Re: Down to serious business -- possible fugitive from vb6

Bob Dawson wrote:

I'm inclined to believe that as a for-profit organization, CG will attempt
to fix the bugs with the greatest ROI for the effort, basically a triage
based on several factors, among them
--how bad is the bug?
--how many customers does it effect?
--how much effort will it take to fix it?
--what negative side effects would fixing it have? (ie, would correcting
the problem at this point break currently working user code?)

bobD



Based on the past actions of Borland, many of whom are now with CodeGear, I don't believe that the correct decisions will necessarily be made. In many cases, I think that data has been skewed to represent the views that management WANTED to be shown (i.e., survey questions designed to elicit specific responses, etc.).

In a perfect world, the above would be sufficient, both for CG and its customers, but just take a look at your above questions...

'how bad is the bug?' - IMHO, Borland has not had a good track record of evaluating that question. CG seems to be doing better, but there are still obvious examples in QC where 'bad bugs' still exist.

'how many customers does it effect?' - just as many folks from Borland/Codegear claim that us end-users can't see the 'reality' that they see, I would submit that CodeGear doesn't see the reality that we see. They're just taking a SWAG at this number, and human nature dictates that they will evaluate/choose a value that is most beneficial to them. In other words, it's easy to look at a 'difficult' problem and to shrug it off as 'well, it doesn't affect too many folks...'

'how much effort will it take to fix it?' - this one is easy... Much more than any original estimate <G>. The bottom line is that if they don't budget a certain amount of resource dollars towards maintenance, then sufficient maintenance won't occur -- it's that simple. I understand that allocating such dollars will inevitably lead SOME problems to fall through the cracks, but (again, IMHO) Borland has historically allocated too small a dollar amount for maintenance. Happily, CodeGear is doing MUCH better in this regard, but it remains to be seen as to whether they'll stick to this path...

'what negative side effects would fixing it have?' -- the problem is that both Borland and CodeGear use this excuse far more than addressing the OTHER problem ('what negative side effects does NOT fixing the bug produce'). The second part of this equation seems to inevitably get ignored. Further, as is evidenced by many of the regression bugs experienced with CodeGear, and some of the statements made by QA folks, it's only 'recently' that they've become serious about testing these things -- so they've really only got themselves to blame.

David Erbas-White
.

Relevant Pages

  • Re: Delphi QC Top Voted
    ... That priorization policy sounds logical for a developer who release a ... He could fix some now and the rest in a later SP. ... Borland released Delphi a year and a half ago, and no say that they will fix ... Take for example D7 IDE bugs. ...
    (borland.public.delphi.non-technical)
  • Re: Marcos latest Blog
    ... I know, they can't fix every bug, but the important ones, they need too. ... If I take the time to track down a big bug, Borland will actually fix it ... Just to give you an example, I have complained repeatedly about dbExpress drivers meta data bugs. ... Delphi would roll up their sleeves and fix the problems. ...
    (borland.public.delphi.non-technical)
  • Re: Proposal: Delphi VCL+RTL as open source
    ... -Delphi volunteers and freaks can fix and test much more bugs than Borland. ... Anyone can modify it and submit bugs and fixes to QC regarding it. ... Improvements made by external Borland developers (like Fast-code ...
    (borland.public.delphi.non-technical)
  • Re: Give me reasons not to use Delphi
    ... I keep hearing this argument that Borland doesn't fix their bugs. ... fix if you have a solid report with a good description and reliable ...
    (borland.public.delphi.non-technical)
  • Re: Another take on Embarcadero and CodeGear
    ... CodeGear to say "do you want these bugs fixing, ... CodeGear will have to do both. ... they can't fix ALL bugs, ... affected play a role, as well as the effort required to fix it, and the ...
    (borland.public.delphi.non-technical)