Title: .NET: Another point of view
Question: What about .NET?
Answer:
I have seen the article of Romeo Lefter where he gives his opinion about the .NET framework. Since my view is rather different I decided to write an article about this too.
Before I start I just want to explain what I my interest. I am not a Microsoft fan. When Delphi appeared I immediately dropped using of Visual Basic. I never liked MFC although I have programmed in it. I was and am a Borland fan. I have experience with Java so I am not blind about other technologies. I am stating that just to prove I am not one-sided.
First of all I will speak about the IDE. It is not important for the clients but for us as software developers it is very important. What about it? It is perfect. Yes, the Delphi IDE is good. It has many great features - Borland is making very good IDEs for more then decade. But you just have to see it. It not only looks good but working with it is really comfortable. Not to mention that Microsoft found a better way for the non visual components not on the form like in Delphi, but on a separate area of the IDE. There is much I can write about the IDE and only good stuff.
Now lets talk about the languages They are interchangeable. You can write in C#, VB.NET and many of the others languages which will be available for the .NET platform. But I dont consider this as important at all. C# will rule here since it is especially designed for this framework. And you cant just take a VB program and start it in .NET. You have to change much. You have to teach all the VB programmers stuff like classes, inheritance and other conceptions new for them. VB was widely spread only because of its simplicity. But this spread it has already may enable it to remain important in the .NET framework. But if you want to program for the .NET with Microsoft tools C# is the obvious choice. Actually the languages are not very different because they all support the same framework and they have the same classes. In fact Microsoft noticed this problem and is going to make the languages more different in future.
The framework: .NET is nothing revolutionary despite the Microsofts will to prove the opposite. It is a way to tell Sun The IT industry is ours! It has the interpreted byte code and the just-in-time compiler, it has the garbage collector it has all the Java stuff. Well, Microsoft calls them different but this does not change the facts. But regardless of the statements by some that the interpreted code is bad, Java already proved that it is a very good choice. So this is not the weak branch in their strategy. Whether or not the .NET framework spread fast I think this will not be a problem for them. They have conquered the browser field and this will not be a problem for them. You just have to download 20MB from internet and install it. With these fast connections nowadays it will stop no one. They will redistribute it with their new OSes (Windows XP has not the framework). They can redistribute it with Internet Explorer, they can even make it part of Windows Update Having already written this I found that they ACTUALLY did this a few days ago. The .NET framework is now a recommended updates. And this problem is really a problem only if you develop Windows client applications. Delphi is still the best choice here.
The .NET apps will remain slower than the compiled ones. But let me explain here more. Now the processor power is too much. Believe me. I can run every client applications on the cheapest CPU available on the market and they will run well. Microsoft Office often accused to be a bloat ware run nice on 300 MHz CPU with 64MB RAM. And it is impossible to find such machine these days except as second-hand. So an office product like Word or Excel can run very well on a modern PC. Keep in mind that if a .NET application is two times slower it will still run at a decent speed.
I am amazed how anyone can consider the decompilation as a serious threat. There are many Java decompilers and this has never been a problem for the Java developers. Very rarely you put something secret. Even a big product like MS Word is built by well known pieces and one very good developer can do it by himself if he has the time to do it a few centuries may be. The real value in a software product is the combination of good design, written code and debugged working source code. And if you are developing components you are giving the source code anyway. And if you are still a bit paranoiac you can use the so called obfuscators so even the decompiled source code will remain pretty cryptic.
The future of .NET is very, very unpredictable. Everything looks so nice not typical for Microsoft product. They opened the source of their C# compiler under a shared source license.
All of you know of Borland and its .NET strategy. They will provide product which will produce applications compatible with the .NET platform. If it is based on CLX it will be great because the same app will be able to compile as native Windows code, native Linux code and .NET CLI.
There are other promising projects. Such as dotGNU and iNET. The first one is an attempt to build a .NET framework for Linux and other OSes and the second one is the .NET framework implemented in Java (both of the projects are still in progress). If such project succeeds, we are all going to benefit from this. Everyone except Microsoft.
But what will happen is not possible to be predicted. As Delphi developers we may stay with Borland because yet again they will be offering us an alternative to Microsoft.