On Wednesday, July 25, 2001, at 11:03 PM, Jason van Zyl wrote:
> On 7/25/01 5:22 PM, "robert burrell donkin" <[EMAIL PROTECTED]> wrote:
>
>> On Wednesday, July 25, 2001, at 09:43 PM, Daniel Rall wrote:
>>
>>> robert burrell donkin <[EMAIL PROTECTED]> writes:
>>>
>>>> is jakarta-turbine-torque going to have it's own mailing list?
>>>
>>> Currently, all traffic is part of the Turbine lists. However, you
>>> might mail [EMAIL PROTECTED] and suggest that.
>>>
>>
>> if there's enough interest in source generation (along the proposed
>> lines)
>> from non-turbiner's then it might be a good idea. anybody interested?
>
> Torque does in fact use texen as the basis for source generating Ant
> tasks, but Torque is moving toward being a persistence layer. Torque
> is commonly known among turbine developers as a generative tool, but
> that's just texen. Only the database specific tasks have been moved
> into the Torque repository.
>
> If you wanted for example to take an XML document and produce a
> corresponding bean and even digester rules for processing I would
> say that's a different project. It could be something that torque
> utilizes because right now torque doesn't use the digester for
> turning the database model into a java object it uses straight
> SAX. I would like to use the digester, so if you started a project
> for generating beans/digester rules from XML documents I would also
> be very interested.
cool :)
i've taken a good look at digester and using digester is definitely
possible (but digester does need some more work if it's going to cope with
complex schema).
i'm pretty keen to take a look at replacing that rudimentary stack-based
parser used in that xml-schema generation material with digester. the idea
would be to use the generation technology to generate a complete
xml-schema object model complete with digester-based processor which can
then be used to generate more complex models.
i'd favour a project with a broader scope (at least to begin with) -
something like a library of useful texen generation tools/scripts/examples.
we could start with datamodel, that xml-schema generation stuff (one
example which people might find interesting is generation of DOM
implementations from xml-schema), that texen-digester stuff and also basic
texen examples.
one reason why i'm keen to have a generation project with greater scope is
that i don't think that each project (by itself) would have a viable
developer-base. i pretty much stopped developing the xml-schema stuff as
soon as it did what i needed it for (ie. refactoring ecs). datamodel has
only two proposed committers (and only Michael Gerdau has contributed to
the debate). i think that you need a big enough community to get
cross-fertilization of ideas. i'd like a commons project that was a broad
enough to encourage anyone who has an interest in generation to come
aboard and contribute.
if you'd actively support a project of this kind (and maybe bring on board
some of the turbineers interested in texen) then i'm sure that it'd be
worth pursuing :)
- robert
