abdera-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Davanum Srinivas <dava...@gmail.com>
Subject Re: Abdera status and future
Date Tue, 05 May 2009 21:55:33 GMT

If i get some breathing room :) i'll hop in.


On 05/05/2009 05:06 PM, James M Snell wrote:
> Davanum Srinivas wrote:
>> James,
>> +1 to nuke axiom. No problems at all from me. Definitely recommend
>> doing that.
> Wanna help? :-)
> - James
>> thanks,
>> dims
>> On 05/05/2009 04:32 PM, James M Snell wrote:
>>> Ok, so I've got a fix for the ant build pending commit... for some
>>> reason the svn.apache.org server is not responding right now so I will
>>> commit that as soon as possible.
>>> In the meantime, I'd like to discuss Abdera plans moving forward.
>>> I know the 1.0 release was voted on but was never actually cut. I'm
>>> going to try to get that pushed out in the next week or two.
>>> Moving foward, there are a number of changes and refactorings I'd like
>>> to look at:
>>> 1. Axiom does a good job of handling the XML infoset stuff Abdera needs
>>> but it brings along with it a number of dependencies and extra stuff
>>> that we just don't need (like all the SOAP stuff). In an ideal world,
>>> Abdera would be entirely self-contained. I would like to start
>>> investigating what it would take to remove the Axiom dependency from
>>> Abdera altogether. This would mean duplicating a large part of the Axiom
>>> functionality, but we'd be able to do so in a way that was highly
>>> optimized for Abdera's purposes. This would mean a pretty much complete
>>> rewrite of the FOM* implementation classes but should not lead to any
>>> significant API changes. It would give us significantly greater control
>>> over the parsing process, would allow us to make certain dependencies
>>> (like Jaxen) optional, and would reduce our overall install footprint.
>>> Initially, the new implementation could sit along side the Axiom based
>>> implementation through it's development, and would eventually replace
>>> the Axiom implementation as the default. I know there are some folks
>>> here that are fans of Axiom (hello Dims! :-)...) and I don't want to
>>> upset them, but I think there are a number of important benefits to
>>> replacing Axiom that cannot be overlooked. Thoughts welcome.
>>> 2. The Abdera i18n code (IRI, unicode, uri templates, etc) should be
>>> moved to its own subproject under the Abdera Top Level Project. It is
>>> extremely useful independent of the rest of Abdera and should be
>>> developed and promoted independently.
>>> 3. I'm going to be working to integrating the more fully featured RSS
>>> capabilities contributed by a fellow IBMer. I'm not really all that
>>> happy with the way it was implemented in the offered patch so I'm going
>>> to go back through it. Unfortunately I didn't have the opportunity to
>>> dig into the patch before, but now that I have time to devote to Abdera
>>> again, I will go through it in more detail.
>>> 4. Simplification! I wish to continue working to simplify the Abdera
>>> implementation and APIs and to refine the implementation, configuration,
>>> etc. In some cases, this may mean API changes, some of which may not be
>>> backwards compatible but will continue the trend towards a better,
>>> easier to use, more functional Atom implementation.
>>> Whatcha think...
>>> - James

View raw message