plinehan wrote:

>
>
> OK, so Interbase was a kind of
> foster cousin within the Borland
> "family"?

Paul Beach knows more about the history of InterBase
at Borland than I do, except the very first and very
last parts.

>
> What was the rationale behind "sunsetting" IB?
>

Remember that American corporations tend to have a
three month window - six months for really forward-
looking executives. Since they could eliminate most
of the cost of InterBase without affecting the revenue
for six months or more, that would help their stock
value and the price Corel would pay for the company.

>
>
> OK, which gets back to your (and my) point(s)
> about Borland not considering IB as a primary
> product.

Their primary products were - and are - developer
tools for windows development, sold shrink wrapped.
InterBase's typical sale was > $80K, had a four
month sales cycle, and required lots of pre-sales
work. Borland didn't understand that market and
tried to push InterBase into the DBase/Paradox
market, where its strengths were incomprehensible
to the customer base.

>
>
>>The database market is pretty tough. Oracle
>>has always maintained very good account control,
>
> i.e. "you shut up about Interbase, and we'll give
> you access to our db API's"?

More like "if you want to push InterBase as a free
introductory database with your tools, that's OK
with us, but if you go beyond that, no analyst we
employ or magazine we advertise with will every
recommend a Borland product again."

>
> AFAICS, big screwup (not that you could be
> mind-readers at the time) was sale to a
> company for whom the db was rather an
> afterthought - i.e. Ashton-Tate,

We needed mind share and money. They needed
technology. Their plans for DBase4 and DBase5
were way beyond what they could build. They
did want to get into the transactional database
business. But at the same time they were
negotiating a timed buyout of InterBase they
were negotiating a technology deal with Sybase,
and Sybase in turn was negotiating with Microsoft
to be the basis for SQL Server. All the deals
happened at about the same time - within a month,
I think.

> who then
> resold to another company for which the
> db was also not core, i.e. Borland.

That was not their choice. Ashton-Tate could be
acquired or go out of business. It was the
odd case of what was perceived to be a big fish
eaten by a small fish with the end result being
an even smaller fish.

> Sybase - always a db company, though I will
> never, for the life of me, understand why
> they sold their db to Widoze.

That was not their intention. They expected to
do a licensing deal of some sort - and certainly
never intended to give up control of the product.
They basically underestimated Microsoft's technical
capability.

>
> (the above 4 are, for my money, the
> most successful db companies around,
> feel free to differ).

Ingres used to be pretty significant.

>
> But, anyone who can code and knows SQL can
> use Oracle effectively can/could...
>

You are joking, right?

> My point is that *_GIVEN THE LACK OF OTHER
> REASONABLE METRICS_*, the only *_SOLID_*
> quantitative data ... is
> job postings out there on boards.

And I don't believe it because that identifies
people with skills in databases that require
specialized skills - e.g. Oracle and its wondrous
tuning parameters - and ignores more standard
compliant, non-biting databases.

>
> I'm not a bigot on this and am more than
> willing to listen to anyone who has
> (an)other idea(s) about how one should
> measure db market penetration.

I don't know that there is any reliable way of
measuring that, but suggest that if we do choose
one of the less reliable ones, we might as well
choose one that makes us look better rather than
worse.

>
> From what I've been reading (and
> on the whole I believe that FB list
> contributors are telling the truth), FB
> is establishing itself as a significant
> player (Goldmine, German TNT, German
> Press Agency, in particular SAS) are
> all impressive clients for any firm/entity.

This week I got calls from more US companies than
I did in the first two years of IBPhoenix. Not
all of them turn into contracts, but the contracts
are bigger than they were. For nearly three years,
most of our paid work was from people who had
corrupted InterBase 5.6 databases. Now, they need
help tuning large Firebird applications. SMP and
better index garbage collection may put IBPhoenix
out of business, but they won't be the end of
Firebird... So don't measure Firebird by us..

Cheers,


Ann