Printable View
OUR clients ? How is that THEY mess up industry, and yet pursuade OUR clients ? Aren't those clients THEIR clients ? I think you ought to worry about YOUR clients, with your own bugs, and let THEM deal with THEIR clients.Quote:
therwise we will continue to be flooded with VB6-like programmers who mess up the industry and pursuade our clients that applications are full of bugs and are generally awful.
Yes, it is slower, but if correctly used (that means, critical routines written in C++ or assembler, and global manipulation done with .NET), it can slow down the application by a so little factor that it is not noticeable.Quote:
Originally Posted by BlackWindzx
And the code is more maintainable, less buggy, and have no memory leaks.
And for the user point of vue this application is really better, because many programmers don't remove all memory leaks (personally, i remove all memory leaks, even smallest, and i deallocate all remaining memory at application exit, even if it is not necessary, because virtual space is automatically deallocated, but all programmers are not like me).
Moreover many programmers write buggy applications and sell them.
Of course, with .NET, there will remain buggy applications, but with less bugs, and no memory leak.
And, yes, the fact that bad coded applications on .NET are slow, but runs, is really a problem, because many bad programmers will not notice that their design is bad, and sell slow applications that need the last processor on the market to correctly run.
But is .NET bad for this reason?
Good programmers can use it, and not doing so will not resolve the problem!
Moreover, i think that programming in C#, for example, is better that programming in VB (VB promotes static structures and non-OO projects).
VB has already introduced bad programmers in the programmer world, and new bad programmers will directly use C#, or VB.NET, and that is better than old VB.
What I meant was the programming community is all getting the same reputation.Quote:
OUR clients ? How is that THEY mess up industry, and yet pursuade OUR clients ? Aren't those clients THEIR clients ? I think you ought to worry about YOUR clients, with your own bugs, and let THEM deal with THEIR clients.
Firstly, VB6 programmers aren't inherently bad. I've seen some really good programmers who use VB6.
However I've seen a lot of bad ones.
Again, the C++ community have bad programmers too. However since C++ is harder to learn and use than VB6 I like to think that we have fewer bad ones.
Unfortunately VB6 is used a great deal in commercial applications directed at PCs. Therefore we are all "getting tarred with the same brush" i.e. the whole industry's reputation is affected by all the bad code being produced.
.NET doesn't solve this issue, and neither does Java. I don't really think that there's a 'holy grail' out there. The whole industry is in somewhat of a mess.
I'm just doing my best to prove that it is possible to produce high quality software rapidly : and I think .NET is a better tool to do this than C++.
I personally NEVER have memory leaks in code. A memory leak in my mind is a show-stopping bug and should be fixed immediately.
However from my experience most companies ignore memory leaks and 'blame the operating system' when their applications fall over.
So I go back to my previous statement that we should have standards of practice the same as all the other professions have.
Darwen.
Isn't that what m$ kinda sorta did with the whole "xp certification" thing (whatever it's called), which hardly anyone seems to be doing anyway?Quote:
Originally Posted by darwen
No. It's not.
With Doctors, Lawyers, Surveyors, Nurses and every other profession governments have set up bodies to regulate their industries and prescribe exams for people wanting to go into their professions.
The exams are set by an external body which decides the standards which must be met by all people.
Microsoft is a company. Besides the fact that to attain a Microsoft qualification (apart from MVP) costs a great deal of money.
I'm suggesting that we should have bodies such as the "Royal College of Architects" or "The Royal College of Law" or the equivalent in every country in the world. I would like the governments of the world to see that programming is a profession - and should be treated as such.
If that happened we won't have companies employing large numbers of people producing rubbish code, and charging ridiculous amounts of money for what will in the end be a bug-ridden solution.
Darwen.
Well, not sure how things go where you are, but in the USA, the government would truly screw things up for everybody if they got involved with the formation of any required standards for programmers. It would cost huge amounts of money, and provide no value in return. In short; a bad idea. That's my take on it.
Some questions for you darwen:
In your example of a program written with .net, utilizing a C++ dll and so forth, It would be interesting to know the following:
1) Does any part of your code access win32?
2) If (1) is yes, can you do without, and at what cost in speed, and how practical would it be to do?
3) What impact on speed would there be if all of it was written in .net, and can you still avoid using win32?
It is in fact written 99.9% in .NET. I have machine language for some realtime bitmap manipulation as well as using DirectShow (with an interface written in .NET, no C++ involved).
I couldn't see any real speed improvement by using Win32 : in fact some of the things I'm doing which involve events, reflection and GDI+ would be considerably more complex to do in C++.
Darwen.
but is it not supossed that every .NET app acces the winapi? or it's not the same saying win32 and winapi?Quote:
Originally Posted by WizBang
But GDI+ is still part of the current API, just a newer part presumably because the functions couldn't be added to GDI without messing it up.Quote:
Originally Posted by darwen
Well, yes. Since the winAPI as of now is win32. This is sorta the point of the thread, which is if .net will somehow replace win32, right? So, my questions are aimed at the idea that if the application were to be run on a win32-less, .net version of windows, would it still work? I find it extremely hard to imagine phasing out win32 unless .net somehow eventually becomes faster. Even then it will take a long time. What I think is more likely, is that .net and/or longhorn will "pretend" that win32 is being used when it isn't. That is, when an app tries to call a win32 function, the system will re-route it to whatever the replacement is. This will inevitably, or even purposly result in a loss of speed, and the users will cry for a new version. Then the author will be forced to make the updates, or loose customers. Basically, will it not have to benefit the programmer to avoid win32 in longhorn? Lets face it, if calling a win32 function in longhorn still achives a speed increase, we're gonna keep using it.Quote:
Originally Posted by BlackWindzx
Maybe I'm working on the first commerically saleable application in .NET.
So far I'm really impressed by the speed and inherent object orientation of .NET.
I just made a leap forwards towards deployment today (i.e. doing things speedily, because I've designed the application well) - and again .NET has proven its worth.
I really can't say much more than this. WizBang - why don't you just try C# ? I think it'll surprise you.
Anyway that's it. I think I've said all I'm going to say on the subject. You can take my posts as either a Microsoft advertisement or as they are intended as a long-term C++ programmer thinking that it's a very useful technology.
We can fight about this forever. But I'm going to leave it now.
Darwen.
Never heard of such incidents. Nowadays whoever manages to ship a medium sized product, implies understanding that the platform doesnt just throw out bugs, even though their app is 'perfect'. I remember such things in days of Windows 95 and 98, but with NT 5.0 core, i never heard of people ignoring their own memory leaks. Besides, whoevers fault it is, usually the vendor of the product fixes the product either way. If WinSock 2.0 is not stable, use WinSock 1.1 - simple as that and wait until the ws232.dll will be stable.Quote:
Originally Posted by darwen
I wouldnt assume this for every and each programmer. For me, programming is a form of art, and standards do not apply to me automatically, unless they are good and useful. I have yet to see a standard like this. Even ID softwares code, whom many consider really professional coders, looks crap, with all the macro bullshit and C style chaos. .NET is indeed a step forward from all that stuff.Quote:
So I go back to my previous statement that we should have standards of practice the same as all the other professions have.
I agree with you on everything else you said though.
Here is a thread which deals with more of what m$ has in store for us and .net: http://www.codeguru.com/forum/showthread.php?t=338407
Well, I think thats an interesting concept, but what advantage would subscription applications have over desktop applications?
No actual advantage from my point of view, but .net developers will cry for some way to sell their apps and not have them stolen, which they can't do if distributed.Quote:
Originally Posted by Ben_Phillips