Re: Continued development of XMLPull

Dennis Sosnoski <[email protected]> Fri, 14 May 2004 22:57:36 -0700
Newsgroups gmane.text.xml.xmlpull.devel
Message-ID <[email protected]>
Aleksander Slominski wrote:

>Dennis Sosnoski wrote:
>
>>I'd think an XMLPull 2 API that made use of the CharSequence approach 
>>(not only for getText(), but also for things like attribute values) 
>>would be a great next step. 
>>
>>    
>>
>i think this can be done first as an addon and maybe have this permanent 
>and required part of API when hitting XmlPull2
>
That'd probably work. I may take a look at how much performance can be 
gained by going this way. I've got an internal parse event stream 
representation of documents (currently going to and from SAX2, because I 
wanted compatibility with other tools) that I could easily modify to use 
the CharSequences.

>>I'd also love to see an XML processing API 
>>that actually showed some sense in the use of names and namespaces 
>>(using QName and Namespace objects, rather than the damned text strings 
>>that everybody has to process and reprocess). 
>>
>>    
>>
>:) i came to the same conclusion in XB1 though i only use XmlNamespace 
>(it can have prefix though)
>
Yes, this has been a major screw-up in XML API design since namespaces 
got pasted on top of XML. I hate to think how many separate 
implementations of namespace tracking code have been written - I'd 
suspect it's in the dozens, at least.

>>I'd expect that an API of 
>>this type would be faster than the current low-level API while also 
>>reducing overhead for the applications that build on top of the parser.
>>
>> 
>>
>>    
>>
>agreed.
>
>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 ...
>
>what you think?
>
>  
>
I'm not all that big on StAX, given the glacial progress, but if it 
doesn't hurt the size and performance of using the base code there's not 
a problem. Can you layer it so that the StAX API uses separate classes? 
That way different build targets could build just the base XMLPull (+ 
the changes we're discussing) API or the full StAX implementation.

  - Dennis



 
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/