Re: Continued development of XMLPull

Stefan Haustein wrote:

> > Alek wrote:
>  
>
>>i see no problem in using XPP3 (or XPP4) as a vehicle to test such 
>>ideas. i am also toying with an idea if taking back modifications to 
>>XPP3 that happened in StAX RI (it is based on XPP3) and merge them back 
>>to XPP4 codebase so it would implement both XmlPull and StAX ...
>>    
>>
>
>
>I think using XPP (and kXML) as test vehicle for new features is a very 
>good idea. Supporting both interfaces is probably the best option for 
>XPP, but may not be ideal for kXML. The problem I see with improving the 
>xmlpull interface significantly is that xmlpull basically is a common 
>interface for XPP and kXML, but what's good for XPP is often not ideal 
>for kXML.
>
>...
>
>
>I think we should position XmlPull.org in the future
>
>- as a maintenance resource for the existing xmlpull interface
>
>- for general Pull parsing promotion/education (including StAX);
>   Sample code fragments may even work for both APIs w/o modification
>
>- as a resource for discussing new ideas and concepts related
>   to pull parsing
>
>- as java namespace and project location of a free implementation of the
>   event object API and Factories, based on the Stax XMLStreamReader.
>
>- as java namespace and project location of StAX and XmlPull
>   based utilities and samples such as
>   - providing an XML stream from a DOM tree
>   - a DOM builder
>   - SAX adapter
>   - JUnit tests
>
>  
>
I think these are great ideas. I don't see why they'd be incompatible 
with defining an experimental XMLPull2 interface, though. Even if XPP3 
is the vehicle for testing and refining that interface, it's still good 
to end up with an interface that's separate from the implementation.

Even though I'm on a pair of JSR Expert Groups (JAXB 2.0 and JAX-RPC 
2.0) I've become increasingly skeptical of the whole Soviet-style 
"planned economy" Sun has implemented for Java. This often seems to lead 
to specifications that are cumbersome and difficult to understand while 
often failing to provide developers with the support they could really 
use. I think open source is a better path for the evolution of Java 
technologies. Seeing what's become of StAX just reinforces that - I 
don't see where it adds anything that's really useful to XMLPull, and it 
actually takes some things away. I recently had it pointed out that it 
doesn't even provide access to the declared namespace stack, for instance.

Are there any plans to add CharSequence to the mobile platform? I'd 
think that the ability to dramatically reduce the garbage generation in 
a parse would be even more valuable on mobile systems than on standard 
ones. That might justify eventually supporting this in kXML as well as 
XPP3. Meanwhile, an XPP3 implementation could also make use of the NIO 
code for even better performance, since that goes hand-in-hand with 
CharSequence.

BTW, I don't think CharSequence is ideal - they probably should have 
included the toArray() and getChars() methods in the interface (and in 
CharBuffer). But living in a centrally planned economy I suppose we 
learn to live with what we're given...

  - Dennis

-- 
Dennis M. Sosnoski
Enterprise Java, XML, and Web Services
Training and Consulting
http://www.sosnoski.com
Redmond, WA  425.885.7197




 
Yahoo! Groups Links

<*> To visit your group on the web, go to:
     http://groups.yahoo.com/group/xmlpull-dev/

<*> To unsubscribe from this group, send an email to:
     [email protected]

<*> Your use of Yahoo! Groups is subject to:
     http://docs.yahoo.com/info/terms/