In fact, I have not programmed under any Windows programming language.
That includes Visual C++, J++ (if you have heard of that), C#, etc., etc., etc.
I can only suppose that you haven't programmed in anything then. Because certainly C++, smalltalk, python, perl, ruby and even Cobol, Fortran and Lisp are "Windows programming language[s]".
Also includes Java, especially so, given that Java was released first on windows. (And with a thread model that was so dependent on windows that it took several years to correctly implement something in Solaris.)
Perhaps you use Snobol or Algol? I don't recall seeing those on windows.
As to the crap called RDBMS -- which is taught in all the colleges, universities and technical institutes so that we can all become slaves of Oracle/Sybase/SQLServer, what none of you realize is that Oracle on IBM mainframes was implemented as a set of KSDS data sets, which provides unique primary keys and multiple non-unique secondary keys for any record. Anything other than that in a RDBMS is syntactic sugar that gets translated into these fundamentals.
Yes but everyone knows that that is just a conspiracy by the Illuminati to subjugate the world and that the incredibly vast amount of research on relational databases is all fabricated by super genius aliens that come from Betelgeuse.
My point was that .NET did not exist in those days. So all your experience in C# and J++ would be irrelevant.
That is why I put a date of 1995.
I agree that DOS-like interfaces are enough for most applications. OTOH, having a simple GUI allows not-so-technical users to work with software. No, I do not miss DBase IV (if you have heard of that), nor WP5.1.
..and yes, I think World of Warcraft is awesome. I'd like to see you do that with 1995-technologies
As to the crap called RDBMS
Most people I encounter dislike databases. And cascading deletes. And triggers.
Most people dislike what they don't understand. I got the same with ASP.NET, being a complete other beast than the WinForms I'm used to. No, that doesn't make me label it as "crap".
But, thinking is not the forte of the current crop of programmers who run after the latest and greatest scripting language/framework/any other bullcrap peddled by software vendors.
Ah, so you're blaming the engineers of said vendors? IIRC, there's a white collar that makes those decisions, and always prioritizing the looks above functionality. Why? Simple; a program has to look up to date, meaning that it needs to have the same colors as the latest version of Windows or Office. Why? Because those colors are the "sexy colors" for that particular year.
Perhaps by coming up with a new marketing phrase that means nothing?
That means one has to abandon the path we have collectively taken for the 35 years or more. Specifically, abandon the crap called relational DBMS.
Joke right? You are supposing that the OP question can be achieved and is not now a reality solely because people use relational databases? I suppose you figure that people that use NoSQL are already 'computer whispers'?
Take any Windows program and see how much glue you have in it. Figure out how to remove the glue.
Follow the thought process all the way through.
Following through I can only suppose that you are claiming Unix developers using Java are now 'computer whispers'?
You will realize that we could have stopped programming around 1995.
I suspect that some people should have stopped programming then.
And since then some others probably shouldn't have started in the first place.
But that is because they have no aptitude for it.
But other than fantasy (not science fiction) novels no one can stop programming now. And based on current research there doesn't seem to be any indications it will be possible even in any mid term future scenario. I wouldn't be surprised that due to complexity issues it will never occur.
I'm trying to develop a server, using C#, on which are connected a printer and clients. Whenever a client want to print a document, the server must recognize him to verify some permissions before continuing to printing.
Could anyone tell me when the client click on "print", the server will know about that (will he receive that request as the printer's driver is installed only there) ?
or he have to check the Queue of the printer to know the creator of every process ?
Thank you in advance!
If the user starts the print-action from, say, his browser, then the local computer will display a list of all available printers. If the printer is in the network, and the local computer has the correct drivers, it'll be able to print to there. Permissions are granted over the Active Directory.
Printing is not an action that "raises an event". You could add code that instructs the server to print something on behalf of the user.
I have rehosted the workflow designer and I've added a few custom activities along with default activities in it. It is going to be used by users to create and edit control flows. All I want to do is to plug the designer in a predefined architecture which includes a DLL, BLL and presentation layer. What architecture do you suggest for the new case including workflow designer?
That is common practice with devices like TV set top boxes. Many use the Linux kernel and some required OS specific programs (often BusyBox) from the open source side and implement the main function in closed source applications.
The source of the kernel and all code based on open source must be published according to their licenses. Each program written from scratch by you must not be published.
A special case is linking open source libraries. These must have a GPL linking exception or are published under a license allowing such linkage like LGPL.
Assuming you want to use the Linux kernel but implement nearly all other code yourself, I think it is possible. But it would be a lot of work and requires checking the licenses of the used open source components.
I am reading the bridge design pattern from the GoF book, there is a variant defined which the author calls "Sharing implementors" illustrated by handle body / idiom in c++. I did some research to find its implmenetation but i am unable to see Reference counting technique to allow multiple objects with the same value to share a single representation of that value. can any body help me please to learn this?
for our project we use Windsor Castle to do IoC. This means that we also let the IoC container resolve the needed constructor arguments. But for some cases (wrappers, view models, etc.) we also need to pass some ctor arguments which are only known at runtime and from my opinion are needed to create a valid object. For this we use generic factories like:
For this to work I have IoC registered that factory as well as IMyWrapper, where the implementation of IMyWrapper only has one ctor overload with the string argument.
Problem: we also have created an integration test which instantiates all registered components to ensure that everything can be resolved. But this fails for obvious reasons with IMyWrapper, as the string argument is not registered!
I would be very happy if you could tell me what you think about it.
1. Is there a way to let certain components exclusively be instantiated by generic factories so the test could differ and it would not be possible to acquire an instance directly without using the needed string argument?
2. Am I wrong and it is generally bad practice to provide runtime parameters to constructors?
3. If not, would you limit your freedom of doing so just to make the integration test check 100% well?
My question is whether there are best practices around color-coding flow-charts and similar diagrams. Do certain colors belong to certain symbols? (such as Yellow for "decisions" while green for "data") If so, what are they? Are there resources online that you know of regarding this?
I have an Ethernet connected motion controller that I need to get information from. I'm using WPF (C# flavor) for the project. The goal is to get data from the device as fast as possible so the UI doesn't seem sluggish.
My thought was to spin off a background task that constantly polled the device for information as fast as it could run. It would fire events when certain groups of data were updated. Another class would monitor for these events and then set dependency properties that would update the UI.
I have a couple questions:
1) What is the best way of starting the background task? I want it to close when the main thread is closed and not stop the main thread from being closed.
2) Any helpful advice as to how data can be retrieved on one thread that will eventually be displayed on a different thread.
I have only toyed with multi-threaded applications up to this point. I'd rather ask and do it right than to find out later there was a better way.
This is on a closed network on the machine. It is not going over a corporate network.
The device does not fire events. It has an API that allows data to be requested and then it replies. Right now I am using a DispatchTimer to request updates at a regular interval. It is running at 25ms without any issues. I was looking at trying to free my UI from my polling routine in case the device locks up or stops responding.
If you think you can, you will.
If you think you can't, you won't.
Either way, you're right.
Last Visit: 31-Dec-99 18:00 Last Update: 22-Sep-23 9:34