Thread: .NET woes

>I hope I'm wrong.
I think you are. Open source projects have a habit of sprouting where they're most needed. If C++/CLI grows to the point where people badly want those tools but don't want to pay for them, you'll see the free stuff showing up more and more.

>Oh... and say goodbye to portability.
In a way, portability is a myth.

>It's the possibility that they may push C++/CLI as the replacement for ISO
>C++ programming that scares me.
The chance of that is vanishingly small. I can't begin to fathom why you would think that Microsoft's version of C++, that breaks every guideline for the ISO standard, could possibly replace it.

>believing that the only option to be marketable in the job industry is to learn C++/CLI, seems
>to me rather naive.
He never said it was the only option, just the most popular. If you remove the most popular options, getting a job is considerably more difficult, and failing to recognize that is naive.

Originally Posted by Prelude

The chance of that is vanishingly small. I can't begin to fathom why you would think that Microsoft's version of C++, that breaks every guideline for the ISO standard, could possibly replace it.

Because in the future, interesting new features of Windows might only be accessible through a CLI API that has no native equivalent.

>Because in the future, interesting new features of Windows might only be
>accessible through a CLI API that has no native equivalent.
Are you suggesting that the standards committee will feel obligated to support a proprietary API as an exception? Given past decisions by the committe about features and additions, I predict that the C++ standard will simply ignore these "interesting new features" and let implementations deal with them as non-portable extensions.

Originally Posted by Mario F.

Oh... and say goodbye to portability.

This is probably my favorite "I won't use .Net" myth to debunk. Why should I say goodbye to portability? The Mono project is in the process of implementing the CLR and the standard class libraries on a number of different OS's. If an OS has those two components, then it can run any CLS-compliant .Net application. Then the question becomes "Is a CLS-compliant application capable of doing anything useful?" Everything developed by the company I work for is CLS-compliant with the sole exception of registry operations (which could be handled by a config file) so I'd say so.

Hmm... let me make this clear. I'm not going to use .Net because... you got it, i don't want to.
Not because I think it's better, worse, more colorful, dull, cute or ugly.

If you prefer, I'm not going to use .Net because it is a personal choice. Just like, for instance, I don't feel interested in learning Delphi. I just... don't want to.

As for Mono... it has been around since the beta versions of .Net. Or close, I can't remember. It will always be a step behind Microsoft implementation of .Net. If I remember correctly, it took Mono 5 months to keep up with the .Net Framework 2.0 and it is still struggling to. And this is just for ASP.Net.

It's a viable solution, sure. But one that comes at the cost of lenghty compatibility updates. And I'm sure Microsoft will not be waiting for Mono to come up with new versions... nor it will think of Mono when changing function locations and whatnot like it so much enjoys to do with the DirectX API.

>>Oh... and say goodbye to portability.

Why is portability so important?



Approx 90% of home PCs run windows so limited extra return for a portable app (compared to a MS only app).

The biggest expense (IMO) in development is time costs (wages). If an app can be developed in significantly less time using a MS product (or targeting only a MS OS), then BAs will push that type of development.

I have coded, in a commercial capacity, for very high volume online trading sites, embedded XP for industrial devices and desktop apps.

Never needed any of these projects to be portable, as the target OS was always specified in the business requirements or by the hardware used.

Nor has specializing in MS technologies limited the projects I can work on.

I do agree that portability is playing a lower role. That has to do though with a lot of factors. Lack of vision from customers/solution providers, lack of knowledge, commercial companies pushing their "way of life", developer lazyness...

However portability is important despite the number of systems running MS Windows. If for anything else because I'm sure nobody in here would want to see a Windows only world. Lack of portability only serves the commercial brands. When Microsoft started their hype with the .Net Beta version, they did it by exacerbating the portable nature of the framework. Back then, 5 years ago or something, Linux was still posing a serious threat to Microsoft. Now that the waters have calmed, portability is no longer an issue. In fact, it's an enemy of a company that wants "every computer in america to run microsoft windows" (sic).

But... just as an example. 3 years ago The Portuguese Ministry of Higher Education (concentrates on universities) wanted to migrate to Linux. The cost was huge. The whole Visual Basic base had to be rewriten, Microsoft SQL Server databases migrated and queries rewritten and many other factors. I don't know them all. I was simply in charge of evaluating the VB migration. The idea was to move to a more cost effective IT solution.

It had to be abandoned. The time needed for the migration and the costs involved with hiring coders were simply too overwhelming. And to this day, the Ministry still has to pay Microsoft for every single computer, every single server. Bulk licenses, of course.

Meanwhile, on a small room, just like in the movies, an old computer still ran some legacy C code. That computer migration cost was calculated in 160 Euros. The cost of buying a new HD and of the guy installing linux and compiling the C code.

I try and make most of my code portable not because linux > windows or anything like that (I do believe that is true in a lot of cases) but more so that down the road, when the leading OS has changed, the API for it has changed, and all the other changes, I will still be able to use the same programing habits. I seriously considered learning to use DirectX due to it being the API that Microsoft supports and 90% of the time (atleast) my code will be executed on a MSOS.

But being forced into .NET to use it, I would rather not. I am not a professional game developer or anything, and that isn't my goal, learning game programming is more for a greater understanding.

>>3 years ago The Portuguese Ministry of Higher Education (concentrates on universities) wanted to migrate to Linux. The cost was huge.

I helped port one of Australia's largest banks online share trading site from Oracle based system to a MS SQL Server + white label solution.

Cost was huge over the 18 months it took to complete.

A senior developer's time is charged (by the company) at nearly Au$170/h and I was only one of a large team (>30 coders and testers).


Six months later, a merger ment they had to port everything to JAVA ....
(one bank had a contract stating that JAVA would be used, the others did not == move to JAVA)

Moral: Choices in technologies used in major development are rarely based on the best fit for the project at hand and are usually mede by people who know very little about the merits of one over the other.

I don't understand why people took to .NET so fast, half the people in my programming classes love it, and I can't see why.

Because it's RAD without the nasty VB behind it. C# can create a productive commercial solution in probably half the time. I love C++, but I cannot justify charging double rates just because I like C++ better. C# gets the job done in 90% of business needs and in 50-80% of the time needed for a native C++ application. And don't anyone dare to tell me that a Win32 project or something involving DirectX, MFC, ATL, COM would be any more portable. The interface to our banking software a a windows DLL. The interface to our inhouse accounting software is an OCX. The interface to the post code labelling software is a windows DLL. Portability is for people who want to do it right and for people who have too much time on their hands. Businesses are neither. They care for ROI, right now.

>Am I to understand now that MS is forcing us to use .NET to do mail, internet, messaging, etc, etc?

But for other things, the Win API seems the only way to do things:

http://www.itwriting.com/blog/?postid=486

So, nvoigt, are you saying there is no market for C++? That it is just a toy for those who like it and it can only be justly applied on systems where a framework the likes of .Net was not created?

I formulate my opinions on this matter mostly on what I read here and there. However you guys use these tools at a professional level. I am curious as to if I'm learning a programmin language that, from what I'm reading here, is dying.

>Linux is not free.
You know perfectly well what I meant. Linux is free in that there's not necessarily a direct cost in legally acquiring it. Sure, some distros are sold retail, and there's certainly the indirect cost of net access for downloads and the time spent installing and learning the system. But with that line of thinking, we'll spend more time specifying our definition of the word "free" rather than talking about more interesting things.