SiegfriedN <
XXXX@XXXXX.COM >wrote:
Quote
Jeff Overcash (TeamB) wrote:
>
>Name the one open source project the complexity of Kylix that
>has a single expert covering all the areas I mentioned and I'll
>continue this conversation. It does NOT work in open source
>projects this way. Open source projects of this complexity
>have many experts covering a wide variety of areas, not one
>single expert covering them all.
There is no denying that Kylix/Delphi 7 is a very complex product.
Lazarus/FreePascal is an open indication and there a few people are
working on it. What was trying to be said was that '..if one person can
fix sooo many Kylix bugs in his spare time..' (-QA) what is delaying
Borland who potentially have enough resources and time (able to hire
resources) - (why don't they hire Andreas?) But then this is old news to
you and you are fighting the same battle as us :)
I don't believe Borland has abandoned Kylix. One CAN develop
applications in Kylix. (Though not so pretty with QT2.) I am porting one
of my D7 apps to Kylix at the moment in Kylix. (CLX only) You just need
a lot of patience. Andreas's work is an incredible help!
My point is that Andreas has done a great job on VCL issues,
but I highly doubt those with dbExpress problems would be happy
if patches (obviously non QA'd patches if it is a single person
only) didn't address their problems in database connectivity.
There is a C++/Delphi compiler bug that prevents me from
releasing an IBX for the C++ side of K3, I'm sure all the other
people having compiler issues won't be happy with patches only
to certain areas of Clx. What about IDE stability?
If Borland is going to do right by the Kylix customers they
have to dedicate more than a single developer. They have to
dedicate some QA resources + many other Dev resources. Trying
to do a single person type of patch that addresses only one
area of a product is like only treating the broken arm of a
person in a car accident and ignoring the broken legs,
internal injuries, head trama etc. You might save the arm, but
the patient still dies.
Even though there were 7+ individual updates to D7, people still
complain that certain areas were never touched (IDE, compiler
and de{*word*81}). Many of the same areas complained about for D7
are being ignored in this suggestion too. I don't see this
making many happy with the state of Kylix. The only thing that
would make them happy is Kylix being given full attention by
a team large enough to make a real difference - and that that
work go through a full QA process. If you don't need a full QA
of patches, then Andreas' current solution would be just fine,
but people seem to want more than just the unofficial patches.
Borland will not call a patch official if it does not go
through QA, so the only diffeence would be Borland releasing
an unofficial patch or Andreas doing so. In the end there would
be no real difference.
I'm not happy about the way Kylix has been handled, but I
recognize that Borland has to get a good ROI on their resource
investment to stay in business. Kylix (or maybe more appropriately
the Linux developer market) was not giving them a good enough ROI to continue, at least in 2004, to work on Kylix. They may be
making it impossible to ever revive Kylix in the future because
of their treatment today, but unless there is a sudden change
in the Linux market (which there really is not any sign of
happening soon) that's not necessarily a bad thing (from
Borland's business standpoint).
Also remember that from a Linux viewpoint Borland already has JBuilder there, they have CBX there (so while no real RAD yet, that has been promised) for the C++ line and if Mono takes off they will have Delphi for .NET there. So IOW, two years from
now they might have basically everything they provide in Kylix
in Linux anyways, but in a more resource friendly manner. For
each of those they almost get Linux resource free which greatly
increases the Linux lines' ROI without a needed increase in
Linux market itself that Kylix needs to increase its ROI.