avalon-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Leo Simons <leosim...@apache.org>
Subject Re: Verifying Meta Tags
Date Tue, 08 Apr 2003 20:39:10 GMT
Berin Loritsch wrote:
> Leo Simons wrote:
>> Berin Loritsch wrote:
>>> Am I misunderstanding anything?
>> you might be.
> The tags, man, the tags.  What about the tags.  It should be a subset
> of the Phoenix tags is all I am saying.

+1 to that (I think you mean superset, but +1 to that as well if its all 
you are saying). We are already on the same page there I think.

It is just that I think you got them slightly wrong still (see msgs from 
Peter and Peter :D).

> Once folks are used to generating the metadata with ANT, it really
> doesn't matter how it is stored.

I disagree again here. What is important is the contracts surrounding 
the data format (whether that's javadoc tags or some other kind of 
storage), not what tool is used to convert between formats or the format 
itself. That still boils down to the point you were making though...

IIUC, we are trying to figure out what semantic "metainformation" is 
most commonly associated with a java class or interface within the 
context of the avalon container-component contract, and you are saying 
we should get that figured out before we talk about how to encode that 
information into some kind(s) of data format(s).

- I agree with the goal and the road chosen to get there.
- I also agree with starting off with the current (well, 4.1 so cvs 
head) phoenix metainformation contracts (though some, like phoenix:mx, 
should obviously be excluded from any common namespace).
- I agree with the earlier conclusion that further contracts should stay 
container specific and namespaced appropriately, until we come to the 
conclusion that they are/should not be container specific, as far as the 
containers inside avalon which have metainfo capability are concerned.

Having that settled, moving on to the next step of actually talking 
about the format again...

- I also agree that it makes sense to use (or at least support) javadoc 
tags, like has been done in phoenix and in/for many other projects 
outside of avalon.
- I also agree that it makes sense to use a common javadoc tag format 
where possible.
- I think it makes sense to follow the lead of the biggest (and very 
successful, and coming to apache) project which deals with this stuff, 
namely xdoclet.
- I think it makes sense to use some sort of namespacing.
- I think it makes sense to have both a common avalon namespace and 
container-specific namespaces.
- I think it makes sense to KISS.

The xdoclet general format is

@namespace.tag-name attribute-name="attribute value"

where you can have stuff like

  * @ejb.bean
  *     name="bank/Account"
  *     type="CMP"
  *     jndi-name="ejb/bank/Account"
  *     local-jndi-name="ejb/bank/LocalAccount"
  *     primkey-field="id"

The other mayor player in this area is QDOx, which IIUC can support such 
a format well, and completely. I'm not sure whether xdoclet works with 
the ':' phoenix uses, but that's a detail.

If we can accommodate all of the above (and I think that's the idea), I 
don't care about implementation details. I do think it is a bad idea to 
have so many implementations lying around, which led to my earlier message.


- Leo

To unsubscribe, e-mail: dev-unsubscribe@avalon.apache.org
For additional commands, e-mail: dev-help@avalon.apache.org

View raw message