db-torque-user mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Geoff Fortytwo <...@ign.com>
Subject Re: Torque project references, stability, reliability
Date Wed, 02 Apr 2003 20:07:50 GMT
I agree on all these points. I'd like to add a few points of my own. I 
recently started using JAXB for xml parsing and it uses a system similar to 
torque. You specify a schema that describes the xml file and it generates 
java files using ant which are used to represent that data. However, it 
also allows for additional validation and I've always wished that torque 
had a pluggable validation capability. Errors that come from the database 
are not standard, so I can't write code that'll work with multiple 
databases if I write code that checked the error code in the SQLException. 
OTOH, if Torque had optional pluggable validation that validated each field 
as the set method was called for it, then I could write code that caught 
and dealt with those exceptions in a portable way. We use both Oracle and 
SQL Server here, so I need to be able to write code that works for both. 
This point has been answered before that validation should be at the db 
level, but that's just not portable.

At 12:33 PM 4/1/2003, you wrote:
>We've used Torque extensively in the last 12 months as the O/R wrapper for
>both a web site and a server application.  This includes use both within
>Turbine and in separate applications.
>I like these features
>-- robust
>-- reasonably well-documented
>-- does the basics well (simple queries, insert/update, delete)
>-- source code is very readable
>-- generates Java source code from XML schema in a way that allows me to
>easily update both
>-- the Torque-user list is very helpful
>-- Torque integrates very nicely with a MVC approach
>A note on the last point.  A lot of our code involves a servlet processing
>velocity pages.  The servlet can do a query, which generates a list of
>objects mapped to records.  My velocity page can iterate through those
>objects and display the properties with some very simple presentation layer
>statements.  I've found this combination (Torque/Velocity/Servlet) to be
>highly productive for basic data-backed applications.
>I've been frustrated on these points
>-- Doing queries with "ands" and "ors" is confusing and not robust.  The
>"criteria" object is not easy to use for complex queries.
>-- All the generated Java OM source code files inherit from a common Torque
>class, making it hard to add new methods across the board.  This is
>additonally difficult in the Peer classes in which the important methods
>(e.g. doSelect) are static.
>-- I've submitted a few small patches for bug fixes and simple feature
>enhancements, and have had difficulty getting any of the committers to pay
>attention.  (to be fair, I haven't tried in a while).
>Hope this is useful.  I'm currently doing a project using Hibernate, just to
>compare and see the difference.
>Be glad to answer any questions on my experience.
>Best, WILL
>P.S.  A couple of example apps:
>http://broadcast.forio.com/discuss  (fairly simple discussion board)
>http://broadcast.forio.com (business simulation development/hosting,
>requires registration)
>Forio Business Simulations
>Will Glass-Husain
>To unsubscribe, e-mail: torque-user-unsubscribe@db.apache.org
>For additional commands, e-mail: torque-user-help@db.apache.org

View raw message