This blog has moved to ericnelson.wordpress.com

We have not had good experiences moving from VB6 to VB.net; much better was the move from VB6 to c#; however, some "developers" must move back into Office/VBA world. Or stick with VB6, which explains why that route will hang on. Read on...

There is no point in pretending that VB6-style code has any place in .Net. The "VB6 developers" that were also pure C/C++, also wrote COM objects via ATL, and WinUI(C/C++) via MFC (as well as C/C++ calling Win32 API directly) prefer and are more successful in C# once they move to the .Net world. The "vb6 developers" that were *only* VB6 developers (with a smattering of Office VBA and maybe some sharepoint) find they are fish out of water in .Net and quickly either give up or produce VB.net code that has no chance of succeeding, worse than VB6.

On the other hand, we have had *some* success building libraries and business frameworks in C#, and allowing "VB6 Developers" to develop an application using these libraries in VB.Net.

VB is a fine CLR language; does not fit in too well with rest of the development picture (Java, JavaScript, C/C++), but that is OK. I have no problem with VB.net,

There are many excellent VB6 developers; we have found that they also work fine in c#. What we are losing are the lower quality developers that made the jump from Office/VBA to VB6, but cannot go any further. What can we do with them?

Please don't confuse this post with "vb bashing" or other class warefare. We need most developers to be fluent in a variety of tools on a variety of platforms - An engineer who "knows c#" is no better than one that "knows vb.net" or "vb6". However, we also need a pack of "second tier" developers that can provide less sophsticated software solutions for less money, or implement the designs of the core software engineering team. Microsoft has no platform for them, now, because VBA inside office is not enough and .Net is too much. It would have been better if MS continued VB6, and left VB.net out of the picture. In my opinion.

Maybe if there were more stability with the language and tools of .Net it might look more appealing. I've stepped up to .net 2 and VS2008, but can't be fussed with the new features of v3, nevermind v4!
Wold be nice of the development train slowed down a little, give people time to adapt and develop.

I like your approach of VB6 Myths. I've been writing a blog called The Myth of Visual Basic Migration / Upgrade (http://blogs.artinsoft.net/fzoufaly/ ) precisely trying to eliminate any fears in this direction.

In my opinion the first wave is a migration of skills, which is well underway, and the second wave is the migration of VB6 code. Even if we've participated in many many projects over the years, it is clear that the second wave is accelerating and now that companies are confortable with the .net platform they are moving their code at an increasing rate.

From the perspective of someone who has only used VB6 rarely and with reservation and wasn't really interested in developing on Windows until C# came along I have mixed feelings about Microsoft abandoning VB6.

When I was a kid it was being able to write fairly complex GUI applications in Apple's Hypertalk that got me in to programming in the first place (that and Basic on the Intellivision). Wanting to extend the functionality I ended up moving on to Metrowerks CodeWarrior (C/C++) and Visual Cafe (Java).

Today, C# is the best fit for most applications for me (certainly on Linux and Windows and with porting to the Mac in mind), but it's not what I would have been looking for when I was starting out. For the guy who isn't a really a programmer but still wants to be able to knock up an application that can help him (as is the case for many VB6 developers) neither C# or VB.NET is what they are looking for - some might feel ready to move to VB.NET after years of VB6, but certainly not all and of course the next generation might not want to start out on .NET.

Even though I don't really like the idea of hordes of amateur developers - with little to no knowledge of interface design or appreciation for good software architecture - writing what often turn out to be large, unmaintainable hydras of broken code, often used as sticking plaster for businesses that may be trying to patch holes in the wrong places I recognise there is an unmet need, a void currently filled by VB6.

While I can understand the affection many feel for VB6 (because it lets people "get their stuff done" and I appreciate the need for that) I don't think fondly of it. I think VB6 sits uncomfortably between something like Hypertalk and VB.NET and that there is a place for a good scripting language in the world for RAD but that we can do much better than VB6.

I would like to see an open specification (ala C#) scripting language that very easy to use but that encouragings best practice and introduces OO concepts by stealth - with an IDE that, while making RAD easy and painless, even for inexperienced and novice programmers goes some way to guide them in the right direction when it comes to structuring their code, unobtrusively and without them even realizing it.

While I don't quite agree with Jo that Microsoft should have kept VB6 alive, I do agree with his point and think that the "second tier" developers are not quite catered for by .NET (and not even by VB.NET) and that they need something that has access to much the power of the API's in .NET but the syntax of a very high level ('goal orientated') language, rather than passible-but-not-quite-satisfying middle ground VB6 inhabits.

I can understand some people sticking with .NET 2.0 over 3.x for now (our own analysis showed that ~80% of site visitors using MSIE already had .NET 2.0 installed according to their user agents) but I don't know of any compelling reason to still be deploying .NET 1.1 applications (unless they have truly monstrous .NET 1.1 apps that are hard to port - but that seems unlikely).

80% of clients having dotnet v2.0 = F A I L if you want to distribute an app beyond the in-house/server.

Get over to c++ and stop telling people to reboot 3 times then buy a new system.