Re: XML namespaces on the Web

Leif Halvard Silli scripsit:

> John Cowan claimed there to be (I think) 22 fatal error categories in
> XML. And an not easy to count amount of -in theory- not fatal errors
> - which still, according to John, are treated as fatal errors in all
> known XML parsers.

I did a careful investigation of XML 1.x MUSTard.  After removing the
MUSTs that refer processor behavior rather than document content, the
redundant MUSTs in well-formedness constraints, the MUSTs in validity
constraints which apply only to validating parsers, and so on, I wound
up with only the four following errors that are technically non-fatal:

    A reference is made to an entity that is undefined in the document,
    but *might* be defined in a part of the DTD that the parser is not
    processing, typically the external DTD subset.

    A document is in UTF-16 encoding but does not begin with a BOM.

    A document or external entity is not in UTF-8 or UTF-16 encoding,
    the encoding is not externally specified (as by a MIME header),
    and there is no encoding declaration.

    A built-in entity has an incorrect entity declaration.

I may have overlooked or misread some others, but not many.

> It is not clear to me how most in the XML community want to keep it -
> whether they want to keep all not fatal errors as fatal, or if they
> are open to discern between the kinds of errors.

I think XML folks are just as happy to treat these as fatal.

> One should think that A) browser vendors could have a look at the
> countless non-fatal errors an make sure that they are not treated as
> fatal.

I think only the first of these is of practical importance.

-- 
John Cowan  cowan@ccil.org  http://ccil.org/~cowan
If I have seen farther than others, it is because I am surrounded by dwarves.
        --Murray Gell-Mann

Received on Thursday, 19 November 2009 17:19:22 UTC