Subject Re: [Firebird-general] Interbase vs Firebird..
Author steve summers
--- cowmix3 <mmarch@...> wrote:
> I was as some trade show this past day and a vendor there told me
> that
> Firebird has been horrible for reliability for his clients and that
> Interbase is much more stable. Can anyone spell out the differences
> between the modern Interbase vs Firebird? I thought they were
> basically the same.
>
The way I interpreted the history as I saw it, the first thing the
Firebird team did when they got ahold of the open-sourced 6.0 code was
to plow through it, figure out how to make it build (since they weren't
given a buildable version), and fix hundreds of blatent, ugly (and in
Jim's view, stupid, braindead, moronic...) bugs. The only real feature
addition was to make it handle > 4GB files. (There might have been
more, but that's all I remember in 1.0.)

Meanwhile, the first thing Borland did when they realized that having
open-sourced a product that was bringing in revenue was to create a
closed-source proprietary 6.5, which added support for big files, and
new features. Assuming that they didn't expand the staff and at-best
replace the people they annoyed enough to leave, they couldn't have
ALSO fixed a whole bunch of bugs.)

From there, the firebird team concentrated on cleaning up the source
code so it could be compiled in C++ compilers, in preparation for
object-orienting it. Meanwhile, the Borland camp concentrated on
back-porting Firebird bug fixes back into Interbase, and adding more
features (like performance monitoring tables.) Remember- existing
customers want bug fixes, but NEW customers want features- and NEW
customers pay 100% to buy the product, not 15% to get maintenance
updates- so who do you concentrate on pleasing if you're Borland?)

Now, the firebird team has been concentrating on fixing architectural
limitations, like the index size limits, garbage collection
inefficiencies, temporary tables, etc (with Jim Starkey cleaning up the
architecture in Vulcan) - while the Borland team has continued to focus
on... new features.

When I was faced with making a choice between following Borland down
the "Features" path with typical corporate reluctance to discuss their
flaws publically, and following Firebird down the "Stability" path with
open-source transparency, globally accessible bug list and source code,
etc, it was an easy decision. Our customers don't care about new
database engine features that make it easier for US to build the
product we sell to THEM- they want their databases not to crash. We
have no experience with versions of IB past 5.6, but I can definitely
say that FB 1.5 has been MANY TIMES more stable than IB 5.6 was.

Maybe the person you talked to has had different experiences, but
logically, the product that has been focusing on quality should have
produced a product with more quality than the company who has been
focusing on features, and that has DEFINITELY been our experience.