Re: XML 1.1 backwards compatibility

> The only backwards incompatibility in XML 1.1 is the fact that we
> don't allow the C1 control characters, right? And we only removed
> them because for a brief period we allowed the C0 control characters,
> right? But then we removed the C0 control characters, right?

XML 1.0 allows literal C1 controls, but no C0 controls except TAB, CR
and LF, even as character references.

XML 1.1 allows references to both C0 (except NUL) and C1 controls,
but disallows literal C1 controls (except NEL).

So XML 1.1 added references to C0 controls, and removed literal C1
controls.  To some extent, requiring references for C1 controls was a
quid pro quo for allowing references to C0 controls.  An explicit
reason was to catch encoding errors (particularly Microsoft
proprietary encodings mislabelled as iso-8859-1), but there were also
clearly two camps: one favouring more rigorous text-only enforcement
and the other favouring easier inclusion of non-text characters.

I don't recall us ever allowing literal C0s.

> So what happens if we erratum/second edition the 1.1 spec so that
> it allows the C1 control characters?

We outrage the people who have already had to fix things once?

-- Richard

Received on Friday, 18 February 2005 15:54:47 UTC