Lists Home |
Date Index |
> Michael Kay wrote:
> >This is not an erratum. It is a change proposal.
> Since there is no corrigendum category at W3C, it is the only bucket.
Erratum and corrigendum have almost identical definitions in my
dictionary, so I don't understand the point you are making. (Both,
surprisingly to me, are defined as "errors" rather than "corrections of
errors", but this probably reflects the fact that they are intended to
refer to "typos" where the correction is obvious once the error is
In a very loose sense, you can say that if there is a usability problem
with a spec, then that is an error by the designers. But if W3C or
anyone else changed its specs whenever someone came up with a bright new
idea, no one would touch them with a bargepole. The world needs
stability. Retrospective legislation is a really bad thing, in this
sphere as in any other.
> I guess my view of what a change may be different from yours,
> because for the last couple of years almost all the XML
> programming I have had to do for our products involved making
> non-standard parsers, to help bring in information to let
> people make it WF or Valid.
I have no objection to anyone trying to find a solution to this problem,
only to solutions that break architectural interfaces. In my view the
right solution is something that sits between the character decoding and
the XML parsing, and that is context-free. Something a bit like \uuuu in
Java. It doesn't have to go into XML, it can be defined as a
preprocessor that takes any stream of Unicode characters as input and
produces a different stream of Unicode characters as output. And if it
doesn't need to be in XML, then it shouldn't be in XML.