""Rob Schuff"" <rob@...> wrote in message
news:014901c1affe$decdfc80$0200a8c0@picard...

> Well, that's a mighty big bite! But it seems to me that postgres and

mySQL

> people are those who are looking for alternatives to the MS/Oracle, so why
> not agressively go for that "market" first? A worthy question is: Why

are

> those who are looking for alternatives to MS/Oracle not as often choosing
> Firbird? It seems that going "straight to the top" given that so few have
> even heard InterBase let alone Firebird, might not be the best way to go.

There are not only technical reasons, but "political" reasons, as you know.
I'm not speaking about the Congress <g>, but about decisions that are taken
inside organizations and that are not only based on crude technical merits.

Try to act as the IT manager's advocate, for a few seconds. Do you select IB
if you know that:

- The product never was marketed appropriately in the past so it appeared as
if the company owning it was a bit ashamed of it?
- The owning company was accused of killing the product at the end of 1999?
- The product's open sourcing story was plenty of shakes and recriminations
among several parties?
- Only the commercial version is actively backed by a company the size of
Borland?
- There're still no clear plans nor clear route for the future WRT the IB OE
flavor? How does it fit in the lanscape? Does it receive only bug fixes? How
often? Does it receive features but in a delayed way? Will it exist always?
Will it become frozen at some time in the future so it turns irrelevant?
- There's a free version named Firebird, done by a bunch of crazy people and
backed by a tiny company named IBPhoenix?
- The product wasn't born as open source, but it became open source as a
result of some decisions the owning company seems to regret today? (As Bill
Todd himself wrote, "nobody knows exactly why top management decided to open
source IB initially".)
- The owning company doesn't publish list of bugs for either the paid of fre
e versions and the main fork doesn't have a guaranteed critical mass of
developers because nobody pays them?
- We've been reading the forums and there appears to be quite a bunch of
bugs. How opportunely are they acknowledged and fixed? Where are the
reliable contacts we can address in case of those problems, that doesn't
entail purchasing a support ticked immediately?
- It seems that Borland's revenues come exclusively from selling their
closed version following the old good commercial model and IBPhoenix'
revenues come from optimizing applications and repairing databases. Is this
model profitable so that those companies won't drop the product? BTW, by
reading the forums we came to the conclusion that IB databases become
corrupt often and it's not uncommon that they cannot be restored, too.
- Instead of trying to push a common project, those IB/FB people work each
one for their respective projects and there's no cross-pollination nor
cooperation. More often than seldom we see guerrillas. How weak will the
original engine became in terms of market share if more forks continue
appearing and its small market fragments even more among the different
forks? There are already 4 known flavors!
- We would like to take a risk if we know another company of some importance
that has decided on IB. Where are those companies listed? It seems that some
customers like Boeing and a processor manufacturer are history only.
- In comparison with other open source and closed, commercial projects, IB
has developed few notable improvements in the last years. It seems the
project lost momentum. Most of its features advertised as unique in the past
can now be found on PostGres or MsSql. We are not impressed with a lame XML
support or a lame ANSI substring support.
- Borland has worldwide presence but only a few partners handle IB
adequately. IBPhoenix has points of contacts in a few strategic zones only.
Who's our local and proficient support company? How many ISP's risk
installing IB for their users or offer IB-based services?

When I wrote IB, I was mostly referring collectively to all the flavors, IB,
IB OE, Firebird and Yaffil.

> I am not saying that MS/Oracle should be the ultimate targets but rather

I'm

> arguing that the first step should be for FireBird to become the deFacto
> standard RDBMS of choice in hte open source world. That and only then do

There's no need to bite the neighbors. Look, even Del Yocam practiced that
when he was Borland's CEO: he never mentioned IB, he always praised Oracle
only. Nobody is going to demote PostGres and MySQL. Even MS is unable to do
that. Recently, I read an article where the MySql's architect downplays ACID
features as unnecesary for most modern applications, specially for Internet
applications. It seems that several users agree or don't care at all.
PostGres has a long open source tradition. There are others less known or
experimental engines. (There's a link to free dbs in one of the pages at my
site.) To be true, we are the newcomers here. Pretending to behave in an
arrogant and aggressive way to capture users is going to backfire. And
there's SAPdb, too.

Our hope is to replace MsSql where it makes sense to replace it. I mean
where you have mostly OLAP requirements to crunch hundreds of GBs, IB
doesn't make sense because you will starve sitting in front of your desk
before getting an answer from IB (and furthermore, the sorting will be wrong
<g>).

We can replace Oracle where it's a waste or resources of all kinds to run a
800 pounds gorilla to have only a few big queries and the rest is
interactive editing (where traditional engines usually shout for deadlock).
We can't replace Oracle to take advantage of 3 GB RAM to drive a monster db
just because IB doesn't know how to do with such amounts of RAM.

We can replace DB/2 where people are tired of low level tweaks typical of
IBM, like making stored procs with an an external C compiler and prefer
instead a simple and basic built-in language that solves the most immediate
needs. Interestingly, our Proc/Trig language is near to the SQL
specification.

We can replace other engines by working ad-hoc, namely, by paying attention
to the scenario and being sincere if the engine is able to undertake the
task. Otherwise, the attempt only backfires. Don't promise what you can't
do.

> I
> believe that FB will have the critical mass of user acceptance and media
> support to drive it towards the ultimate target.

I found a casual poll in some German site where FB beta was a few percent
points below IB6, but the poll didn't differentiate between IB6 offerings:
the old one that Borland still doesn't delete from their site, the free one
compiled by Rob and the commercial one.

What's interesting to observe is if IB+FB take market share. But open source
advocates will want to count only IB-OE, Firebird and Yaffil.

> rob (suddenly appearing again from the ether)

I would be interested in learning how to vanish and reappear from the ether.
Is there any doco available? Maybe a paper written by a magician?

C.
--
Claudio Valderrama C. - http://www.cvalde.com - http://www.firebirdSql.org
Independent developer
Owner of the Interbase® WebRing