[Zope-DB] Interested in continued Oracle support in Zope? Raise your hand!

[Bogdan M. Maryniuck]

Hi.

On Thu, Aug 26, 2004 at 08:07:09AM +0200, Dario Lopez-K�sten wrote:
> 1) Does anyone else out there have a business case for
>    using Oracle with Python/Zope?

Yes.

> 2) Will those from (1) suffer badly, ie. will switch
>    from Python or Zope to something else, if Oracle
>    support in Python/Zope would dissapear?

I think Oracle has more commertial heavyness than Zope and it is really more
successfull (hey, flame down, I know that "MySQL -- rulezzz!" for somebody,
but not for serious bankers). So this is really obvious that if Zope/Python 
community will be Oracle-less, this means that the word "Enterprise" could 
be dissapear as well very soon (Sybase -- maybe, but Oracle -- much better
:-P ). Pity.

> 3) Will those from (1) suffer badly, ie. will switch
>    from Python or Zope to something else, if Oracle
>    support is not improved? (ie less crashes, better
>    throughput, etc).

My idea in our company was to bring Zope/Python instead of Java.
And my boss, all of them supports me. But unfortunately, there is
really just impossible have really fully-functional things. Therefore
we already switching to Java, because there everything works really OK. 
Although I use Jython there because of crappy Java C-like syntax. :-) 
But, unfortunately, this is really impossible to use Zope/Python
well with Oracle. Always something missing, something unsupported, 
something crashy and works wrong. For example, BLOB/CLOB's: they are
not supported *well*, procedures/functions can not return it and driver is
not able to handle it properly (DCOracle2 and cxOracle both of them).
NOTE: Maybe cxOracle can, but there are other serious set of various kind
of asspain).

> 4) Of those who say yes to at least one of 1-3, which
>    ones can invest actual cash to create and maintain
>    good quality Oracle Support in Python/Zope?

We'd rather tend to buy good software with good support instead of
invest lots of money to hire somebody to write a driver, which MIGHT BE 
better. What, if not? And I don't think that hack-based driver could be better
than original one. Sure, if Oracle would opensource his RDBMS, than maybe
whole world will be happy and a toy-MySQL (which I hate), finally dies. :-) 

Ideally it would be nice if Oracle Inc. go to support Python. Than we
will not care of kind of License or so: just buy good, fully 
functional driver for few USD and that's all the story what we really 
need.

> 5) Of those who say yes to at least one of 1-3, which
>    ones can invest *time* to create and maintain good
>    quality Oracle Support in Python/Zope?

Me -- not.

> Also, I seem to be one of those that really stretch the limits on 
> DCOracle2 - we use it super-extensivly for thousends of users, and seem 
> to encounter all kinds of problems with it, ranging from lockups to 
> connectivity loss.

BTW, what about mxODBC? I never tried it. At least I guess it might
work with Oracle, yes?

-- 
WBR Bogdan M. Maryniuk
Software developer
FORBIS ltd