Re: Rough sketch for an I-D (a successor of RFC 3023)

Henry Thompson writes:
>I strongly believe that we need a frag-id syntax for */xml, and that it
>should be as recommended by the XLink WG as was [1].  I will argue the
>point in whatever forum is agreed to be appropriate.
>
>[1] http://www.w3.org/TR/xml-fragid/#d0e135

I am no longer an editor of RFC 3023 or its successors, but I will argue
against Henry's position in whatever forum he chooses to present it. 

XPointer is a misbegotten beast, a Recommendation with virtually no
popular support and with deep flaws in its basic scheme-based
infrastructure. It has, unfortunately, snuck through the W3C process,
apparently approved out of mercy so that its Working Group could finally
retire.  

There were substantial late changes at PR that I do not think should
have been permitted at all under purported W3C process. That those
changes reflected hysterically broken decisions like requiring namespace
declarations (which themselves contain URIs) inside of URIs for any
scheme that didn't happen to come from the W3C doesn't help.  

The flaws are architectural, not merely syntactic, and lurk in the
XPointer Framework itself, making adoption of that document in any form
a dangerous precendent to future work.

XPointer should be laid to rest and deprecated rather than considered a
foundation of any further XML 'progress'.  I would strongly encourage
the TAG to avoid explicit reference to XPointer in any of its work.

Simon St.Laurent
http://simonstl.com

Received on Sunday, 2 November 2003 13:02:41 UTC