I just had some stuff occur to me, partly prompted by the fact I'm
working simultaneously in RB 5.5.5, Visual Studio 2003 and Object
Master (with CodeWarrior Pro5).
1) The scary data-loss problems at the moment are most likely either
framework bugs or design/coding bugs in the IDE. It is particularly
significant that some major new bugs (the saving and menu stuff) seem
to be introduced by successive versions.
2) The 2005 IDE was written by RB programmers using RB5.5.9 (old IDE,
new compiler & frameworks) - it proves nothing for or against RB2006
as a development environment. It will take someone writing something
of similar complexity from scratch in RB2006 to prove if the new IDE
is good for developing complex programs. The ongoing support of RB
using the new IDE doesn't prove anything in greenfield development -
maintenance is not the same as writing from scratch.
3) Whilst I believe RB is a better language in some ways than C++,
especially for working with frameworks, it is not magical and hasn't
exhibited any inherent virtue in this example of making it
significantly easier to produce a robust, complex GUI app.
What I think would be interesting and rather reassuring would be a
public post-mortem by RS on the IDE project to-date, especially some
of these recent bugs, focussing on:
i) the development practices that led to these bugs, if any in
particular can be identified
ii) any practices or features of RB that particularly helped in
identifying the bugs
iii) changes in practices they are initiating to improve the
software quality of future versions
iv) ideas for how the IDE (IF they have started using RB2006 to debug
the IDE) can be improved to help mitigate such problems in future.
Andy
_______________________________________________
Unsubscribe or switch delivery mode:
<http://www.realsoftware.com/support/listmanager/>
Search the archives of this list here:
<http://support.realsoftware.com/listarchives/lists.html>
