Nissim and Sean,
I wish you guys would try to integrate my OPL into
Turbine. Seriously, it would be much easier to deal
with db differences using a system which has a mapping
layer like Castor or OPL. OPL would be easier for you
guys because it supports criteria objects and is a
much smaller system.
Recently, many of the development issues have some
connection -- no pun intended ;-) -- to the database
access.
I will have the new OPL in a public CVS within 2
weeks.
While adding this, you could rework Turbine so it has
"hooks" for data access so you could plug in either
the Peer model, OPL, or some other system. By making
this more open with clean interfaces, you broaden
Turbine's appeal.
Since my last posting of source to Turbine users, I've
made the following changes / additions to OPL:
1. Replaced use of attributed interface with vectors.
This uses fewer objects so should be faster, more
efficient.
2. Modeling Oleg Nitz's code in Castor, I added key
generation for max key, auto-increment, and sequence.
3. Added auto-update of foreign key values in
dependent tables.
4. Added Castor's (Assaf Arkin's) type conversion
features. I'm packaging Assaf's code for type
conversion in the distribution. It's so clean, I can
use it with zero code changes.
You can see that I really like Castor. Problem for me
with Castor is that Castor is moving toward being more
O-O to be consistent with Sun's standards. Castor
doesn't have support for Criteria objects, and it's
unlikely you'll ever have the low-level access that's
in OPL.
While programming, you have the criteria data. With
Castor, you take this criteria and formulate an OQL
query, which is given to an OQL parsing engine, and
then the parsing engine builds criteria objects which
are used to build the query expression. With OPL, you
build these objects directly.
A note to Turbine / Castor users is that OPL has
support for multi-valued keys and using connection
pooling. In the code, you aren't aware of the
pooling; it's implemented with configuration.
For users, the major problem in using Castor or OPL is
editing the class maps. I will write a schema for the
mapping, and, hopefully, you could use Xeena to edit
the maps.
- george
__________________________________________________
Do You Yahoo!?
Get Yahoo! Mail - Free email you can access from anywhere!
http://mail.yahoo.com/
------------------------------------------------------------
To subscribe: [EMAIL PROTECTED]
To unsubscribe: [EMAIL PROTECTED]
Search: <http://www.mail-archive.com/turbine%40list.working-dogs.com/>
Problems?: [EMAIL PROTECTED]
