Re: Generating Chemistry Examples

On Thu, 03 Dec 2015 00:05:36 +1000, Fred Esch <fesch@us.ibm.com> wrote:

> Volker,
>
>> However, what I can not see yet, is how I could represent my current
>> navigation model in this format. In particular I struggle to understand
>> -- how elements can be shared between groups (e.g, we have R1-O-R2 where
>
>> both groups share O and I want to navigate from [R1-O]-R2 to R1-[0-R2],
>> if that makes sense).
>
> This is an interesting use case that has parallels in chord charts.  
> Without connectors, I don't know how you can solve it.

With chord charts it seems reasonable to solve it with a link from each  
side, and a bit of SVG animation.

But that's because the two links can be seperated easily enough. For  
overlapping chemistry nodes, the most obvious approach I can think of is  
to actually make two groups, which means in this instance you have a  
second oxygen node, and that gets messy:

g@role=listitem: R1 - linked to - O, the central oxygen atom
g@role=listitem: O, the central oxygen atom, linked to R2

But there are two elements that are O. Navigation of that might become  
tricky if you want to allow exploring the content of each group, and  
having the links from each group.

Although maybe it would just work, if you ge tthe linking right, and the  
two rendered O atoms are in fact genuine clones of each other so you can't  
tell which you are looking at.

Until you try to extract the underlying chemistry…

Hmmm.

cheers


>
>     Regards,
>    Fred Esch
>  Watson, IBM, W3C
>   Accessibility
> IBM Watson       Watson Release Management and Quality
>
>
>
>
>
>
> From: "Chaals McCathie Nevile" <chaals@yandex-team.ru>
> To: "Volker Sorge" <v.sorge@progressiveaccess.com>
> Cc: "SVG-A11y TF" <public-svg-a11y@w3.org>
> Date: 12/02/2015 06:10 AM
> Subject: Re: Generating Chemistry Examples
>
>
>
> On Wed, 02 Dec 2015 20:30:21 +1000, Volker Sorge
> <v.sorge@progressiveaccess.com> wrote:
>
>> Dear Chaals,
>>
>> Thank you for those links. I had a look and I can see that it should be
>> quite straight forward to generate something similar automatically from
>> my molecule examples.
>
> Whee! Seriously, that's nice to hear... although we of course need to  
> work
>
> through the details…
>
>> However, what I can not see yet, is how I could represent my current
>> navigation model in this format. In particular I struggle to understand
>> -- how elements can be shared between groups (e.g, we have R1-O-R2 where
>
>> both groups share O and I want to navigate from [R1-O]-R2 to R1-[0-R2],
>> if that makes sense).
>
> I don't currently have a way to do that :( It might be possible using
> aria, but of course that only works with screenreaders - and even then,  
> in
>
> quite a limited way at the moment ;(
>
>> -- how hierarchical navigation is supported.
>
> Technically, it isn't.
>
> In the composed-tree example I have sort of faked it - tabindex gives a
> "breadth-first" navigation of a tree, but you can follow a link to jump
> down to the next level. Unfortunately going back is only possible for
> links - if you moved a bit with tabindex, you don't get more history.
>
> It would be possible to hack that by adding stuff to the history tree,  
> but
>
> then you make it very painful to use the back button for going back
> completely - but this isn't a problem that is even unique to SVG let  
> alone
>
> accessibility :(
>
> I have also used role=list / role="listitem" to describe the hierarchy -
> more thoroughly in composed-tree than in the chem-BV-ox example. Again,
> *IFF* you have a screenreader you can navigate that structure, albeit
> still not very effectively.
>
> The fastest answer seems to me a treewalking extension for a browser -  
> but
>
> this may have some interesting compatibility problems with sites that
> think they are offering helpful shortcuts.
>
> The most sensible answer would be to revise navigation fairly thoroughly,
> but of course that will take some serious implementation work, and isn't
> going to be quick ;(
>
> cheers
>
> Chaals
>
>> Best,
>> Volker
>>
>>
>>
>> On 2015-11-29 20:07, Chaals McCathie Nevile wrote:
>>> Dear Volker,
>>>  On Fri, 27 Nov 2015 23:22:15 +1000, Volker Sorge
>>> <v.sorge@progressiveaccess.com> wrote:
>>>
>>>> My apologies for never engaging again after I had shared the
>>>> navigatable  chemical diagrams initially. I was simply too busy this
>>>> year. But I  tried to keep an eye on the list.
>>>  And my apologies for not engaging more directly with the stuff that  
>>> you
>>> showed - for the same reason.
>>>
>>>> Now that there is a draft spec available, I was wondering if there is
>
>>>> any interest in programmatically generating molecule diagram examples
>
>>>> for the spec. If there is, I can have a look into that. (I can't
>>>> promise  anything with respect to delivery time though!)
>>>  I'm interested - I've been working with a habdful of techniques that
>>> *sort
>>> of* work now - if you have the right browser and a bit of luck, as well
>
>>> as
>>> getting the code right.
>>>
>>>> I am currently also looking at diagrams from Biology (phylogenetic
>>>> trees) and Physics (circuit diagrams) that might serve as useful
>>>> examples in the future.
>>>  Both of those would be great kinds of examples. I mostly got together  
>>> a
>>> simple tree diagram (there are a couple of bugs, but there is enough
>>> right
>>> to show the idea) with some notes on what I did at
>>> svg-access-w3cg.github.io/use-case-examples/composed-tree-notes.html
>>>  I have a diagram of a reaction process I am also working on, but it's  
>>> a
>>> mess really - I need to spend a day or so with it so it shows at least
>>> what I am trying to achieve. I'll put something up tomorrow at
>>> svg-access-w3cg.github.io/use-case-examples/chem-BV-ox-notes.html to
>>> explain what I am thinking about it…
>>>  cheers
>>>  Chaals
>>
>
>
> --
> Charles McCathie Nevile - web standards - CTO Office, Yandex
>   chaals@yandex-team.ru - - - Find more at http://yandex.com
>
>


-- 
Charles McCathie Nevile - web standards - CTO Office, Yandex
  chaals@yandex-team.ru - - - Find more at http://yandex.com

Received on Wednesday, 2 December 2015 17:42:39 UTC