Re: minor suggestions on Public Editor's draft of MathML

On 16/05/2012 20:34, Joe Java wrote:
> Nit-picking at the latest MATHML editors draft
>
>
> 1.  Some of the examples appear to use theHyphen-minus (-)
> instead of the− (−) character to represent the minus
> sign.  For an example see "3.5.5.8 MathML representation of an
> alignment example" in the editors draft.
>
> Section "7.7.1.1 Minus" does say that MathML renderers should treat
> Hyphen-minus as a minus sign,

Yes and no, I think it's no bad thing to show that + and - can be types
just on an ASCII keyboard, and also of course using - gives a little
test of your renderer, using " & m i n u s ;" makes it too easy:-)
>
> but perhaps the defining document of MathML should use the correct
> character defined as the minus sign.

Perhaps or perhaps not. I could be persuaded either way but I'm inclined
to stick with - but this reply is a personal response the WG might
agree. It would be simple enough to change the sources. Perhaps we
should change some of them...

>
> 2. The examples use the Unicode number followed by a commented out
> long Unicode name for non-ASCII characters. Now that the "XML Entity
> Definitions for Characters" is out and the latest HTML5 draft has a
> table of all the character entity names, it might be better to only
> have the entity names used rather than the Unicode number of the
> non-ASCII characters/entities.

I wondered about that in the html version but....
For older browsers (including current Opera unless you have a test
version I think) that do not pre-define the full html5 entity set the
characters would be mis-parsed if I used entity references.

Also it's technically easier for me to use the expanded form (although
that's no excuse really and I could change this if necessary) but new
html5 version is actually a post-processed version of the xhtml version
and that has to have the numeric references rather than entity
references, so the entity names (which are used in the source) are no
longer available. I could of course change the pipeline to preserve the
names or I could just pick one of the entity names for that character.
But it may still be a bit early given browser support for html5 parsing.

The xhtml version has to use numeric references because browsers do not
fetch external dtd so it would not be well formed otherwise. the html5
spec could say that xhtml documents should be parsed with a catalog that
preloads the entities. I have an open bug on html5 requesting that, but
that is the current situation.

https://www.w3.org/Bugs/Public/show_bug.cgi?id=13409

(but I think the bugs database is down for a while)

>
> 3.  The document HTML5 + MATHML
> (http://www.w3.org/Math/draft-spec/mathml.html) does not validate
> with the W3C validator.

It validates with the validator.nu based validator at the w3c except
that the public instance of that currently uses a mathml2 schema for its
math fragment so the mathml3 features don't validate. I pinged the
maintainers of that earlier in the week and there is a development
version updates the mathml schema used so this will be fixed once that's
ready to roll out. But thanks for checking:-)

http://validator.w3.org/nu/

>
> It would be easy to fix this problem, but the documents' HTML code
> is very poorly indented and spaced.

>
> For example see line 24605 of the HTML5 + MATHML document.
Ah the index, yes that's mechanically generated of course I could put
newlines rather than spaces between entries. It's never bothered me, but
I suppose some editors (or people) get a bit stressed by 56737 character
lines:-)
>
> Any decent code editor can easily assist in proper indentation and
> line spacing.

Yes but that file as served there is never touched by an editor or
manually changed in any way. I think that's an important property to
keep when publishing in multiple formats.
>
> Joe
>
Thanks for your comments, I'll see what I can do about the formatting.

David

Received on Wednesday, 16 May 2012 20:30:41 UTC