Re: Open Source Oracle?

From: Howard J. Rogers <hjr_at_dizwell.com>
Date: Wed, 15 Sep 2004 09:45:45 +1000
Message-Id: <414782a8$0$20581$afc38c87@news.optusnet.com.au>

ctcgag_at_hotmail.com wrote:

[snip]

>> The availability of a free version, in short, does not preclude the
>> existence of the paid-for version. And whilst Oracle might lose out on
>> the small-time operators and installations, they might take the
>> opportunity to enhance their large-scale operations... and charge a
>> (further!) premium for doing so... so there's money in it for them, too.

>
> I don't see how open-sourcing old code would aid in the endeavor to
> enhance
> their large-scale operations. It pretty much seems like two independent
> things.

Look at Red Hat. They now don't "do" freebie Linux. They only charge an arm and a leg for the 'Enterprise' class stuff. Same deal for Oracle, I was thinking: small end of town, OpenOracle, we don't support it, save some costs etc etc. Big end of town, Oh yes, we do EnterpriseOracle, that will be both arms, one leg and a lifetime supply of bank guarantees, please.  

>> Alternatively, consider that little companies which start with freebie
>> software have a habit of growing up to be big-ish companies that can pay
>> good money for their software.

>
> Well, some fairly small percentage of them do.

Of course. A much larger percentage goes bust. But a slice of a small percentage is better than nothing at all, which is what they get at the moment when it's a toss-up between MySQL and Oracle. Actually, for the sort of people I have in mind, it never actually gets to *be* a toss-up, since Oracle is "well-known to be too expensive" from the outset.  

>> Wouldn't it be a good idea to "infect" the
>> minds of these sorts of people in the early, el-cheapo days, and clean up
>> when the hard cash starts being passed around?

>
> If they really need Oracle1 10g, they will buy it, regardless of what they
> previously used.

OK, this one could be a good debate on its own. Tell me one feature that means Oracle 10g is a must-have database for a company with, say, a staff of 20 and a turnover of, say, US$200,000. That will effectively arm-twist that company into buying 10g, because they "really need it".

I can't think of a single one. ASM, ADDM, and whatever other flavour of alphabet soup you care to stir the pot with, are all 'nice to have if I have the cash'. Not 'must-have so I will find the cash'.

OK, so now I'll contradict myself a little: flashback in 10g is *almost* worth the cash. But only almost, so I avoid the contradiction, I think!

10g is all lace and crenelation, not earthworks and foundation. If I'm not of the Oracle mindset from an early stage, there are a lot of alternatives I could choose.

> If they only kind-of-sort-of need Oracle 10g, then
> Oracle would only clean up if that growing company had a good deal of
> loyalty to
> Oracle. I would contend that companies like that are the kinds of
> companies that would have been willing to pay in the first place, even
> when they were small.

I've just been dealing with someone on the Suse newsgroup where that is evidently not the case. I realise one anecdote doesn't make a novel, but there you go...

I also recall starting work for a company of 3, when Oracle would have been a complete luxury, and leaving it 6 years later with a workforce approaching 30 (OK, hardly big time), an annual turnover of a couple of million pounds sterling, and a willingness to spend cash on a "proper" database at last.

So it does happen. We can leave it hanging as to how much and how often.

Regards
HJR Received on Tue Sep 14 2004 - 18:45:45 CDT