorithms), used Java a little bit in another class (had to write a ping client program, pretty easy) and used Java for some simple Project Euler problems to help learn syntax and such when writing the program for the class. I've also used C# a bit for my own simple personal projects (simple programs, one which would just generate an HTTP request on a list of websites and notify if one responded unexpectedly or not at all, and another which just held a list of things to do and periodically reminded me to do them), things I would've written in vb6 a year or two ago. I've just started using Qt C++ for some undergrad research I'm working on. Now I've had some formal education, I [think I] understand organization in programming a lot better (I didn't even use classes in my vb6 programs where I really should have), how it's important to structure code, split into functions where appropriate, document properly, efficiency both in memory and speed, dynamic and modular programming etc. I was looking for some input on which language to pick up as my "primary". As I'm not a "real programmer", it will be mostly hobby projects, but will include some 'real' projects I'm sure. From my perspective: QtC++ and Java are cross platform, which is cool. Java and C# run in a virtual machine, but I'm not sure if that's a big deal (something extra to distribute, possibly a bit slower? I think Qt would require additional distributables too, right?). I don't really know too much more than this, so I appreciate any help, thanks! TL;DR Am an avocational programmer looking for a language, want quick and straight forward development, liked vb6, will be working with database driven GUI apps- should I go with QtC++, Java, C#, or perhaps something else?

    Read the article

  • What makes great software?

    - by VirtuosiMedia
    From the perspective of an end user, what makes a software great rather than just good or functional? What are some fundamental principles that can shift the way a software is used and perceived? What are some of the little finishing touches that help put an application over the top? I'm in the later stages of developing a web app and I'm looking for ideas or concepts that I may have missed. If you have specific examples of software or apps that you absolutely love, please share the reasons or features that make it special. Keep in mind that I'm looking for examples that directly affect the end user, but not necessarily just UI suggestions. Here are some of the principles and little touches I'm trying to use: Keep the UI as simple as possible. Remove absolutely everything that isn't necessary. Use progressive disclosure when more information can be needed sometimes but isn't needed all the time. Provide inline help and useful error messages. Verbs on buttons wherever possible. Make anything that's clickable obvious. Fast, responsive UI. Accessibility (this is a work in progress). Reusable UI patterns. Once a user learns a skill, they will be able to use it in multiple places. Intelligent default settings. Auto-focusing forms when filling out the form is the primary action to be taken on the page. Clear metaphors (like tabs) and headings indicating location within the app. Automating repetitive tasks (with the ability to disable the automation). Use standardized or accepted metaphors for icons (like an "x" for delete). Larger text sizes for improved readability. High contrast so that each section is distinct. Making sure that it's obvious on every page what the user is supposed to do by establishing a clear information hierarchy and drawing the eye to the call to action. Most deletions can be undone. Discoverability - Make it easy to learn how to do new tasks. Group similar elements together.

    Read the article

  • Image animation problem in silverlight

    - by Jak
    Hi followed " http://www.switchonthecode.com/tutorials/silverlight-3-tutorial-planeprojection-and-perspective-3d#comment-4688 ".. the animation is working fine. I am new to silver light. when i use dynamic image from xml instead of static image as in tutorial,.. it is not working fine, please help me on this. i used list box.. for this animation effect do i need to change listbox to some other arrangement ? if your answer yes means, pls give me some sample code. Thanks in advance. Xaml code: <ListBox Name="listBox1"> <ListBox.ItemTemplate> <DataTemplate> <StackPanel> <Image Source="{Binding imgurl}" HorizontalAlignment="Left" Name="image1" Stretch="Fill" VerticalAlignment="Top" MouseLeftButtonUp="FlipImage" /> </StackPanel> </DataTemplate> </ListBox.ItemTemplate> </ListBox> My C# code: //getting image URL from xml XElement xmlads = XElement.Parse(e.Result); //i bind the url in to listBox listBox1.ItemsSource = from ads in xmlads.Descendants("ad") select new zestItem { imgurl = ads.Element("picture").Value }; public class zestItem { public string imgurl { get; set; } } private int _zIndex = 10; private void FlipImage(object sender, MouseButtonEventArgs e) { Image image = sender as Image; // Make sure the image is on top of all other images. image.SetValue(Canvas.ZIndexProperty, _zIndex++); // Create the storyboard. Storyboard flip = new Storyboard(); // Create animation and set the duration to 1 second. DoubleAnimation animation = new DoubleAnimation() { Duration = new TimeSpan(0, 0, 1) }; // Add the animation to the storyboard. flip.Children.Add(animation); // Create a projection for the image if it doesn't have one. if (image.Projection == null) { // Set the center of rotation to -0.01, which will put a little space // between the images when they're flipped. image.Projection = new PlaneProjection() { CenterOfRotationX = -0.01 }; } PlaneProjection projection = image.Projection as PlaneProjection; // Set the from and to properties based on the current flip direction of // the image. if (projection.RotationY == 0) { animation.To = 180; } else { animation.From = 180; animation.To = 0; } // Tell the animation to animation the image's PlaneProjection object. Storyboard.SetTarget(animation, projection); // Tell the animation to animation the RotationYProperty. Storyboard.SetTargetProperty(animation, new PropertyPath(PlaneProjection.RotationYProperty)); flip.Begin(); }

    Read the article

  • What are some things you'd like fresh college grads to know?

    - by bradhe
    So I proposed this to the Reddit community and I'd like to get SO's perspective on this. This is pretty much the copypasta of what I put there. I was thinking about this last night and thought it would be neat to compile a list. I'm still a pretty fresh college grad -- been in industry for 2 years -- but I think that I might have a few interesting things to lend. You don't know as much as you think you do. Somehow, college students think they know a lot more than they do (or maybe that was just me). Likewise, they think they can do more than they actually can. You should fairly assess your skills. QA people are not out to get you. Humans introduce bugs to code. It's not (nescessarily) a personal reflection on you and your skills if your code has a bug and it's caught by the QA/testing team. Listen to your senior (developers). They are not actually fuddy duddies who don't know about the new L337 hax in Ruby (okay, sometimes they are, but still...). They have a wealth of knowledge that you can learn from and it's in your best interest to do so. You will most likely not be doing what you want to for a while. This is mostly true in the corporate world -- startups are a different matter. Also, this is due to more than just the economy, man! Junior devs need to earn their keep, so to speak. Everyone wants to be lead dev on the next project and there are a lot of people in line ahead of you! For every elite developer there are 100 average developers. Joel Spolsky, I'm looking at you. Somehow this concept of ninja coders has really ingrained itself in our culture. While I encourage you to be the best you can be don't be disappointed if people aren't writing blog posts about you in the near future. Anyone else have anything they would see added to this list?

    Read the article

  • plane bombing problems- help

    - by peiska
    I'm training code problems, and on this one I am having problems to solve it, can you give me some tips how to solve it please. The problem is something like this: Your task is to find the sequence of points on the map that the bomber is expected to travel such that it hits all vital links. A link from A to B is vital when its absence isolates completely A from B. In other words, the only way to go from A to B (or vice versa) is via that link. Notice that if we destroy for example link (d,e), it becomes impossible to go from d to e,m,l or n in any way. A vital link can be hit at any point that lies in its segment (e.g. a hit close to d is as valid as a hit close to e). Of course, only one hit is enough to neutralize a vital link. Moreover, each bomb affects an exact circle of radius R, i.e., every segment that intersects that circle is considered hit. Due to enemy counter-attack, the plane may have to retreat at any moment, so the plane should follow, at each moment, to the closest vital link possible, even if in the end the total distance grows larger. Given all coordinates (the initial position of the plane and the nodes in the map) and the range R, you have to determine the sequence of positions in which the plane has to drop bombs. This sequence should start (takeoff) and finish (landing) at the initial position. Except for the start and finish, all the other positions have to fall exactly in a segment of the map (i.e. it should correspond to a point in a non-hit vital link segment). The coordinate system used will be UTM (Universal Transverse Mercator) northing and easting, which basically corresponds to a Euclidian perspective of the world (X=Easting; Y=Northing). Input Each input file will start with three floating point numbers indicating the X0 and Y0 coordinates of the airport and the range R. The second line contains an integer, N, indicating the number of nodes in the road network graph. Then, the next N (<10000) lines will each contain a pair of floating point numbers indicating the Xi and Yi coordinates (1 No two links will ever cross with each other. Output The program will print the sequence of coordinates (pairs of floating point numbers with exactly one decimal place), each one at a line, in the order that the plane should visit (starting and ending in the airport). Sample input 1 102.3 553.9 0.2 14 342.2 832.5 596.2 638.5 479.7 991.3 720.4 874.8 744.3 1284.1 1294.6 924.2 1467.5 659.6 1802.6 659.6 1686.2 860.7 1548.6 1111.2 1834.4 1054.8 564.4 1442.8 850.1 1460.5 1294.6 1485.1 17 1 2 1 3 2 4 3 4 4 5 4 6 6 7 7 8 8 9 8 10 9 10 10 11 6 11 5 12 5 13 12 13 13 14 Sample output 1 102.3 553.9 720.4 874.8 850.1 1460.5 102.3 553.9

    Read the article

  • Is there a programming language with be semantics close to English ?

    - by ivo s
    Most languages allow to 'tweek' to certain extend parts of the syntax (C++,C#) and/or semantics that you will be using in your code (Katahdin, lua). But I have not heard of a language that can just completely define how your code will look like. So isn't there some language which already exists that has such capabilities to override all syntax & define semantics ? Example of what I want to do is basically from the C# code below: foreach(Fruit fruit in Fruits) { if(fruit is Apple) { fruit.Price = fruit.Price/2; } } I want do be able to to write the above code in my perfect language like this: Check if any fruits are Macintosh apples and discount the price by 50%. The advantages that come to my mind looking from a coder's perspective in this "imaginary" language are: It's very clear what is going on (self descriptive) - it's plain English after all even kid would understand my program Hides all complexities which I have to write in C#. But why should I care to learn that if statements, arithmetic operators etc since there are already implemented The disadvantages that I see for a coder who will maintain this program are: Maybe you would express this program differently from me so you may not get all the information that I've expressed in my sentence Programs can be quite verbose and hard to debug but if possible to even proximate this type of syntax above maybe more people would start programming right? That would be amazing I think. I can go to work and just write an essay to draw a square on a winform like this: Create a form called MyGreetingForm. Draw a square with in the middle of MyGreetingFormwith a side of 100 points. In the middle of the square write "Hello! Click here to continue" in Arial font. In the above code the parser must basically guess that I want to use the unnamed square from the previous sentence, it'd be hard to write such a smart parser I guess, yet it's so simple what I want to do. If the user clicks on square in the middle of MyGreetingForm show MyMainForm. In the above code 'basically' the compiler must: 1)generate an event handler 2) check if there is any square in the middle of the form and if there is - 3) hide the form and show another form It looks very hard to do but it doesn't look impossible IMO to me at least approximate this (I can personally generate a parser to perform the 3 steps above np & it's basically the same that it has to do any way when you add even in c# a.MyEvent=+handler; so I don't see a problem here) so I'm thinking maybe somebody already did something like this ? Or is there some practical burden of complexity to create such a 'essay style' programming language which I can't see ? I mean what's the worse that can happen if the parser is not that good? - your program will crash so you have to re-word it:)

    Read the article

  • Primary language - QtC++, C#, Java?

    - by Airjoe
    I'm looking for some input, but let me start with a bit of background (for tl;dr skip to end). I'm an IT major with a concentration in networking. While I'm not a CS major nor do I want to program as a vocation, I do consider myself a programmer and do pretty well with the concepts involved. I've been programming since about 6th grade, started out with a proprietary game creation language that made my transition into C++ at college pretty easy. I like to make programs for myself and friends, and have been paid to program for local businesses. A bit about that- I wrote some programs for a couple local businesses in my senior year in high school. I wrote management systems for local shops (inventory, phone/pos orders, timeclock, customer info, and more stuff I can't remember). It definitely turned out to be over my head, as I had never had any formal programming education. It was a great learning experience, but damn was it crappy code. Oh yeah, by the way, it was all vb6. So, I've used vb6 pretty extensively, I've used c++ in my classes (intro to programming up to algorithms), used Java a little bit in another class (had to write a ping client program, pretty easy) and used Java for some simple Project Euler problems to help learn syntax and such when writing the program for the class. I've also used C# a bit for my own simple personal projects (simple programs, one which would just generate an HTTP request on a list of websites and notify if one responded unexpectedly or not at all, and another which just held a list of things to do and periodically reminded me to do them), things I would've written in vb6 a year or two ago. I've just started using Qt C++ for some undergrad research I'm working on. Now I've had some formal education, I [think I] understand organization in programming a lot better (I didn't even use classes in my vb6 programs where I really should have), how it's important to structure code, split into functions where appropriate, document properly, efficiency both in memory and speed, dynamic and modular programming etc. I was looking for some input on which language to pick up as my "primary". As I'm not a "real programmer", it will be mostly hobby projects, but will include some 'real' projects I'm sure. From my perspective: QtC++ and Java are cross platform, which is cool. Java and C# run in a virtual machine, but I'm not sure if that's a big deal (something extra to distribute, possibly a bit slower? I think Qt would require additional distributables too, right?). I don't really know too much more than this, so I appreciate any help, thanks! TL;DR Am an avocational programmer looking for a language, want quick and straight forward development, liked vb6, will be working with database driven GUI apps- should I go with QtC++, Java, C#, or perhaps something else?

    Read the article

  • help with generating models from database for many to many in doctrine

    - by ajsie
    im using doctrine and i have set up some test tables to be generated into models: I want a many-to-many relationship models (3 tables converted into 3 models) (things are simplified to make the point clear) mysql tables: user: id INT // primary key name VARCHAR group: id INT // primary key name VARCHAR user_group: user_id INT // primary and foreign key to user.id group_id INT // primary and foreign key to group.id i thought that it would generate these models (from the documentation): // User.php class User extends Doctrine_Record { public function setTableDefinition() { $this->hasColumn('id'); $this->hasColumn('name); } public function setUp() { $this->hasMany('Group as Groups', array( 'refClass' => 'UserGroup', 'local' => 'user_id', 'foreign' => 'group_id' ) ); } } // Group.php class Group extends Doctrine_Record { public function setTableDefinition() { $this->hasColumn('id'); $this->hasColumn('name); } public function setUp() { $this->hasMany('User as Users', array( 'refClass' => 'UserGroup', 'local' => 'group_id', 'foreign' => 'user_id' ) ); } } // UserGroup.php class UserGroup extends Doctrine_Record { public function setTableDefinition() { $this->hasColumn('user_id') ); $this->hasColumn('group_id') ); } } but it generated this: // User.php abstract class BaseUser extends Doctrine_Record { public function setTableDefinition() { $this->hasColumn('id'); $this->hasColumn('name'); } public function setUp() { $this->hasMany('UserGroup', array( 'local' => 'id', 'foreign' => 'user_id')); } } // Group.php abstract class BaseGroup extends Doctrine_Record { public function setTableDefinition() { $this->hasColumn('id'); $this->hasColumn('name'); } public function setUp() { $this->hasMany('UserGroup', array( 'local' => 'id', 'foreign' => 'group_id')); } } // UserGroup.php abstract class BaseUserGroup extends Doctrine_Record { public function setTableDefinition() { $this->hasColumn('user_id'); $this->hasColumn('group_id'); } public function setUp() { $this->hasOne('User', array( 'local' => 'user_id', 'foreign' => 'id')); $this->hasOne('Group', array( 'local' => 'group_id', 'foreign' => 'id')); } } as you can see, there is no 'refClass' in the 'User' and 'Group' models pointing to the 'UserGroup'. the 'UserGroup' table in this case is just another table from Doctrine's perspective not a reference table. I've checked my table definitions in mysql. They are correct. user_group has 2 columns (primary keys and foreign keys), each one pointing to the primary key in either User or Group. But i want the standard many-to-many relationship models in Doctrine models. I'd appreciate some help. I have struggled to figure it out for a half day now. What is wrong? Thanks!

    Read the article

  • Using Selenium-IDE with a rich Javascript application?

    - by Darien
    Problem At my workplace, we're trying to find the best way to create automated-tests for an almost wholly javascript-driven intranet application. Right now we're stuck trying to find a good tradeoff between: Application code in reusable and nest-able GUI components. Tests which are easily created by the testing team Tests which can be recorded once and then automated Tests which do not break after small cosmetic changes to the site XPath expressions (or other possible expressions, like jQuery selectors) naively generated from Selenium-IDE are often non-repeatable and very fragile. Conversely, having the JS code generate special unique ID values for every important DOM-element on the page... well, that is its own headache, complicated by re-usable GUI components and IDs needing to be consistent when the test is re-run. What successes have other people had with this kind of thing? How do you do automated application-level testing of a rich JS interface? Limitations We are using JavascriptMVC 2.0, hopefully 3.0 soon so that we can upgrade to jQuery 1.4.x. The test-making folks are mostly trained to use Selenium IDE to directly record things. The test leads would prefer a page-unique HTML ID on each clickable element on the page... Training the testers to write or alter special expressions (such as telling them which HTML class-names are important branching points) is a no-go. We try to make re-usable javascript components, but this means very few GUI components can treat themselves (or what they contain) as unique. Some of our components already use HTML ID values in their operation. I'd like to avoid doing this anyway, but it complicates the idea of ID-based testing. It may be possible to add custom facilities (like a locator-builder or new locator method) to the Selenium-IDE installation testers use. Almost everything that goes on occurs within a single "page load" from a conventional browser perspective, even when items are saved Current thoughts I'm considering a system where a custom locator-builder (javascript code) for Selenium-IDE will talk with our application code as the tester is recording. In this way, our application becomes partially responsible for generating a mostly-flexible expression (XPath or jQuery) for any given DOM element. While this can avoid requiring more training for testers, I worry it may be over-thinking things.

    Read the article

  • Device drivers and Windows

    - by b-gen-jack-o-neill
    Hi, I am trying to complete the picture of how the PC and the OS interacts together. And I am at point, where I am little out of guess when it comes to device drivers. Please, don´t write things like its too complicated, or you don´t need to know when using high programming laguage and winapi functions. I want to know, it´s for study purposes. So, the very basic structure of how OS and PC (by PC I mean of course HW) is how I see it is that all other than direct CPU commands, which can CPU do on itself (arithmetic operation, its registers access and memory access) must pass thru OS. Mainly becouse from ring level 3 you cannot use in and out intructions which are used for acesing other HW. I know that there is MMIO,but it must be set by port comunication first. It was not like this all the time. Even I am bit young to remember MSDOS, I know you could access HW directly, becouse there ws no limitation, no ring mode. So you could to write string to diplay use wheather DOS function, or directly acess video card memory and write it by yourself. But as OS developed, there is no longer this possibility. But it is fine, since OS now handles all the HW comunication, and frankly it more convinient and much more safe (I would say the only option) in multitasking environment. So nowdays you instead of using int instructions to use BIOS mapped function or DOS function you call dll which internally than handles everything you don´t need to know about. I understand this. I also undrstand that device drivers is the piece of code that runs in ring level 0, so it can do all the HW interactions. But what I don´t understand is connection between OS and device driver. Let´s take a example - I want to make a sound card make a sound. So I call windows API to acess sound card, but what happens than? Does windows call device drivers to do so? But if it does call device driver, does it mean, that all device drivers which can be called by winAPI function, must have routines named in some specific way? I mean, when I have new sound card, must its drivers have functions named same as the old one? So Windows can actually call the same function from its perspective? But if Windows have predefined sets of functions requored by device drivers, that it cannot use new drivers that doesent existed before last version of OS came out. Please, help me understand this mess. I am really getting mad. Thanks.

    Read the article

  • What is the best approach in SQL to store multi-level descriptions?

    - by gime
    I need a new perspective on how to design a reliable and efficient SQL database to store multi-level arrays of data. This problem applies to many situations but I came up with this example: There are hundreds of products. Each product has an undefined number of parts. Each part is built from several elements. All products are described in the same way. All parts would require the same fields to describe them (let's say: price, weight, part name), all elements of all parts also have uniform design (for example: element code, manufacturer). Plain and simple. One element may be related to only part, and each part is related to one product only. I came up with idea of three tables: Products: -------------------------------------------- prod_id prod_name prod_price prod_desc 1 hoover 120 unused next Parts: ---------------------------------------------------- part_id part_name part_price part_weight prod_id 3 engine 10 20 1 and finally Elements: --------------------------------------- el_id el_code el_manufacturer part_id 1 BFG12 GE 3 Now, select a desired product, select all from PARTS where prod_id is the same, and then select all from ELEMENTS where part_id matches - after multiple queries you've got all data. I'm just not sure if this is the right approach. I've got also another idea, without ELEMENTS table. That would decrease queries but I'm a bit afraid it might be lame and bad practice. Instead of ELEMENTS table there are two more fields in the PARTS table, so it looks like this: part_id, part_name, part_price, part_weight, prod_id, part_el_code, part_el_manufacturer they would be text type, and for each part, information about elements would be stored as strings, this way: part_el_code | code_of_element1; code_of_element2; code_of_element3 part_el_manufacturer | manuf_of_element1; manuf_of_element2; manuf_of_element3 Then all we need is to explode() data from those fields, and we get arrays, easy to display. Of course this is not perfect and has some limitations, but is this idea ok? Or should I just go with the first idea? Or maybe there is a better approach to this problem? It's really hard to describe it in few words, and that means it's hard to search for answer. Also, understanding the principles of designing databases is not that easy as it seems.

    Read the article

  • How to calculate where bullet hits

    - by lkjoel
    I have been trying to write an FPS in C/X11/OpenGL, but the issue that I have encountered is with calculating where the bullet hits. I have used a horrible technique, and it only sometimes works: pos size, p; size.x = 0.1; size.z = 0.1; // Since the game is technically top-down (but in a 3D perspective) // Positions are in X/Z, no Y float f; // Counter float d = FIRE_MAX + 1 /* Shortest Distance */, d1 /* Distance being calculated */; x = 0; // Index of object to hit for (f = 0.0; f < FIRE_MAX; f += .01) { // Go forwards p.x = player->pos.x + f * sin(toRadians(player->rot.x)); p.z = player->pos.z - f * cos(toRadians(player->rot.x)); // Get all objects that collide with the current position of the bullet short* objs = _colDetectGetObjects(p, size, objects); for (i = 0; i < MAX_OBJECTS; i++) { if (objs[i] == -1) { continue; } // Check the distance between the object and the player d1 = sqrt( pow((objects[i].pos.x - player->pos.x), 2) + pow((objects[i].pos.z - player->pos.z), 2)); // If it's closer, set it as the object to hit if (d1 < d) { x = i; d = d1; } } // If there was an object, hit it if (x > 0) { hit(&objects[x], FIRE_DAMAGE, explosions, currtime); break; } } It just works by making a for-loop and calculating any objects that might collide with where the bullet currently is. This, of course, is very slow, and sometimes doesn't even work. What would be the preferred way to calculate where the bullet hits? I have thought of making a line and seeing if any objects collide with that line, but I have no idea how to do that kind of collision detection. EDIT: I guess my question is this: How do I calculate the nearest object colliding in a line (that might not be a straight 45/90 degree angle)? Or are there any simpler methods of calculating where the bullet hits? The bullet is sort of like a laser, in the sense that gravity does not affect it (writing an old-school game, so I don't want it to be too realistic)

    Read the article

  • Parallelism in .NET – Part 10, Cancellation in PLINQ and the Parallel class

    - by Reed
    Many routines are parallelized because they are long running processes.  When writing an algorithm that will run for a long period of time, its typically a good practice to allow that routine to be cancelled.  I previously discussed terminating a parallel loop from within, but have not demonstrated how a routine can be cancelled from the caller’s perspective.  Cancellation in PLINQ and the Task Parallel Library is handled through a new, unified cooperative cancellation model introduced with .NET 4.0. Cancellation in .NET 4 is based around a new, lightweight struct called CancellationToken.  A CancellationToken is a small, thread-safe value type which is generated via a CancellationTokenSource.  There are many goals which led to this design.  For our purposes, we will focus on a couple of specific design decisions: Cancellation is cooperative.  A calling method can request a cancellation, but it’s up to the processing routine to terminate – it is not forced. Cancellation is consistent.  A single method call requests a cancellation on every copied CancellationToken in the routine. Let’s begin by looking at how we can cancel a PLINQ query.  Supposed we wanted to provide the option to cancel our query from Part 6: double min = collection .AsParallel() .Min(item => item.PerformComputation()); .csharpcode, .csharpcode pre { font-size: small; color: black; font-family: consolas, "Courier New", courier, monospace; background-color: #ffffff; /*white-space: pre;*/ } .csharpcode pre { margin: 0em; } .csharpcode .rem { color: #008000; } .csharpcode .kwrd { color: #0000ff; } .csharpcode .str { color: #006080; } .csharpcode .op { color: #0000c0; } .csharpcode .preproc { color: #cc6633; } .csharpcode .asp { background-color: #ffff00; } .csharpcode .html { color: #800000; } .csharpcode .attr { color: #ff0000; } .csharpcode .alt { background-color: #f4f4f4; width: 100%; margin: 0em; } .csharpcode .lnum { color: #606060; } We would rewrite this to allow for cancellation by adding a call to ParallelEnumerable.WithCancellation as follows: var cts = new CancellationTokenSource(); // Pass cts here to a routine that could, // in parallel, request a cancellation try { double min = collection .AsParallel() .WithCancellation(cts.Token) .Min(item => item.PerformComputation()); } catch (OperationCanceledException e) { // Query was cancelled before it finished } .csharpcode, .csharpcode pre { font-size: small; color: black; font-family: consolas, "Courier New", courier, monospace; background-color: #ffffff; /*white-space: pre;*/ } .csharpcode pre { margin: 0em; } .csharpcode .rem { color: #008000; } .csharpcode .kwrd { color: #0000ff; } .csharpcode .str { color: #006080; } .csharpcode .op { color: #0000c0; } .csharpcode .preproc { color: #cc6633; } .csharpcode .asp { background-color: #ffff00; } .csharpcode .html { color: #800000; } .csharpcode .attr { color: #ff0000; } .csharpcode .alt { background-color: #f4f4f4; width: 100%; margin: 0em; } .csharpcode .lnum { color: #606060; } Here, if the user calls cts.Cancel() before the PLINQ query completes, the query will stop processing, and an OperationCanceledException will be raised.  Be aware, however, that cancellation will not be instantaneous.  When cts.Cancel() is called, the query will only stop after the current item.PerformComputation() elements all finish processing.  cts.Cancel() will prevent PLINQ from scheduling a new task for a new element, but will not stop items which are currently being processed.  This goes back to the first goal I mentioned – Cancellation is cooperative.  Here, we’re requesting the cancellation, but it’s up to PLINQ to terminate. If we wanted to allow cancellation to occur within our routine, we would need to change our routine to accept a CancellationToken, and modify it to handle this specific case: public void PerformComputation(CancellationToken token) { for (int i=0; i<this.iterations; ++i) { // Add a check to see if we've been canceled // If a cancel was requested, we'll throw here token.ThrowIfCancellationRequested(); // Do our processing now this.RunIteration(i); } } With this overload of PerformComputation, each internal iteration checks to see if a cancellation request was made, and will throw an OperationCanceledException at that point, instead of waiting until the method returns.  This is good, since it allows us, as developers, to plan for cancellation, and terminate our routine in a clean, safe state. This is handled by changing our PLINQ query to: try { double min = collection .AsParallel() .WithCancellation(cts.Token) .Min(item => item.PerformComputation(cts.Token)); } catch (OperationCanceledException e) { // Query was cancelled before it finished } PLINQ is very good about handling this exception, as well.  There is a very good chance that multiple items will raise this exception, since the entire purpose of PLINQ is to have multiple items be processed concurrently.  PLINQ will take all of the OperationCanceledException instances raised within these methods, and merge them into a single OperationCanceledException in the call stack.  This is done internally because we added the call to ParallelEnumerable.WithCancellation. If, however, a different exception is raised by any of the elements, the OperationCanceledException as well as the other Exception will be merged into a single AggregateException. The Task Parallel Library uses the same cancellation model, as well.  Here, we supply our CancellationToken as part of the configuration.  The ParallelOptions class contains a property for the CancellationToken.  This allows us to cancel a Parallel.For or Parallel.ForEach routine in a very similar manner to our PLINQ query.  As an example, we could rewrite our Parallel.ForEach loop from Part 2 to support cancellation by changing it to: try { var cts = new CancellationTokenSource(); var options = new ParallelOptions() { CancellationToken = cts.Token }; Parallel.ForEach(customers, options, customer => { // Run some process that takes some time... DateTime lastContact = theStore.GetLastContact(customer); TimeSpan timeSinceContact = DateTime.Now - lastContact; // Check for cancellation here options.CancellationToken.ThrowIfCancellationRequested(); // If it's been more than two weeks, send an email, and update... if (timeSinceContact.Days > 14) { theStore.EmailCustomer(customer); customer.LastEmailContact = DateTime.Now; } }); } catch (OperationCanceledException e) { // The loop was cancelled } Notice that here we use the same approach taken in PLINQ.  The Task Parallel Library will automatically handle our cancellation in the same manner as PLINQ, providing a clean, unified model for cancellation of any parallel routine.  The TPL performs the same aggregation of the cancellation exceptions as PLINQ, as well, which is why a single exception handler for OperationCanceledException will cleanly handle this scenario.  This works because we’re using the same CancellationToken provided in the ParallelOptions.  If a different exception was thrown by one thread, or a CancellationToken from a different CancellationTokenSource was used to raise our exception, we would instead receive all of our individual exceptions merged into one AggregateException.

    Read the article

  • Another Marketing Conference, part two – the afternoon

    - by Roger Hart
    In my previous post, I’ve covered the morning sessions at AMC2012. Here’s the rest of the write-up. I’ve skipped Charles Nixon’s session which was a blend of funky futurism and professional development advice, but you can see his slides here. I’ve also skipped the Google presentation, as it was a little thin on insight. 6 – Brand ambassadors: Getting universal buy in across the organisation, Vanessa Northam Slides are here This was the strongest enforcement of the idea that brand and campaign values need to be delivered throughout the organization if they’re going to work. Vanessa runs internal communications at e-on, and shared her experience of using internal comms to align an organization and thereby get the most out of a campaign. She views the purpose of internal comms as: “…to help leaders, to communicate the purpose and future of an organization, and support change.” This (and culture) primes front line staff, which creates customer experience and spreads brand. You ensure a whole organization knows what’s going on with both internal and external comms. If everybody is aligned and informed, if everybody can clearly articulate your brand and campaign goals, then you can turn everybody into an advocate. Alignment is a powerful tool for delivering a consistent experience and message. The pathological counter example is the one in which a marketing message goes out, which creates inbound customer contacts that front line contact staff haven’t been briefed to handle. The NatWest campaign was again mentioned in this context. The good example was e-on’s cheaper tariff campaign. Building a groundswell of internal excitement, and even running an internal launch meant everyone could contribute to a good customer experience. They found that meter readers were excited – not a group they’d considered as obvious in providing customer experience. But they were a group that has a lot of face-to-face contact with customers, and often were asked questions they may not have been briefed to answer. Being able to communicate a simple new message made it easier for them, and also let them become a sales and marketing asset to the organization. 7 – Goodbye Internet, Hello Outernet: the rise and rise of augmented reality, Matt Mills I wasn’t going to write this up, because it was essentially a sales demo for Aurasma. But the technology does merit some discussion. Basically, it replaces QR codes with visual recognition, and provides a simple-looking back end for attaching content. It’s quite sexy. But here’s my beef with it: QR codes had a clear visual language – when you saw one you knew what it was and what to do with it. They were clunky, but they had the “getting started” problem solved out of the box once you knew what you were looking at. However, they fail because QR code reading isn’t native to the platform. You needed an app, which meant you needed to know to download one. Consequentially, you can’t use QR codes with and ubiquity, or depend on them. This means marketers, content providers, etc, never pushed them, and they remained and awkward oddity, a minority sport. Aurasma half solves problem two, and re-introduces problem one, making it potentially half as useful as a QR code. It’s free, and you can apparently build it into your own apps. Add to that the likelihood of it becoming native to the platform if it takes off, and it may have legs. I guess we’ll see. 8 – We all need to code, Helen Mayor Great title – good point. If there was anybody in the room who didn’t at least know basic HTML, and if Helen’s presentation inspired them to learn, that’s fantastic. However, this was a half hour sales pitch for a basic coding training course. Beyond advocating coding skills it contained no useful content. Marketers may also like to consider some of these resources if they’re looking to learn code: Code Academy – free interactive tutorials Treehouse – learn web design, web dev, or app dev WebPlatform.org – tutorials and documentation for web tech  11 – Understanding our inner creativity, Margaret Boden This session was the most theoretical and probably least actionable of the day. It also held my attention utterly. Margaret spoke fluently, fascinatingly, without slides, on the subject of types of creativity and how they work. It was splendid. Yes, it raised a wry smile whenever she spoke of “the content of advertisements” and gave an example from 1970s TV ads, but even without the attempt to meet the conference’s theme this would have been thoroughly engaging. There are, Margaret suggested, three types of creativity: Combinatorial creativity The most common form, and consisting of synthesising ideas from existing and familiar concepts and tropes. Exploratory creativity Less common, this involves exploring the limits and quirks of a particular constraint or style. Transformational creativity This is uncommon, and arises from finding a way to do something that the existing rules would hold to be impossible. In essence, this involves breaking one of the constraints that exploratory creativity is composed from. Combinatorial creativity, she suggested, is particularly important for attaching favourable ideas to existing things. As such is it probably worth developing for marketing. Exploratory creativity may then come into play in something like developing and optimising an idea or campaign that now has momentum. Transformational creativity exists at the edges of this exploration. She suggested that products may often be transformational, but that marketing seemed unlikely to in her experience. This made me wonder about Listerine. Crucially, transformational creativity is characterised by there being some element of continuity with the strictures of previous thinking. Once it has happened, there may be  move from a revolutionary instance into an explored style. Again, from a marketing perspective, this seems to chime well with the thinking in Youngme Moon’s book: Different Talking about the birth of Modernism is visual art, Margaret pointed out that transformational creativity has historically risked a backlash, demanding what is essentially an education of the market. This is best accomplished by referring back to the continuities with the past in order to make the new familiar. Thoughts The afternoon is harder to sum up than the morning. It felt less concrete, and was troubled by a short run of poor presentations in the middle. Mainly, I found myself wrestling with the internal comms issue. It’s one of those things that seems astonishingly obvious in hindsight, but any campaign – particularly any large one – is doomed if the people involved can’t believe in it. We’ve run things here that haven’t gone so well, of course we have; who hasn’t? I’m not going to air any laundry, but people not being informed (much less aligned) feels like a common factor. It’s tough though. Managing and anticipating information needs across an organization of any size can’t be easy. Even the simple things like ensuring sales and support departments know what’s in a product release, and what messages go with it are easy to botch. The thing I like about framing this as a brand and campaign advocacy problem is that it makes it likely to get addressed. Better is always sexier than less-worse. Any technical communicator who’s ever felt crowded out by a content strategist or marketing copywriter  knows this – increasing revenue gets a seat at the table far more readily than reducing support costs, even if the financial impact is identical. So that’s it from AMC. The big thought-provokers were social buying behaviour and eliciting behaviour change, and the value of internal communications in ensuring successful campaigns and continuity of customer experience. I’ll be chewing over that for a while, and I’d definitely return next year.      

    Read the article

  • Transparency and AlphaBlending

    - by TechTwaddle
    In this post we'll look at the AlphaBlend() api and how it can be used for semi-transparent blitting. AlphaBlend() takes a source device context and a destination device context (DC) and combines the bits in such a way that it gives a transparent effect. Follow the links for the msdn documentation. So lets take a image like, and AlphaBlend() it on our window. The code to do so is below, (under the WM_PAINT message of WndProc) HBITMAP hBitmap=NULL, hBitmapOld=NULL; HDC hMemDC=NULL; BLENDFUNCTION bf; hdc = BeginPaint(hWnd, &ps); hMemDC = CreateCompatibleDC(hdc); hBitmap = LoadBitmap(g_hInst, MAKEINTRESOURCE(IDB_BITMAP1)); hBitmapOld = SelectObject(hMemDC, hBitmap); bf.BlendOp = AC_SRC_OVER; bf.BlendFlags = 0; bf.SourceConstantAlpha = 80; //transparency value between 0-255 bf.AlphaFormat = 0;    AlphaBlend(hdc, 0, 25, 240, 100, hMemDC, 0, 0, 240, 100, bf); SelectObject(hMemDC, hBitmapOld); DeleteDC(hMemDC); DeleteObject(hBitmap); EndPaint(hWnd, &ps);   The code above creates a memory DC (hMemDC) using CreateCompatibleDC(), loads a bitmap onto the memory DC and AlphaBlends it on the device DC (hdc), with a transparency value of 80. The result is: Pretty simple till now. Now lets try to do something a little more exciting. Lets get two images involved, each overlapping the other, giving a better demonstration of transparency. I am also going to add a few buttons so that the user can increase or decrease the transparency by clicking on the buttons. Since this is the first time I played around with GDI apis, I ran into something that everybody runs into sometime or the other, flickering. When clicking the buttons the images would flicker a lot, I figured out why and used something called double buffering to avoid flickering. We will look at both my first implementation and the second implementation just to give the concept a little more depth and perspective. A few pre-conditions before I dive into the code: - hBitmap and hBitmap2 are handles to the two images obtained using LoadBitmap(), these variables are global and are initialized under WM_CREATE - The two buttons in the application are labeled Opaque++ (make more opaque, less transparent) and Opaque-- (make less opaque, more transparent) - DrawPics(HWND hWnd, int step=0); is the function called to draw the images on the screen. This is called from under WM_PAINT and also when the buttons are clicked. When Opaque++ is clicked the 'step' value passed to DrawPics() is +20 and when Opaque-- is clicked the 'step' value is -20. The default value of 'step' is 0 Now lets take a look at my first implementation: //this funciton causes flicker, cos it draws directly to screen several times void DrawPics(HWND hWnd, int step) {     HDC hdc=NULL, hMemDC=NULL;     BLENDFUNCTION bf;     static UINT32 transparency = 100;     //no point in drawing when transparency is 0 and user clicks Opaque--     if (transparency == 0 && step < 0)         return;     //no point in drawing when transparency is 240 (opaque) and user clicks Opaque++     if (transparency == 240 && step > 0)         return;         hdc = GetDC(hWnd);     if (!hdc)         return;     //create a memory DC     hMemDC = CreateCompatibleDC(hdc);     if (!hMemDC)     {         ReleaseDC(hWnd, hdc);         return;     }     //while increasing transparency, clear the contents of screen     if (step < 0)     {         RECT rect = {0, 0, 240, 200};         FillRect(hdc, &rect, (HBRUSH)GetStockObject(WHITE_BRUSH));     }     SelectObject(hMemDC, hBitmap2);     BitBlt(hdc, 0, 25, 240, 100, hMemDC, 0, 0, SRCCOPY);         SelectObject(hMemDC, hBitmap);     transparency += step;     if (transparency >= 240)         transparency = 240;     if (transparency <= 0)         transparency = 0;     bf.BlendOp = AC_SRC_OVER;     bf.BlendFlags = 0;     bf.SourceConstantAlpha = transparency;     bf.AlphaFormat = 0;            AlphaBlend(hdc, 0, 75, 240, 100, hMemDC, 0, 0, 240, 100, bf);     DeleteDC(hMemDC);     ReleaseDC(hWnd, hdc); }   In the code above, we first get the window DC using GetDC() and create a memory DC using CreateCompatibleDC(). Then we select hBitmap2 onto the memory DC and Blt it on the window DC (hdc). Next, we select the other image, hBitmap, onto memory DC and AlphaBlend() it over window DC. As I told you before, this implementation causes flickering because it draws directly on the screen (hdc) several times. The video below shows what happens when the buttons were clicked rapidly: Well, the video recording tool I use captures only 15 frames per second and so the flickering is not visible in the video. So you're gonna have to trust me on this, it flickers (; To solve this problem we make sure that the drawing to the screen happens only once and to do that we create an additional memory DC, hTempDC. We perform all our drawing on this memory DC and finally when it is ready we Blt hTempDC on hdc, and the images are displayed in one go. Here is the code for our new DrawPics() function: //no flicker void DrawPics(HWND hWnd, int step) {     HDC hdc=NULL, hMemDC=NULL, hTempDC=NULL;     BLENDFUNCTION bf;     HBITMAP hBitmapTemp=NULL, hBitmapOld=NULL;     static UINT32 transparency = 100;     //no point in drawing when transparency is 0 and user clicks Opaque--     if (transparency == 0 && step < 0)         return;     //no point in drawing when transparency is 240 (opaque) and user clicks Opaque++     if (transparency == 240 && step > 0)         return;         hdc = GetDC(hWnd);     if (!hdc)         return;     hMemDC = CreateCompatibleDC(hdc);     hTempDC = CreateCompatibleDC(hdc);     hBitmapTemp = CreateCompatibleBitmap(hdc, 240, 150);     hBitmapOld = (HBITMAP)SelectObject(hTempDC, hBitmapTemp);     if (!hMemDC)     {         ReleaseDC(hWnd, hdc);         return;     }     //while increasing transparency, clear the contents     if (step < 0)     {         RECT rect = {0, 0, 240, 150};         FillRect(hTempDC, &rect, (HBRUSH)GetStockObject(WHITE_BRUSH));     }     SelectObject(hMemDC, hBitmap2);     //Blt hBitmap2 directly to hTempDC     BitBlt(hTempDC, 0, 0, 240, 100, hMemDC, 0, 0, SRCCOPY);         SelectObject(hMemDC, hBitmap);     transparency += step;     if (transparency >= 240)         transparency = 240;     if (transparency <= 0)         transparency = 0;     bf.BlendOp = AC_SRC_OVER;     bf.BlendFlags = 0;     bf.SourceConstantAlpha = transparency;     bf.AlphaFormat = 0;            AlphaBlend(hTempDC, 0, 50, 240, 100, hMemDC, 0, 0, 240, 100, bf);     //now hTempDC is ready, blt it directly on hdc     BitBlt(hdc, 0, 25, 240, 150, hTempDC, 0, 0, SRCCOPY);     SelectObject(hTempDC, hBitmapOld);     DeleteObject(hBitmapTemp);     DeleteDC(hMemDC);     DeleteDC(hTempDC);     ReleaseDC(hWnd, hdc); }   This function is very similar to the first version, except for the use of hTempDC. Another point to note is the use of CreateCompatibleBitmap(). When a memory device context is created using CreateCompatibleDC(), the context is exactly one monochrome pixel high and one monochrome pixel wide. So in order for us to draw anything onto hTempDC, we first have to set a bitmap on it. We use CreateCompatibleBitmap() to create a bitmap of required dimension (240x150 above), and then select this bitmap onto hTempDC. Think of it as utilizing an extra canvas, drawing everything on the canvas and finally transferring the contents to the display in one scoop. And with this version the flickering is gone, video follows:   If you want the entire solutions source code then leave a message, I will share the code over SkyDrive.

    Read the article

  • So, how is the Oracle HCM Cloud User Experience? In a word, smokin’!

    - by Edith Mireles-Oracle
    By Misha Vaughan, Oracle Applications User Experience Oracle unveiled its game-changing cloud user experience strategy at Oracle OpenWorld 2013 (remember that?) with a new simplified user interface (UI) paradigm.  The Oracle HCM cloud user experience is about light-weight interaction, tailored to the task you are trying to accomplish, on the device you are comfortable working with. A key theme for the Oracle user experience is being able to move from smartphone to tablet to desktop, with all of your data in the cloud. The Oracle HCM Cloud user experience provides designs for better productivity, no matter when and how your employees need to work. Release 8  Oracle recently demonstrated how fast it is moving development forward for our cloud applications, with the availability of release 8.  In release 8, users will see expanded simplicity in the HCM cloud user experience, such as filling out a time card and succession planning. Oracle has also expanded its mobile capabilities with task flows for payslips, managing absences, and advanced analytics. In addition, users will see expanded extensibility with the new structures editor for simplified pages, and the with the user interface text editor, which allows you to update language throughout the UI from one place. If you don’t like calling people who work for you “employees,” you can use this tool to create a term that is suited to your business.  Take a look yourself at what’s available now. What are people saying?Debra Lilley (@debralilley), an Oracle ACE Director who has a long history with Oracle Applications, recently gave her perspective on release 8: “Having had the privilege of seeing a preview of release 8, I am again impressed with the enhancements around simplified UI. Even more so, at a user group event in London this week, an existing Cloud HCM customer speaking publically about his implementation said he was very excited about release 8 as the absence functionality was so superior and simple to use.”  In an interview with Lilley for a blog post by Dennis Howlett  (@dahowlett), we probably couldn’t have asked for a more even-handed look at the Oracle Applications Cloud and the impact of user experience. Take the time to watch all three videos and get the full picture.  In closing, Howlett’s said: “There is always the caveat that getting from the past to Fusion [from the editor: Fusion is now called the Oracle Applications Cloud] is not quite as simple as may be painted, but the outcomes are much better than anticipated in large measure because the user experience is so much better than what went before.” Herman Slange, Technical Manager with Oracle Applications partner Profource, agrees with that comment. “We use on-premise Financials & HCM for internal use. Having a simple user interface that works on a desktop as well as a tablet for (very) non-technical users is a big relief. Coming from E-Business Suite, there is less training (none) required to access HCM content.  From a technical point of view, having the abilities to tailor the simplified UI very easy makes it very efficient for us to adjust to specific customer needs.  When we have a conversation about simplified UI, we just hand over a tablet and ask the customer to just use it. No training and no explanation required.” Finally, in a story by Computer Weekly  about Oracle customer BG Group, a natural gas exploration and production company based in the UK and with a presence in 20 countries, the author states: “The new HR platform has proved to be easier and more intuitive for HR staff to use than the previous SAP-based technology.” What’s Next for Oracle’s Applications Cloud User Experiences? This is the question that Steve Miranda, Oracle Executive Vice President, Applications Development, asks the Applications User Experience team, and we’ve been hard at work for some time now on “what’s next.”  I can’t say too much about it, but I can tell you that we’ve started talking to customers and partners, under non-disclosure agreements, about user experience concepts that we are working on in order to get their feedback. We recently had a chance to talk about possibilities for the Oracle HCM Cloud user experience at an Oracle HCM Southern California Customer Success Summit. This was a fantastic event, hosted by Shane Bliss and Vance Morossi of the Oracle Client Success Team. We got to use the uber-slick facilities of Allergan, our hosts (of Botox fame), headquartered in Irvine, Calif., with a presence in more than 100 countries. Photo by Misha Vaughan, Oracle Applications User Experience Vance Morossi, left, and Shane Bliss, of the Oracle Client Success Team, at an Oracle HCM Southern California Customer Success Summit.  We were treated to a few really excellent talks around human resources (HR). Alice White, VP Human Resources, discussed Allergan's process for global talent acquisition -- how Allergan has designed and deployed a global process, and global tools, along with Oracle and Cognizant, and are now at the end of a global implementation. She shared a couple of insights about the journey for Allergan: “One of the major areas for improvement was on role clarification within the company.” She said the company is “empowering managers and deputizing them as recruiters. Now it is a global process that is nimble and efficient."  Deepak Rammohan, VP Product Management, HCM Cloud, Oracle, also took the stage to talk about pioneering modern HR. He reflected modern HR problems of getting the right data about the workforce, the importance of getting the right talent as a key strategic initiative, and other workforce insights. "How do we design systems to deal with all of this?” he asked. “Make sure the systems are talent-centric. The next piece is collaborative, engaging, and mobile. A lot of this is influenced by what users see today. The last thing is around insight; insight at the point of decision-making." Rammohan showed off some killer HCM Cloud talent demos focused on simplicity and mobility that his team has been cooking up, and closed with a great line about the nature of modern recruiting: "Recruiting is a team sport." Deepak Rammohan, left, and Jake Kuramoto, both of Oracle, debate the merits of a Google Glass concept demo for recruiters on-the-go. Later, in an expo-style format, the Apps UX team showed several concepts for next-generation HCM Cloud user experiences, including demos shown by Jake Kuramoto (@jkuramoto) of The AppsLab, and Aylin Uysal (@aylinuysal), Director, HCM Cloud user experience. We even hauled out our eye-tracker, a research tool used to show where the eye is looking at a particular screen, thanks to teammate Michael LaDuke. Dionne Healy, HCM Client Executive, and Aylin Uysal, Director, HCM Cloud user experiences, Oracle, take a look at new HCM Cloud UX concepts. We closed the day with Jeremy Ashley (@jrwashley), VP, Applications User Experience, who brought it all back together by talking about the big picture for applications cloud user experiences. He covered the trends we are paying attention to now, what users will be expecting of their modern enterprise apps, and what Oracle’s design strategy is around these ideas.   We closed with an excellent reception hosted by ADP Payroll services at Bistango. Want to read more?Want to see where our cloud user experience is going next? Read more on the UsableApps web site about our latest design initiative: “Glance, Scan, Commit.” Or catch up on the back story by looking over our Applications Cloud user experience content on the UsableApps web site.  You can also find out where we’ll be next at the Events page on UsableApps.

    Read the article

< Previous Page | 45 46 47 48 49 50 51 52 53 54 55  | Next Page >