Re: OCFS2

oracle-l-bounce_at_freelists.org gravou em 2005-08-08 10:24:31:

> > Is that a fault of Linux, GNU, Red Hat or Oracle?

> How about IRS? I'd like blaming them for something.

        That's precisely the point. Only from symptoms you can't pinpoint the cause.

> > Given known history, I'd bet Oracle. Perhaps Oracle, but more
> > probably Oracle.

> Nope. Oracle is unbreakable.

        What? Are you a paid shill of Oracle, are you on dope or what?

> Oracle has its fair share of problems but Linux is far
> from being blameless. Look at the kernel 2.6.12 and how many
> subversions were officially released.

        So what?

        No one should be running 2.6.12 as it is released from Linus. It's been quite some time since the standard procedure is to consider Linus' Linux as just a baseline for distributors to create their own kernels.

        While this sounds like fragmentation, welcome to the world of monolithic kernels and complex hardware. Why do you think we (still) have so many flavours of Unix?

        The real difference is that in free software one admits errors, while proprietary vendors -- Unix, SQL, whatever -- tend to hide them so that innocent (or lazy) people will believe their claims, like Oracle's 'unbreakable'.

> What is SQL standard? SQL 92?

        ISO SQL:1992 is quite old now. One should assume by now Oracle would target at least ISO SQL:1999 if it was interested at all.

> I hate ANSI joins and alike. It gives people the
> impression that it is possible to write "database independent
> applications". That
> has resulted in some of the most abominable applications that I've ever
seen.

        It is not a fault of the standard that vendors refuse to comply with it.

        Or perhaps the standard is too complex, so it is hard to comply with it. But then creating endless variations will hardly make it easier for us users. And I'd really like to shift some work to the vendors.

> Oh, come on! Linux is partly to blame! Look for Torvalds vs.
> Tannenbaum debate and see the
> OS model chosen against the advice of other, more experienced
> people.

        Sorry to break it to you, but if you make a point of it you should start testing the GNU Hurd. All Unices are monolithic. Even OSF/1 and Mac OS X aren't true microkernels since they have single servers on top of Mach.

> As a result, Linux
> kernel is now a monster that is a bit overwhelming.

        Agreed. But how exactly is HP-UX and Solaris better? They run on far less hardware, but this just mask bugs and make them expensive to run.

> Kernel development cannot be run like
> a pyramid scheme: you cannot expect new developers when you need
> them. Telling your friends
> to find new developers will not always work.

        Are you aware how many developers are actually paid to work at Linux, both the kernel in particular and the GNU OS in general?

--
Leandro Guimarães Faria Corcete DUTRA
Administrador de Bases de Dados      +55 (11) 4390 5383
Toyota do Brasil Ltda              ldutra_at_toyota.com.br
São Bernardo do Campo, SP                        BRASIL



This message (including any attachments) is confidential and may be privileged and intended solely for the use of the person/entity to whom it is addressed. If you have received it by mistake please notify the sender by returning via e-mail as well as delete this message from your system. Any unauthorized use or dissemination of this message in whole or in part is prohibited. Please note that e-mails are susceptible to change. TOYOTA DO BRASIL LTDA (including its group companies) shall not be liable for the improper or incomplete transmission of the information contained in this communication, neither for personal, nonbusiness related information nor opinion sent through this email or even for any delay in its receipt or damage to your system. TOYOTA DO BRASIL LTDA (or its group companies) does not guarantee that the integrity of this communication has been kept nor that this communication is free of viruses, interceptions or interference.
--
http://www.freelists.org/webpage/oracle-l