Re: What so special about PostgreSQL and other RDBMS?

Noons <wizofoz2k_at_yahoo.com.au.nospam> wrote in message news:<40b5ca15$0$8988$afc38c87_at_news.optusnet.com.au>...
> Galen Boyer wrote:
>
> >
> >
> > Success of open source isn't solely dependent on the purchaser
> > tinkering. One thing you could do, if it was open source, was
> > hire someone who specialized in that technology to fix it for
> > you. With closed source, the only option you have is Oracle
> > releasing a patch or new version. Thats it.
> >
>
>
> Sorry, no. You are assuming that:
>
> 1- it somehow needs fixing from the word go. Nothing says that
> is the case. When you buy an app using a db, you expect it to work.
> Period. If it does, you pay for the purchase. If it doesn't, you send them
> packing and get another vendor in. What the heck do you need source code
> for?

See all the discussions and lawsuits about MS API's not being truthful, er, "documented correctly."

>
> 2- "Hiring someone specialised" is somehow magically cheaper than
> getting a piece of software of the appropriate vintage. At a time
> when software is getting cheaper and cheaper and development costs
> more expensive by the minute, the entire theory of "source code included"
> is nothing more nothing else than pure unadulterated shite.

You mean Larry and Bill aren't investing heavily in the Far East?

>
> And that is the bottom line. I know that I'd like to have the source
> code and tinker around with it like Kim suggested. But the economic
> reality nowadays is that it is pure financial suicide to rely on source
> code to "fix" anything. The costs of hiring the right people to
> make that code usable (and maintaining it in a usable form) are
> astronomical compared to just getting a new release upgrade.

Depends what's wrong.

>
> Did you know that 30 years ago it was quite common to buy the OS
> for your system WITH the source code included? It turned out to be
> so expensive to even attempt to do anything with that source code that
> the whole idea was abandoned ages ago.

There was a halfway time in the '80s when you could buy source licenses for some OS's. For example, companies that sold monitoring or development tools would do better buying source code than trying to reverse engineer things. Some would do amazing things with this knowledge. For example, there was a product written in BLISS on VMS that could do joins between tables in different RDBMS's. Abandoning that idea set some realy good implementations in the garbage, to where we still haven't caught up - and people try to push products that can't even join tables in one schema!

So now we have laws against reverse-engineering and companies based on reverse-engineering to provide products and information that add value to the RDBMS's. Are we better off? I'd say no.

And we can't say that attempting to do anything with linux source code is too expensive. There are obvious problems, sure. But a large vendor such as Oracle can certainly work with it, and has. Is it a good thing that linux has grown at the expense of other unix rather than Windows? Jury's still out on that one. The most important thing that linux gives us is the ability to influence the direction of the OS more, um, directly.

>
> I'm sorry, but this whole "source code included" rubbish is old hat:
> the industry has been there, done that, and found it was unworkable.
> About time we learned from history instead of just repeating the same
> errors in a never-ending circle.

Unworkable for end-user companies, perhaps, but they are not the only part of the industry. As Jim's post implicitly pointed out, there's need for things between mass-vendors and users. It's difficult to make the business case of such things work long-term though. Just like Walmart coming through and wiping out mom-and-pop stores, Walmart removes variety and quality and good things are lost in the process. This is where social values must work with and override economics, since the demand is not properly informed.

jg

--
@home.com is bogus.
http://www.ducktapeguys.com