Re: draft HTML5: Techniques for the provision of text alternatives

Laura Carlson:
> Hi Steve,
>
> I agree with Olaf that the Lady of Shalott examples may be problematic.
>

The example itself is pretty good, especially because it raises non trivial
markup and semantical questions - and related to the old img+alt method
interesting questions too.
Publishers, who have to care both about semantics, accessibility and authors
rights have to face these problems anyway.

And there should be at least one expert for HTML, another for RDF(a),
yet another for accessibility and maybe one for text/literature/semantics  
in the HyperText Markup Language Working Group to
markup such a real world example together properly ;o)


> Poetry markup is an important yet separate issue that Olaf may want to
> raise a bug [1] [2] [3] [4] to resolve.
>

Well several members of the working group convinced me in the
previous discussions, that HTML5 (HyperText Markup Language Vesion 5)
is not the right place for poetry, literature or text, it cares about other
things (I did not ask or look into details, about what else it cares) 
- therefore it was suggested to use another language to markup
text or to use RDF(a) and a specific vocabulary. Because possible fragments
of vocabulary were spread around, I created LML (literature markup language)
to collect the vocabulary in one compact file and to solve the problem to
markup text properly. I think, it can be used in (X)HTML+RDFa or with
(X)HTML and role indications, as already available in SVG tiny 1.2.
Or authors may embed LML (or another language to markup literature)
directly within (X)HTML as it is proposed to embed vector graphics with 
SVG in HTML5 or formulas with MathML.

However, if the working group changes their ideas about what the
HyperText Markup Language Vesion 5 is - text or not - LML may be
a good starting point and overview about different solutions in other
languages how to markup text ;o) 
In this case, this is more than a bug.


> Until the poetry markup issue is decided, perhaps for 6.1 and 6.2 we
> could use something like we did for our May 2008 HTMLWG Action 54 [5]:
>
> <!-- Full Recitation of Alfred, Lord Tennyson's Poem -->
>
> Then the text alternative issue [6] won't be intertwined with the
> poetry markup issue [7].
>

I think, it appears more often, that images and text/literature are combined
somehow.
This is, what some authors do. If everything is separated in simple and 
trivial examples, authors will still do not know how to solve such non
trivial real world problems properly.  And looking on how they did it
up to now, indeed most did not solve the problem, maybe due to some
limitations of the old HTML and maybe due to personal limitations.
Again, the example is pretty good, but needs some more efforts to be really
helpful for authors ;o) 
I have an art and literature gallery on my own and had already
discussions about accessibility of abstract artworks - well, not trivial,
I switched from raster images to SVG already years ago and  from
img to object to get it somehow better, mainly because SVG has much
more options to solve such problems than raster images and HTML.
Of course, for such reproductions of classical, almost non abstract paintings
raster image are still a compact and effective solution, therefore one has
to face such samples anyway in HTML - and one cannot expect that
many authors will embed the raster images in SVG to be able to
describe it properly and then embedding this again in HTML to work around
the limitations of the old fashioned HTML. As far as I understand, a similar
approach was proposed for video and audio  (that the referenced file 
contains the accessibility features and maybe meta information). Well,
looking on what authors currently really do, I think, this is nothing, what
will really work (embedding video, audio, images in SMIL or SVG tiny 1.2
to add meta information and alternative text/presentations? and then
embedding SMIL or SVG in HTML again?)

Olaf

Received on Tuesday, 19 January 2010 10:21:51 UTC