[
Lists Home |
Date Index |
Thread Index
]
At 12:40 PM 6/11/2002 -0400, Mike Champion wrote:
>So, I take it your approach is to have the application use
>pattern-matching to
>extract the information within an element/attribute value into the form
>you wish to process it? For example let the author type the date in any
>convenient format, then use something like the Javascript Date.parse()
>method (sorry, I'm a DOM/Javascript weenie by default ...I guess you would
>use regular fragmentations and SAX) to make it useful by the application???
Pretty much. I think it's fair to say that my approach is more focused on
local understandings of information than on building the one true global
type and expecting the apps to magically take care of it. There are lots
of ways to get there.
>Hmm, I guess that makes sense a an option; rather than have the GUI
>normalize it, or some post-processor normalize it (making it vulnerable
>to being broken by somebody "fixing" the XML with Notepad), just
>do the "normalization" in the application. I'm personally not crazy about
>it (since I tend to use XPath and/or XSLT to locate parts of a document
>or a collection of docs to process, and XPath doesnt' know about the private
>understanding between author and app writer on what the acceptable formats
>for a date might be).
I think the part that's most important is "private understanding", and how
we share those. Looking at information exchange as a transfer of private
understandings rather than conformance to a public understanding changes my
perspective substantially.
(It also changes the whole way I look at "options" and whether they provide
a foundation for better communications - things that describe local
understandings of types - or worse - things like all the external info in
an XML 1.0 doc disappearing thanks to a non-validating parser.)
At the end of my XML 2001 presentation on Regular Fragmentations, someone
asked if it could be obsoleted by adding regular expression support to XSLT
content-processing. I said I hoped it might be.
>Still, I'm sure we agree on the fundamental point: there is a lot of territory
>between helplessly dealing with <date>June 11, 2002</date> as
>a string that requires human intervention to process, and drinking the
>Schema koolaid
>in order to automatically bind it to a Date object.
Yep, definitely.
Simon St.Laurent
"Every day in every way I'm getting better and better." - Emile Coue
|