[MathML4] Removing some deprecated attributes

Previous Topic Next Topic
 
classic Classic list List threaded Threaded
4 messages Options
Reply | Threaded
Open this post in threaded view
|

[MathML4] Removing some deprecated attributes

Frédéric Wang-2
Hi,

The following attributes are deprecated in MathML3. Some of them are
still implemented in Gecko and WebKit. If they are no longer used, it
would be good to do some clean up by removing the support. Also they
should probably be removed from a specification focusing on core features.

About deprecated features, the recommendation says that "However, all
(MathML-input-conformant) tools are encouraged to support the old forms
as much as possible.". This is bad because supported features encourages
authors to use them.

1) general attribute: other.

2) math attributes: macros, mode. mode is supported in Gecko.

3) mglyph attributes: fontfamily, index.

4) token elements attributes: fontfamily, fontweight, fontstyle,
fontsize, color, background. These are supported in Gecko and WebKit
(although the latter does not make them of lower priority than thei
non-deprecated replacements).

5) deprecated namedspace attributes: veryverythinmathspace,
verythinmathspace, thinmathspace, mediummathspace, thickmathspace,
verythickmathspace, veryverythickmathspace. This used to be implemented
in Gecko but is now removed.

Frédéric


Reply | Threaded
Open this post in threaded view
|

Re: [MathML4] Removing some deprecated attributes

David Carlisle
On 03/08/2016 09:07, Frédéric Wang wrote:
> About deprecated features, the recommendation says that "However, all
> (MathML-input-conformant) tools are encouraged to support the old forms
> as much as possible.". This is bad because supported features encourages
> authors to use them.

Not sure I agree with the claim that this is bad as a general rule.
Certainly in html there are many things deprecated and classed as
invalid in html(5) where the html5 renderer is still specified to
support the old behaviour <a name="xx" etc.

Although for any particular case it's not unreasonable for a renderer
(or a class of renderers such as
mathml-systems-embedded-in-a-css-based-html5-browswer) to drop some
deprecated features if supporting them is costly and they are in
practice not used. We just, in the end, need to specify whatever profile
of MathML is supposed to work for such a class
of systems (and how the MathML interacts with CSS)

David

________________________________


The Numerical Algorithms Group Ltd is a company registered in England and Wales with company number 1249803. The registered office is:

Wilkinson House, Jordan Hill Road, Oxford OX2 8DR, United Kingdom.



This e-mail has been scanned for all viruses by Microsoft Office 365.

________________________________

Reply | Threaded
Open this post in threaded view
|

Re: [MathML4] Removing some deprecated attributes

Frédéric WANG
Le 03/08/2016 à 10:50, David Carlisle a écrit :
> Not sure I agree with the claim that this is bad as a general rule.
> Certainly in html there are many things deprecated and classed as
> invalid in html(5) where the html5 renderer is still specified to
> support the old behaviour <a name="xx" etc.
True, however HTML5 old features were still kept to preserve
compatibility with existing browser implementations and documents
available on the web. Most of the attributes listed here were never
implemented in browsers and/or have very low (inexistent?) usage.


signature.asc (836 bytes) Download Attachment
Reply | Threaded
Open this post in threaded view
|

Re: [MathML4] Removing some deprecated attributes

William F Hammond
Frédéric WANG <[hidden email]> writes:

> Le 03/08/2016 à 10:50, David Carlisle a écrit :
>> Not sure I agree with the claim that this is bad as a general rule.
>> Certainly in html there are many things deprecated and classed as
>> invalid in html(5) where the html5 renderer is still specified to
>> support the old behaviour <a name="xx" etc.
>
> True, however HTML5 old features were still kept to preserve
> compatibility with existing browser implementations and documents
> available on the web. Most of the attributes listed here were never
> implemented in browsers and/or have very low (inexistent?) usage.

I suppose pulling support for 'never implemented' is OK.
Low usage is quite different.  If any document on the web
uses it and support is pulled, then that document is broken.
Hasn't Tim Berners-Lee said that one should not do this (assuming
the document was correct when posted)?

Also: how does one determine low usage?  There can be large archives
unknown to a given user and not found in search engines.

                                    -- Bill