Re: [xml-dev] SGML DTDs for HTML 5.1

[u123724 <u123724@gmail.com>]

> If you need to drop the multiple character once I'd just drop those rather than
> give them different definitions which leads to silent corruption.

Ok point taken. But what do you think of the couple entities that are
just variation sequences of their base code points such as caps/cups
(intersection/union) etc. These are pretty commonly used, and leaving
out U+FE00 VARIATION SELECTOR-1 on those, for all intents and
purposes, is equivalent to their use in variation sequences.

> That's a PoV I don't share!
Ok I'll just leave MathML in there. Subset customization works just
fine as it is, so no need to do anything about it.

Btw, do you know a recent SHORTREF map to use with MathML 3 (probably no ;).
From the MathML site/materials, I get that pre-MathML 1, there was an
inofficial SHORTREF map in use, but of course it was taken away b/c
XML.

I'm asking because years ago (I think it was around 2003) I had to
prepare a handout for a course I held, and remember me wanting to use
some math shorthand; back then I figured I could use pre-MathML
shortrefs, and came up with a notion of "infix" tags to be used in an
operator-precedence grammar (in addition to SGML's SHORTREFs); anyway,
it think it would be cool to have a generally useable shortref map for
casual math once again.

On Fri, Nov 18, 2016 at 6:15 PM, David Carlisle <d.p.carlisle@gmail.com> wrote:
> On 18 November 2016 at 16:40, u123724 <u123724@gmail.com> wrote:
>> Yes, I think you're right and I should leave the entities alone.
>> Thanks for the link.
>>
>> My motivation for removing some entities for the permissive DTD is
>> that the SGML declaration (WebSGML/ISO 8879 Annex K) only allows a
>> single code point to be specified as the replacement for an entity,
>
> for predefined ones does it, must admit it's been a while since I looked at
> annex K:-)
>
>> while HTML's entities also contain some multi-code point sequence
>> (as detailed in
>> http://sgmljs.net/docs/html5.html#html5-named-character-references).
>
> yes I know:-)
>
>> And I really want to align with WebSGML here, as for my application,
>> having to retrieve entity sets over the net kills it performance-wise.
>
> If you need to drop the multiple character once I'd just drop those rather than
> give them different definitions which leads to silent corruption.
>
>> Thus I was thinking that since most of the entities come from MathML,
>> and MathML doesn't really belong in a group next to HTML and SVG
>> anyway from a browser implementation PoV anyway,
>
> That's a PoV I don't share!
>
>>  I could conveniently drop some entities.
>
> As I say dropping them would be better than giving them different definitions.
>
> David