The interesting comparison is not Interbase/Firebird vs. Oracle but
Interbase/Firebird and Firebird vs. MySQL.

Interbase and Sybase were launched in the same week. Interbase was
capitalized with a Steelcase chair and two two-drawer file cabinets to
hold up a pre-existing door, with one full time employee, and two cats.
Sybase got $40M+ in venture money. The products came out about the same
time, but in different markets. Interbase went for early adopters in
the workstation market. Sybase went for enterprise sales in what we
would now call the departmental computing market. Sybase's product was
pretty terrible (page level locking, fixed size log file, all sorts of
nasty restrictions in the API) and never got much better, but they had
lots of money for marketing, advertising, and a very, very expensive
sales force. Interbase had high presence in engineering end the
aero-space business when a couple of bad things happened. First, our
partner and eventual owner, Ashton-Tate, flamed out and went into an
unrecoverable tailspin. Second, Ashton-Tate was bought by Borland.
Borland had some short term successes, but thought like a PC company,
which was the kiss of death. They destroyed the support organization,
decimated the engineering organization, eliminated marketing and sales.
At the same time, they imposed their idea of support, which was give the
page number in the user manual and hang up, on a database product, and
worse, refused to give out any information on upcoming releases.
Borland pretty much destroyed the engineering workstation market in
search for packaged product retail sales. Interbase gradually lost its
US presense; growth was elsewhere. Borland also put out version 5,
which didn't work. Still, Sybase never threatened Oracle, and
disappeared into the shadows. Could Interbase under Borland do any
better?

Borland released Interbase under Mozilla because we believed that
classic was impossible under GPL (and a deep an visceral dislike of the
GPL in general). Perhaps with GPL we could have piggybacked on the
Linux distributions, but at the time open source Interbase was release,
MySQL had that pretty much locked up. MySQL was truly god-awful by
itself, but InnoDB was a respectable transactional engine that covered
up some of the more grotesque of Monty's ideas. Interbase, now
Firebird, had many advantages over InnoDB, but MySQL got mindshare, was
automatically installed by most Linux distributions, and InnoDB blew
Firebird out of the water on performance. MySQL also had an
organization behind it, excellent documentation, and had enough critical
mass for third party support companies to thrive. MySQL also had a lot
of top tier European venture money and a board and CEO with vision,
intelligence, and patience.

Firebird technology is really quite dated with major holes. Classic is
crash resistant, but isn't secure and doesn't scale under load.
Superserver is not crash resistant without forced writes, which does a
number on performance. Superserver also had (has?) a brain dead
security model, was essentially single threaded, and had a layered SQL
implementation that was an architectural nightmare. All these could be
fixed but for a vicious political climate.

Guys, engineering is not a zero sum game. Ideas that are not your own
sometimes have merit. Habitually announcing opinion as conclusions or
decisions is poor project governance. Ignoring backwards compatibility
is a sure-fire way to lose customers. Also, SMP seems to have caught
on. Deal with it.

Personally, I enjoy the interchange on this list. It's a good place to
discuss database architectural ideas. People are open and thoughtful.
But man, the developer climate is poisonous.


Paul Ruizendaal wrote:

>> Very flattering, but I don't think so. Oracle is absolutely
>> first rate at customer control. They also had a head start in
>> the market - well Ingres may have been ahead of them, but Ingres
>> picked QUEL rather than SQL. In the early eighties there was a
>> rumor of a study done of learning QUEL and SQL - end-users
>> favored QUEL slightly, but it could not be taught to programmers.
>> Anyway, Oracle had much more aggressive marketing, did a better
>> job of working with upper management, and used SQL as an effective
>> weapon. The unique InterBase features were available only through
>> GDML and Oracle convinced anyone we competed for that a non-SQL
>> interface was the kiss of death.
>>
>> InterBase and Sybase started at within a month of each other.
>> We turned down venture funding at the beginning because it
>> scared us. They were braver (or more experienced, or maybe
>> just lacked imagination) and got much more exposure.
>>
>
> I was still in school and doing army national service in the 80's, so I am
> not talking from direct knowledge. However, back in 2003 I did a fair bit
> of research into the history of the various database lines and companies.
> My understanding of how it all happened aligns pretty much with what Ann
> says above: Oracle was a very focussed and aggressive sales machine, and
> Larry Ellison was the Formula One racer in the cockpit.
>
> Note that there is a strong "survivor bias" in corporate history. We never
> hear of companies that made similar choices but are not around any more.
> Larry Ellison has at least twice "bet the company" on a particular
> strategy; the bets worked and now he is a very rich man. They might have
> failed and nobody would remember today who Oracle was -- just another
> software graveyard corpse (Oracle almost went bankrupt in the early 90's
> and nearly all Larry's money was in Oracle stock). For every success there
> are 10 failure companies that used exactly the same strategy, but we only
> remember the survivor.
>
> If anything has held back Interbase and Firebird (other than a willingness
> to drive without brakes), I would say that its culture continuously focuses
> on "product" first and seeks customers next. The odds of commercial success
> are far higher if that is reversed: first invest in customer relations and
> then build the product they need. The big benefit of focusing on product
> first is that it leads to, duh, a good product. Nothing wrong with that,
> and as Ann/Jim are living proof: one can have a very comfortable life that
> way too.
>
> For the interested, here is some reading:
> http://www.amazon.com/Real-Story-Informix-Software-White/dp/0972182225
> http://www.amazon.com/Softwar-Intimate-Portrait-Ellison-Oracle/dp/0743225058/ref=pd_sim_b_1
>
> Just my 2c worth.
>
> Paul
>
>
>
> ------------------------------------
>
> Yahoo! Groups Links
>
>
>
>
>

--
Jim Starkey
Founder, NimbusDB, Inc.
978 526-1376



[Non-text portions of this message have been removed]