"Framework Heck" Interpretations and Community Responsibility

Frans seems to be getting awfully riled up over the backlash from his post. I can say this because I had this lesson taught to me as well: As a developer blogging in this community, you have a responsibility to clarify what is opinion and what is fact. What Frans effectively did is back up PeopleSoft's horrendously ridiculous comment that .NET is IT asbestos. By saying "ISV's: do NOT upgrade to Visual Studio.NET 2003! " you are sending not only a bad message to the community, but a message of opinion masked as fact. If versioning issues are your only reason, (pardon my bluntness) sorry but IMO, your reason is not good enough to justify such a harsh stance. The reason I can say anything about this is because I have been guilty of this before, and have since had to change my thinking.

Besides all of this, I have said several times that .NET 1.1 is not an upgrade, it's a sidegrade. VB6 was not an upgrade to VB5... it ran totally separate. Could you build VB5 apps in VB6? Could you build VB6 apps in VB5? So why does anyone expect to be able to build VB 7.1 apps in VB 7.0 or vice versa?

Now, to address the issue of Framework Heck (not quite as bad as hell because everyone totally blows it out of proportion), there has to be a way to intercept the way the Framework loads the assembly and respond by redirecting it programmatically, WITHOUT a *.config file. There has to be a method to redistribute this solution in a format that others can use,  and there HAS to be a way to handle it in future versions. So far I'm the only one that seems to be attempting to come up with a solution. Now, I'm missing a piece to the puzzle here... anyone wanna help pick up where I left off? Most of the logic is there already, it's in the right place. This routine needs to take place when the main assembly is loaded, which may require a patch to the Framework itself.

In the meantime, I'm writing yet another article about tactics for reducing the headaches of dealing with two versions of the Framework. I will make an anouncement when this information is publically available.

3 Comments

  • [quote]


    Besides all of this, I have said several times that .NET 1.1 is not an upgrade, it's a sidegrade. VB6 was not an upgrade to VB5... it ran totally separate. Could you build VB5 apps in VB6? Could you build VB6 apps in VB5? So WHY IN THE BLOODY HELL do you expect to be able to build VB 7.1 apps in VB 7.0 or vice versa?


    [/quote]


    Why in the bloody hell? Well, I'll tell you: BECAUSE I AM NOT IN THE POSITION TO TELL MY CUSTOMERS TO BUY YET ANOTHER IDE FROM MICROSOFT WHICH COSTS MONEY.





    So there. In VB5 you still can use dlls written in VB6 (hell, even VC++ or Delphi). In VS.NET 2002 I can't use assemblies written in VS.NET 2003 unless I add a long list of redirects to a config file. And because I'm on VS.NET 2002, I do not think about that, because I never had to and MS is NOT TELLING ME TO ADD THEM TOO. Only some dork from The Netherlands is, who you flamed out of the water with this posting.





    I changed the title of my article and added a note. Then I read this posting. Thanks a lot, Robert.

  • I totally agree with you...

  • I've posted a brief note on my blog that explains what most developers need to know about this. I don't think it really applies to the case with the Oracle driver, but it certainly covers the general case the Frans was talking about.

Comments have been disabled for this content.