Some thoughts on 'direct access' to XML

A long posting to XML-DEV following up on the original posting I made about
what 'direct access' to XML might mean from a programmer's perspective. This posting is strongly related to my last long post to xml-dev called
'the privilege of XML parsing' http://lists.xml.org/archives/xml-dev/200212/msg00277.html.

Spot the penguin

Make sure you set your browser window size to full screen after you click this. You could read the text or you could spot the penguin in the text. Its up to you.

yetAnotherProceduralProgrammingLanguageInPointiesWithColons

Isn't Len Bullard just totally priceless!

A technical, technical writer required

I have a stack of white papers for Propylon written. Unfortunately they are all in my head :-) I badly need someone who can work closely with me to get this stuff onto paper and into pictures as quckly as possible. I need someone who has a lot more than a glossary level understanding of XML, schema, workflow, pervasive computing, patterns, n-squared integration, MOM, RPC, SOAP/REST/XML-RPC.

Does such a person exist? Perhaps someone who does some IT consulting/programming but writes technical books also?

If you are such a person, or could suggest such a person, please contact me.

What does 'direct access' to XML data mean?

A post to XML-DEV about an interesting Adam Bosworth article has spawned a long thread in XML-DEV. In the article Adam Bosworth points out the clunkyness of SAX/DOM/XPath for accessing XML from, say, a piece of middle tier business logic. He is right of course but the intrigue surrounds what might be meant by a more 'direct' form of access?

Those of us who strongly dissent from the view that "XML is just another notation to store/transmit your objects in", always worry that 'direct access' might be constructed to means direct object serialization/de-serialization. I think that view of XML is a very misguided one. I believe it is the slippery slope to a tightly-coupled, process-specific, data-head-XML, stored-and-transmitted-in-some-'binary'-XML-form (necessary-for-speed-ya-know), all-you-need-is-SOAP, REST-is-a-back-fill-academic-put-on,-N-Squared integration-haven't-we-been-here-before-come-back-CORBA-all-is-forgiven NIGHTMARE.

I try to read Adam Bosworth's stuff as often as I can because he is generally right on the money. I quote him more than anyone else in my day to day work. I sure hope the immediate leap that some might make about the phrase 'direct access' proves to be dead wrong!

The digital dark age

An article well worth reading from the Guardian The digital dark age (thanks for the pointer, Pat).

XML is, of course, not the answer in itself. To stand the raveges of time as an encoding for knowledge, the markup has to be about the knowledge - not about the needs of the application used to edit the knowledge. Examples include XML formats of Wordprocessors, typesetting systems etc. It is good that they are XML from an EAI perspective, but the XML-ness of those formats does not separate your knowledge from the (short lived) application that created it.

Even with properly semantic, open XML encoding, you still have the problem of the digital media encoding - the way it encodes 1's and 0's. It is probably even more prone to the raveges of time than the knowledge encoding.

I remember many years ago, Eliot Kimber pondering how best to ensure that the HyTime standard would still be readable (admittedly in baroque 20 century English) hundreds of years from now. His solution involved acid free paper as the version of last resort!

X# anyone?

I did an interview with searchwebservices.com speculating what X# might look like. Total speculation of course but fun to think about nonetheless. Spawned a thread on slashdot to boot.

Spreadsheets and XML

Since I wrote this piece for ITWorld a number of people have suggested that Excel (XP?) already has very good Excel support. I'm not a big user of XP (or Excel) so I'm not sure how extensive it is (and resources about it have not easily come to hand). However, the main point in the article is that people whose opinions I respect came back from XML2002 raving about the XML support in Office11 - DOC + XLS.

HL-7 and XML

A presentation I have last year at a Health Informatics conference about the relationship between HL-7 and XML. Its in PDF (17 pages). Summary: XML is a syntax. HL-7 is a meta-model. Neither give you interoperability by some magic causal relationship between the acronym and the goal of "interoperability". Interop is only partly a technology problem. The rest of it - and by far the more complex problem - is a human problem - trashing out agreements between parties as to who/what/where/when/how data will be interchanged.