| ||||
order is importantthis is old joke but still funny and good conclusion:"OK, you're up," the Genie says to the manager. The manager says, "I want those two back in the office after lunch." created sat dec 28 2002 9:30am cet [2002/12/28 9:30 CET] permalink Grid, workflows, and other observationsBunch of observations:
created sun nov 3 2002 1:45pm cst [2002/11/3 13:45 CST] permalink why including XML APIs in JDK is bad ideaXML is not about APIs it is about processing XML. if XML was finished and there was no work on standards (DOM3 comes to mind) and implementations were finished and abandoned (Xerces2, Xalan etc.) than maybe it would make sense ... actually that brings another point: java badly needs library (JAR) dependency mechanism in use right now ... created sun oct 27 2002 cst [2002/10/27 CST] permalink XML API that is between DOM and SAXwhat is between DOM and SAX? of course KHR, KHS, LHR or LHS by looking up ASCII codes D=68 S=83 so average is 75.5= K=75 or L=76, then second letters O=79 A=65 so average 72=H and finally M=77 X=88 so average 82.5 R=82 or S=83 :-) created oct 9 2002 cst [2002/10/9 CST] permalink XML and APIsAPIs that are doing essentially the same things are bad ... Corba: why PL independent XML integration ... components as Web Services will help life easier ... cost of monolithic applications ... what is API XML schemas (what is target) SOAP acronym ... no Simple ... no Object ... so what is left ... CORBA has no abstract protocol created oct 9 2002 cst [2002/10/9 CST] permalink XML and abstractiontoo much abstraction is not good ... how to keep close to XML infoset? created oct 9 2002 cst [2002/10/9 CST] permalink streaming and incremental building of XML node treeHow to combine best of SAX and DOM when working on SOAP motivation DOM API (or one of its alternatives) is easier to use for document/literal encoding but streaming is essential for good performance ... solution: two different modes of operation
one goal: have uniform API for both cases (stream is parsed as much as needed to walk tree) second goal: expose enough information in API so app can control tree creation skipping parts of it or using XmlPullParser directly (no tree nodes created) parts of tree can be detached for later GC handling fragments transparently add all ancestro namespace-prefix declarations XmlPullNode
isFinished()
finishParsing()
XmlPullParser asPullParser
XmlPullNodeWalker asNodeWalker
XmlPullNodeWalker
Object nextNode(),
Object nextNodeChild(),
String nextNodeText(),
XmlPullNode nextNodeTag()
XmlPullParser asPullParser();
DOM working as front partial tree - mirrors what is in XmlPullNode however operations directly on XmlPullNode are _not_ mirrored to DOM!!! Possibly related: PullDOM http://www.prescod.net/python/pulldom.html created sept 20 2002 cst [2002/9/20 CST] permalink |
This blog is about: Find more
about
Blogroll:
Projects::
RSS
Filter Entries: |
Disclaimer: personal opinions and observations that may or may not be taken seriously, or even based on shared reality and generally are very unreliable and personal and snapshots of volatile writer mind ...
NOTE: THIS PAGE IS UNDER CONSTANT DEVELOPEMENT