Oracle FAQ | Your Portal to the Oracle Knowledge Grid |
Home -> Community -> Usenet -> c.d.o.server -> Re: Python vs. PL/SQL for Oracle work
dananrg_at_yahoo.com schrieb:
> Thanks for all the responses.
>
> What if the business requirement was not performance, but code
> portability between RDBMSes?
That's hardly a business requirement in itself. It's a risk you have
to estimate the likelyhood of.
Then you can think about sacrifycing "real" business requirements
like data consistency, MTBF, maintenance costs and other stuff in order
to mitigate that risk.
> If the underlying RDBMS might change, and
> I had to assume so, wouldn't it make sense to use a language like
> Python and a generic ODBC Python module if one exists? Different
> question than my original post, I know.
If you really utilize a database, then the switch from oracle to db2
or sqlserver will involve a lot of rewriting anyway.
If you just use the database for persistent storage, you are reimplementing
database features in your application and have to compare the costs.
And customers won't like that.
At least, customers like me. I've got a company pushing something like that
onto me. Somehow they are proud of using the database in its most dumbed
down mode and have written an extra server to handle all the rest.
Their argument is that they can sell to shops with different databases but
each individual shop typically is heavily invested in one particular software
vendor, like microsoft or oracle, with support contracts and everything, has
trained admins and is very very unlikely to change the database in the
foreseeable future. But they all puke over a product that does what their
database does for free, only slower, with more maintenance overhead, less
stable, with a new set of interfaces and procedures and whatnot.
And with your puny customer base you will *never* play in the same league, Oracle, Microsoft or MySQL are playing in terms of reliability and performance, simply because you don't have the means to get it right.
So, if you have a fixed amount of money to spend on a project, don't spend it reimplementing database features. Spend it on reimplementing the same business logic in PL/SQL, ANSI-PLSQL and TransactSQL or whatever Microsoft uses. They all have their specialties that save you a lot of work on their platforms, making it ultimately cheaper for you too. Also, no customer will complain to *you* if their database crashes, losing data, but they will crucify you if they lose data with your server or otherwise nonstandard implementation.
Lots of Greetings!
Volker
Received on Thu Feb 23 2006 - 12:41:32 CST