Board index » cppbuilder » Re: BCB2007?

2007-03-22 11:53:26 AM
cppbuilder99
"David Dean [CodeGear]" < XXXX@XXXXX.COM >wrote in message
The last word I heard was that C++ doesn't even have a Project Manager
yet...has that changed?
 

David Erbas-White wrote:
the best efforts of large numbers lower down. Yes, a shift in top management can make
a huge difference in what goes on.
The change to CodeGear might well make a big difference. I do not know that for a
certainty. But I have seen up close and personal (recently even) how a changing of
the guard can make a big difference. I've also seen how when the transition comes the
customers do not immediately see the benefit because, well, the ship has lots of
leaks and all that.
the day they retire. Then - unless they go to work for a defense contractor - they
suddenly start speaking their real thoughts.
My guess is their bug fixing efforts have sped up a lot. But I do not know how big
some of the bugs are or just how long it'll take to make the stuff heavily mature.
Again, I draw from personal experience: I fix some bugs and then get to hit the ones
that lurk behind the ones I fixed.
You have plenty of reasons to be dissatisfied by past performance and past promises.
I'm not even saying you should expect a huge improvement in the next release. I'm
just saying that lots of trends continue on a path until suddenly the trend stops and
a different path shows a different trend. Maybe this has happened at CodeGear. Maybe
not. We'll find out on the next release.

what solved the help issue. The new tools creates nicer, more useful
content - but the IDE had to be gutted to redo the help system. The
difference you saw in help access and accuracy from the IDE from BCB6 to
BCB2006 was introduced in Delphi 8 - because MS introduced a completely new
API help system, around 2003 we merged our old system and the new MS help
system together to create a hybrid help system. This worked, but we lost a
lot of IDE to Help content lookup accuracy. We worked to improve the hybrid
system over three releases and finally after 2006 we decided that to make a
major improvement in accuracy we had to do major surgery and redo the IDE's
help integration. This required that we work on the new branch - otherwise
we could destabilize a stable 2006. Why? The IDE to Help integration touches
just about every part of the IDE. If you muck with it - it has to be right.
We only take very very measured risks when we patch and hotfix - we do not
take unnecessary risks that could cause destablization and hurt other areas
in a released version that affects hundreds of thousands of users. This
level of change is a deep IDE level project and can really only be done in a
full release cycle and on a new IDE branch. It's not a "patchable" change.
So we could provide improved help content, that is much nicer - but patching
in the new "system" is not practical. I fully expect to get flamed again for
the explanation, but we're telling you like it is. You can take a cynical
view of it and accuse us of being complacent or ignoring 2006, or "choosing"
not to replace the old help system with the new one, but it's just not true,
we've been more responsive to 2006 than any release in recent memory. We're
very happy with the new system and want everyone to have it. If we could
retrofit the changes into 2006 we absolutely would have done it and it would
have been in a hotfix. Looking forward, the next C++ release will be based
on the exact same branch as Delphi 2007 - and will inherit the same help
system and improved content.
I'll absolutely continue to listen to your feedback. Whether I agree or
disagree with your assessment - I care about your experience with the IDEs
and also what you want to see in the future. -Michael
 

{smallsort}

Michael Swindell (CodeGear) wrote:
it. I will not flame you for doing your job and explaining the
reasons for not able to release patches.
I don't know who Borland got to do their QA for BDS2006, but apparently
the testers were very proficient in Delphi, enough not to really bother
about the help. I know I shouldn't be so facetious, but I don't under-
stand how anyone could've missed the help system. I really don't.
What's done is done. I've purchased a system that's not that helpful,
but useful to a certain extent.
I hope that D2007 (when and if I order it) will be as 'useful' and
'helpful' than its predecessors. There's only so many times I get
burnt before I turn away. As the saying goes, Fool me once, shame
on you. Fool me twice. Shame on me.
In any event, I wish the best of luck to CG. Considering the
baggage its carrying and the stigmata that is Borland's handling
of the IDEs, I hope CG does a better job. I mean, you can't do
any worst than what Borland did. (I know. I'm tempting fate.)
Edmund
 

video.google.com/videoplay
 

Dennis Jones wrote:
groups. This is a private server and operates under somewhat different
rules, one of which is to trim quotes to a minimum and another is not
to flame anyone. While the first is not a major offense, many here
find it irritating. The second can get your access to the groups
denied if it goes too far.
Personally I find it annoying if a user quotes so much that all I see
without scrolling is the quote. And inexcusable if I then scroll only
to find something like "Yes" or "I agree" at the bottom.
- Leo
 

"Leo Siefert" < XXXX@XXXXX.COM >wrote in message
- Dennis
 

"David Perkins" < XXXX@XXXXX.COM >wrote in message
fool me twice, shame on me"! But apparently I was wrong!
- Dennis
 

Ed wrote:
It is a product they are actively selling today. In my mind it is their
flagship product (though I suspect that most of them now think of the
single language D2007 as the flagship product). Yet they are already
treating it as a legacy product with no need for maintenance.
I would greatly appreciate seeing some of the bugs I run into fixed. I
would like to see Codegear walk the walk, instead of just talking the talk.
products.
Companies earn reputations for doing a good job of maintaining their
products, or not doing a good job. Customers are happier to pay for a
product they believe will be maintained, and will pay more for it. They
are rightly more hesitant to commit to purchasing a product that they
believe won't be maintained. Furthermore, they will not value it as
much, and hence are not willing to pay as much for it if they do buy it.
I think Borland/DTG/CodeGear has earned a reputation for average quality
and poor maintenance in the wider developer community (including those
who are not customers or users) through its last few releases over the
last few years. They used to have a reputation for good quality and
great support. D8, D2005, CBX, Kylix, and BDS2006 have been {*word*156} its
reputation. So has the existence of long standing bugs that have
survived multiple releases, and the introduction of new bug ridden
features.
It takes long term thinking to appreciate the increased revenue that
results from good product support and maintenance. I'm not sure I'm
seeing much of that from CodeGear.
Dennis Cote
 

"Dennis Cote" < XXXX@XXXXX.COM >wrote in message
mentioning it again.
As a case in point, look at AutomatedQA and the patch releases that are
available for AQTime. There are recent patches available for *old* products
(apparently *ridiculously old* from a Borland perspective). For example,
AQTime 4 was released sometime in 2004. The latest patch was December 2006,
more than two years after its release, and more than a year after the
release of their next version (5.0)!. The same with AQTime 3, which was
released in late 2002 -- the latest patch was available in September of
2004, more than two years after its release, and 6 months after version 4.0
was released. Now *that's* what I call product support!
Borland's habit, on the other hand, has been to release zero-to-few patches,
to a point, but never beyond (or even close to) the release date of their
next version (that I know of). In effect, forcing users to pay for an
upgrade they may or may not need or want, and even then, getting fixes for
the bugs you hate the most is a {*word*99} shoot -- maybe you get it, maybe you
don't.
- Dennis
 

"Duane Hebert" < XXXX@XXXXX.COM >wrote in message
ever fully recover. Many users have already cut their losses and moved on.
I've been using Borland products since the days of TP 1.0 for CP/M (so I'm
definitely an old-timer -- er, loyal user) and I will probably continue to
use their products for the forseeable future, but only because I am highly
vested in VCL and because VCL is still better than MFC and the
still-maturing .NET. But if .NET continues to mature and improve (and
especially if native Windows API coding becomes obsolete), there will be
fewer reasons to stick with Borland/CodeGear, and loyalty may not be enough
anymore.
- Dennis
 

that a lot of sales for each version were from their existing
customer base. They possibly generated more revenue
but not fixing bugs.
This only works for a while though. We basically went through
BCB4->BCB5->BCB6 in an effort to get bugs fixed. When
they went to CBX we went to MS/Qt since there was no chance
of getting BCB6's bugs fixed. There's little chance that
my current company will buy Borland products again.
 

I want to say <g>
As to VCL being better than MFC, there's no question but VCL isn't the
only choice (VS/Qt works for us). At any rate, it doesn't matter much
what I would prefer. My company will not go for it. It's mostly due to the
lack of bug fixes without paying for upgrades. Maybe that's changing.
CG seems to be trying to change things but that won't be enough to
recapture all of the lost clients. We've switched to MS/Qt and are very
productive. Management has no incentive to go back to Borland at
this point.
 

In article < XXXX@XXXXX.COM >,
Dennis Cote < XXXX@XXXXX.COM >wrote:
tell you what we've been working on, but I am not allowed to do so.
continue to push for more patches.
--
-David Dean
CodeGear C++ QA Engineer
<blogs.codegear.com/ddean/>
 

In article <4602b7f3$ XXXX@XXXXX.COM >,
"Dennis Jones" < XXXX@XXXXX.COM >wrote:
(Just a start, but that should demonstrate that we *do* intend to act
differently than Borland)
--
-David Dean
CodeGear C++ QA Engineer
<blogs.codegear.com/ddean/>