Subject RE: [Firebird-general] "DevCo" & Firebird
Author Steve Summers
As I see it, Borland has focused on features, while firebird focused on quality and performance- first converting the code base
to C++ and fixing the bazillion bugs in the crap that was open-sourced, and then fixing the index logic, query optimization,
etc. The benchmarks we did between IB 7 and FB 1.5 a couple years ago, show that FB was ahead in performance, and the latest
comparisons I've seen show that FB 2 is way ahead of IB 7.5. We never deployed any IB 7 sites, but FB 1.5 has been many times
more stable than IB 5.6 was.

I see this as a big difference between commercial software and open source. Borland's development pressures are to address the
checkmarks they don't have, which cost them sales. The features you mentioned are the kind of thing a commercial company does to
make the product more attractive to the decision makers. Meanwhile, if it crashes occasionally, well, that's what your customers
buy support contracts for, right?

Firebird's development pressures are to increase speed and stability, and make the code better, because the people developing it
are (mostly) volunteering and a big part of their compensation is the recognition of being really, really good programmers -
people who care about speed, correctness, elegance, usability, etc- but don't have time to throw at "glitz" factors, especially
if they're ugly kludges under the hood.

Meanwhile, the users giving feedback to the developers are USING the product already. They complain if it's slow, and really
complain if it crashes. But they managed to build their application without performance monitoring or statement canceling
(things our company REALLY wants, but are getting by without), and are OK with using Classic if they need SMP support-
otherwise, they'd have chosen something else. They might ASK for cool features, but since they're not having to pay for it, they
don't feel particularly abused if those features don't arrive right away.

Look at Linux vs. Windows and you'll see the same thing. Windows is flashy, and practically feature ENCRUSTED - there's almost
nothing you can do in it that you can't do in about 5 different ways. But if you need a server that's efficient and just works,
you're better off running Linux.

The big problem with IB as a commercial product is that commercial software is driven by economy of scale- the more you sell,
the smaller the development costs as a fraction of revenue. How can Borland (or "DevCo") possibly compete in the commercial
database market (over the long term) with SQL Server? MS sells so many copies by virtue of the product's first name that they
make giant mounds of money off it, while keeping the price reasonably cheap. Last I checked, IB was pretty similar in price -
and I doubt Borland could charge less and have IB be profitable at all, with its relatively tiny market share. But SQL server
can remain very profitable even with MS spending probably a hundred times more on its development.

The ONLY way to compete with the kinds of features MS can pile on to SQL Server, is by being free, or having an even bigger
market share. DevCo has ZERO chance of becoming bigger than MS or Oracle or IBM, so what chance does IB have?

-----Original Message-----
From: [] On Behalf Of Martijn Tonies
Sent: Friday, April 28, 2006 11:38 AM
Subject: Re: [Firebird-general] "DevCo" & Firebird

> Personally, I think there is potential for better times ahead for both
groups. I can't see how DevCo could consider Interbase,
> which is now "almost as good as Firebird other than the price", as a
viable part of their product line. They'll be selling
> primarily IDE/Compilers, which historically have been database
> agnostic-
and need to be, since focusing on their OWN database
> would make the products viable for maybe 1% of their potential market.
> So why waste programmers on advancing a product whose market share
> will
continue to decline, which is competing with Free (and
> not keeping up in performance or quality)?

I don't want to start a "InterBase does this better than Firebird and vice versa" e-mail exchange, but I find your above remark
a bit short sighted.

Despite whatever you say about Firebird does not mean that InterBase is always following it. InterBase has SMP support, while
Firebird does not. (Vulcan will do this, but it's not finished yet). InterBase has statement cancelling, and although people say
"it's easy", it's still non existing in Firebird. InterBase has monitoring tables. And there's more ...

Martijn Tonies
Database Workbench - development tool for Firebird and more! Upscene Productions My thoughts: Database development questions? Check the forum!

Community email addresses:
Post message:
List owner:

Shortcut URL to this page:
Yahoo! Groups Links