Scott Swigart said:
I recently interviewed Jay Roxe of Microsoft regarding VB6 to VB.NET
migration. I think there's great stuff in this interview, and Dr. Dobbs has
been gracious enough to publish it.
http://www.ddj.com/documents/s=9776/ddj1129914421813/swigart2.htm
And Dr. Dobbs is gracious enough to allow public viewers to read the entire
article, at least for the time being. But look at some of the contents:
JR: They motivations for doing an upgrade are pretty varied. Some of the
motivations just relate to the lifecycle of any application. The
application has reached a point where the code has grown beyond its
original scope, and it's time to rewrite the application anyways, and it
just makes sense to transition to the latest development platform at the
same time.
True. Now what about applications which haven't reached that point yet,
where it isn't time to rewrite the applications anyways? Or even what about
cases where it _is_ time to rewrite the application, but rewriting from
scratch might take almost half as long as trying to do a migration, and a
company which can't afford the development effort to do a migration also
might not be able to spend half that much to do a complete rewrite instead?
JR: [...] What happened on March 31st of this year (2005), is that the VB6
design environment moved from mainstream support to extended support. It's
a seven-year old product that's had six service packs. It's pretty stable.
Oooh. NT4 had six plus A service packs in how many years, also proving
stability. VB5, VS2002, VS2003, and Windows 98 all had fewer service packs,
is that because they're more stable or less stable? And the known bugs
remaining in VB6 SP6 are going to be stable, perpetuated forever, right? We
just shouldn't confuse that kind of stability with a different kind of
stability, one which judges whether the products operate reliably. Oops, or
is it that we _are_ supposed to confuse those two kinds of stability?
It also means that if you have the two warranty cards that came in the box
with VB6, those warranty cards have expired, seven years later.
April 1, 2005 (first day after what happened)
minus
April 1, 2002 (VB6 still on sale)
equals
seven years?
Microsoft, which version of Visual Basic did that calculation for you?
And even though VB2002 is so stable that it needed only one service pack,
it's also so stable that it only needed one bugfix pack called VB2003. So
in that subtraction I just asked about, maybe the second operand should be
2003 instead of 2002.
And VB2003 is so stable that it needs no service packs. I do admit that the
IDE only crashed one time when I typed an apostrophe in expectation of
adding a comment to one line of code. And I do admit that I didn't find as
many bugs in VB2003 as I did in VC++2003, maybe because there haven't been
many calls to use VB2003, I wonder why that could be.
Existing VB6 code still has no migration path. Existing VB6 code still
needs a compiler and library, and the compiler and library still need bug
fixing.
