Re: warnings from Xeena: dagger multipli-defined

David Carlisle <davidc@nag.co.uk> writes in response to Hugh Devlin:

>                   At present it is necessary for the Math WG to produce
> these ISO entity files but one assumes that there will eventually be a
> standard set (perhaps by adopting the MathML ones).

Has ISO been approached?

> It so happens that the names you mention [dD]agger are contained in two
> entity sets (isopub and isoamsb) and so if you load both of these sets
> then the entity is defined twice. The same definition is used both
> times, and even if that were not the case it would not be an XML error
> (the second would be ignored) however if applications are going to
> trouble users with these warnings a pragmatic solution might be to drop
> the symbols from (say) isopub.ent and document somewhere that if a
> document type wants to use ISOPub without ISOAMSA then it will need to
> additionally define the dagger symbols.

(Perhaps the xml definition could use a "precedence level" in connection
with entity set inclusions.)

Since these two files have origin in other different places (ISO and
AMS, is this right, or is the duplication at ISO?), you are reluctant
to make an arbitrary choice.  Is this right?  Possible slight further
complication for xhtml+mathml: I find the following in the file
"HTMLspec.ent" that accompanies version 1.3 (now a bit old) of SP:

<!ENTITY dagger  CDATA "&#8224;" -- dagger, U+2020 ISOpub -->
<!ENTITY Dagger  CDATA "&#8225;" -- double dagger, U+2021 ISOpub -->

Without having checked I trust that the definitions are, at least,
consistent since this appear to be taken from ISOpub.

While the error messages can be ignored if the definitions are
consistent, it does come down to getting things right, and it is an
issue somewhat similar to that of avoiding the ambiguous content
model for elements.

                                  -- Bill

Received on Thursday, 28 October 1999 10:04:43 UTC