Well, I obviously have a gift for understatement when I merely say I
don't see the benefit outweighs the cost. LOL You folks are AWESOME
for providing such impassioned responses. :-) Even before I wrote the
original post, I had pretty much decided that I would persuade the
client to go in a more canonical direction, but having real people
who've tried it tell of their battle scars can only help me make my
case. To sum up, there are major security, maintainability, scaling,
flexibility, and cost issues that I will need to address with the
client before making any decisions to proceed or not.
Thanks to everyone for corroborating my concerns. Have a great
weekend y'all!
- Kevin
On Sep 24, 2009, at 2:21 PM, Alex Mendoza wrote:
Hello All,
Too right! Exact same experience here at NOAA with the Oracle pl/sql
asset management application I took over...to the letter.
I started this position(as an Oracle developer/DBA) with "feature
requests" for that application more than two years old. I soon
found out why they had not been accomplished before I started on the
project. The lumbering beast that was so expensive to license and
maintain (Oracle quoted $300K to replace software)...was near
impossible to quickly adapt to changing business environments
without breaking it completely. The business logic(by design) was so
integrated into the system, we could hardly change without rewriting
large portions of the application/database and testing for weeks.
Today the same application completely re-vamped in ROR takes up
about 54mb of disk and runs free on multi-threaded Postgresql.
Feature request turn over today, with testing, is about a week.
I consider myself to be a recovering Oracle developer now. (thanx
ara, my wise teacher)
ha!
am
ara.t.howard wrote:
On Thu, Sep 24, 2009 at 12:11, Bothari <[email protected]> wrote:
One client I worked for had the same idea. They did their entire
eCommerce site in PL/SQL for Oracle. They didn't do it for
security,
but because PL/SQL was so versatile. It "rendered" pages, parsed
xml,
wrote files, and sent email. It was abandoned because it was so
complicated and fragile that upgrades took forever, and it took 3
full-time testers to keep bugs (mostly) out of production.
i had the same experience at NOAA actually. in the end they canned
the guy running it and started over from scratch on rails ;-)
you might ask alex mendoza about it - he is the one who wrote the new
rails app - bcc'd on this message.
_______________________________________________
Bdrg-members mailing list
[email protected]
http://rubyforge.org/mailman/listinfo/bdrg-members
