[Blindmath] n-lab questions

[Andrew Stacey]

Gosh, so many answers and I haven't even asked a question yet!  I'd better do
so quickly to avoid breaking the 3rd law of L-space ("Do not interfere with
the nature of causality.").

My first barrage of questions are to do with the n-lab, and the instiki
program that runs it.  In short, how do we rate as regards accessibility?

Some people have already started answering that, which is great, and I've also
read back a little in the archives so you can assume I'm familiar with your
recent discussions.  In particular, it seems that we score highly because we
serve MathML rather than inline images or some other hybrid system.  But I'm
sure that accessibility is about more than just the doctype!

First, a little more background to the project.  The n-lab, as I said earlier,
is like an open lab book.  So it's primary focus is to help a group of people
(mainly mathematicians) to do their research.  As such, it's a bit like any
lab book: some coherent notes, some write-ups of seminars, some distilled
wisdom from the literature, and a lot of scribblings.  The key difference is
that it happens in the open so that anyone can stop by, read our scribblings,
and offer advice or get ideas themselves.  In short, we think that "open
source" should apply to mathematics as well as programming.  So for us,
accessibility is important because it means that our scribblings can be read
by as many as possible and thus we can learn from as many people as possible.

A little on the technical side: instiki is written in ruby-on-rails and serves
XHTML with MathML and SVG.  The input format is a hybrid of extended markdown
syntax with itex2mml as the mathematical engine.  Other mathematical engines,
such as blahtex, are possible but we use itex2mml.  The main maintainer of
instiki is also the main maintainer of itex2mml, Jacques Distler.  Whilst not
directly involved in the content of the n-lab, he keeps a close eye on how
we're doing as we're one of the largest examples of his software.

Now on to my questions.  Let me divide them into three categories.


1. The underlying program, instiki.  This is where I am least involved so
don't really know how to phrase a good question.  I know that Jacques Distler
is keen for Instiki to do all it can to be accessible and he would like to
hear of any ideas that anyone has to do that.  Either you can contact him
directly, or I can pass along any comments anyone has.

2. The overall feel of the n-lab.  Certain aspects of the n-lab website are
globally controlled, such as the main CSS and the navigation that is on each
page.  We've been thinking that the style of the n-lab is a little off-putting
- it's a little stark, to say the least!  How do we rate in terms of
  accessibility at the moment?  And if we start playing around with the style,
are there any words of advice you can give us (or point us to) to ensure that
we are more accessible?  In particular, is it safe to simply modify the CSS,
on the grounds that CSS can always be disabled from the client end?  Are there
any obvious things to avoid (I've heard that tables for layout are a big no-no
from an accessibility point of view).  One other thing that springs to mind
are the page names.  At the start, we had the convention that page names
should be drawn from a basic character set, so no fancy unicode characters
like "infinity".  Is this useful, or completely indifferent?  How do screen
readers cope with the range of unicode characters?

3. The content of each page.  There are many contributors to the n-lab and we
don't like to impose rules and regulations but we do feel able to offer
guidance and we do tend to keep the style reasonably homogeneous.  Is there
a simple list of things that we can ask authors to keep in mind as they write?
We do have a group of self-styled "lab elves" who go around cleaning up pages,
and who take more notice of these style conventions.

I also use instiki for a course that I'm teaching, but I think that there's
nothing particular for that which wouldn't be relevant to the n-lab.  We also
have a forum attached to the n-lab.  I'd imagine that that scores very low on
accessibility, but I'm less worried about that; partly because that's for
internal discussions rather than to be read by casual visitors, and also
because I'd like to redesign the software for that from the ground up to make
it easier to integrate with the n-lab proper and it would be easier to put in
accessibility then rather than try to impose it on a piece of software that
I don't really understand the deep workings of.

Some relevant links:

The n-lab itself: http://ncatlab.org/nlab/show/HomePage

Instiki: http://golem.ph.utexas.edu/wiki/instiki/show/HomePage

Jacques Distler's blog: http://golem.ph.utexas.edu/~distler/blog/

Thanks,

Andrew Stacey