See also: IRC log
<tzviya> http://www.w3.org/2015/11/23-dpub-minutes.html
...: "Comments on last week's minutes? ... Minutes approved."
<dauwhe> https://dauwhe.github.io/epub-zero/acme-publishing/
Dave: "As we talk about portable web
publications and browser-friendly formats, epub 3.1... I decided to make
samples..."
...: "I wanted to see if I could build a tiny reading system around some
of these ideas to see how they work on the wild web. The cool stuff is
do to jake - who put the service workers demo site after he talked to us
at TPAC. I took his code and used it to build this little website [link
above] so it's purely an experiment in 'can we make a web-friendly thing
work in a browser'"
<dauwhe> https://dauwhe.github.io/epub-zero/acme-publishing/MobyDick/html/c001.html
<pkra> it's going to be hard to go offline to show the cool stuff though.
...: "... screen sharing ..."
... " ... begins screen share ... scrolls down ... clicks on links..."
<pkra> if you're on Chrome, you could disable the connection for that tab in the developer console (under Network => Throttling)
...: " ... Begins reading moby dick... Loads nav file
for that book... Button in the corner 'make worky offline' it will
'offlinify' and locally cache everything that goes with the book for
offline reading. When complete it says 'offlinification complete'!!!
Link to a little Javascript function that build a header of prev/next
navigation for chapters"
... " ... Arrow keys are mapped to next/previous ... Styling of
background ... Can also download in a zip-like format ... "
... "he built another page that lets you upload one of those zip files
and read them. Peter put in a journal article here that has lots of math
and such. MathJax is in there to render all the mathematics."
... "Not fancy, but functional."
<pkra> not really published but an introductory note by a high-profile mathematician (source is CC-by on the arXiv)
<Ralph> [very nice prototype/demo Dave!]
Ivan: "Looking into jake's code it looks at a zip file, can it work on a zip file and not just a directory?"
Dave: "This is jake's thing - that I could take one of the files I previously downloaded. "
Ivan: "The other question is more, on longer term, to understand the architecture - if my understanding is correct, the impression I have is that you have to add the contents, a reference to a javascript file (which does the service worker)... Is there a way somehow that the content does not have to include this reference."
Dave: "My knowledge of the Javascript is limited."
Ivan: "We'd have to go back to jake to see what's going on there."
Tzviya: "Jake said he'd be happy to answer questions, so we can ask him as well."
<daniel-weck> I'm happy to answer questions (email or during concall)
Ivan: "It's important, as we should be able to build independently of the content"
nick: what is the role of the SW in the demo?
dave: SW makes offline possible
nick: the JS can then be decoupled from
the content completely
... not sure if it's possible to read a completely zipped filed
... using decoupled JS
Daniel: "You do not have to download the entire zip - with Readium we stream the zip and deal with it on the fly"
Ivan: "At the moment, the example code that Jake gave us is such that ... the bootstrapping procedure for a book is a reference to a script in the book itself. To have the whole mechanism working itself, the bootstrapping goes through modifying the content. What you'd like to have is that it can be bootstrapped by something else in the client. Basically can it be loaded without modifying the source."
Dave: "I was on the queue to change the direction of the conversation. A major goal here is to figure out how to make a file format. Not so much the details of how to get the service workers going. I just wanted to show that it at least works for some of our use cases. More interestingly, what implications does this have for the structure of the documents we want."
Ivan: "The goal is that the same content that it can be displayed in a traditional way. What that would require is that the bootstrapping should not go through the content."
Tzviya: "Lets use the email list to continue this discussion
Peter" "It is not yet published... The background is that MathJax itself is pretty old. In 2010 was version 1.0 and design started a year before that. We've been facing a problem that we need to revamp the internals, but the internal plumbing hasn't changed much. We're actually at a place where we can finally re-do some of the plumbing to make it better. For us it is the right time to consider the inner workings. Not having to support legacy browsers, etc."
...: "One of the key considerations for us is that
MathJax itself is not a single-organization, it's a community driven
effort supported by publishers, societies, startups, etc. Originally in
2009 and 2010 was to help move math. Browsers weren't supporting it
because no one was using it, etc... About 5 years later it hasn't really
moved."
... "We considered a couple different paths and we reached out to get
feedback and eventually decided to go with one of them. A) do we hold on
to this goal to push forward with the native goal of MathML. Or do we
want to be a polyfill in the modern land - where a developer can just
drop things in."
... "Then we ran into the problem that it doesn't simplify things for us
- we ran into performance issues as well. The first idea was that
web-components isn't done yet and doesn't have full backing from the
browsers... We had alot of performance issues, inconsistency issues, and
technology bets. The alternative is to embrace the progress we've made
in the past two years and perfecting a CSS and SVG layout."
...: "We're going to work on making a CSS
picture-perfect display of MathML - so make things better because of
this. We also see that the CSS side - is that CSS has progressed
enormously. Things are much more stable and we can do them much better
than we could do before. "
... "You could say that instead of building a polyfill, we're trying to
build a CSS rendering of math. We want to push the standards to work for
us as well. Doing things with grid layout is a good place to start, and
it's the direction we're working towards."
Deborah: "I love the idea about native
CSS for math. My biggest concern is that whenever I'm talking to someone
about switching to MathML, They always say: 'the accessibility has never
been very good.' From the publisher, an image is the best for everything
except accessibility, the publishers tend to go for images. "
...: "As the move towards the new way, will there be a priority of
working with screen-reader makers to make it good enough for
accessibility."
Peter: "Yes, it's a huge issue and
something. If we want to improve CSS, more than anything we want to have
progress on the accessibility side. People seem to think MathML solves
accessibility - but it's not as good as we hope. That impression goes
back to when mathPlayer was the solution, but it's not there. You have
crappy in voiceover, in chromevox, etc..."
...: "We obviously think MathML is the right step to solving that
problem, but the standard needs to improve on that end. It's something
that any other approach needs to solve."
Ivan: "Whether the mapping and usage of CSS/HTML is it done on the client side or the server size? How is MathML appearing on the client side - if at all."
Peter: "One of the changes for our
perspective, is considering server-side rendering as a first-class
output. One advantage (which is already happening) is that you can now
generate the CSS layout on the server, the SVG output can also be
generated on the server. But how can you put the MathML out there -
well, you don't you replace it."
...: "That was the first idea of doing a polyfill - hide all the output
and hope one-day it has a rendering. But we're embracing the idea that
we won't have the MathML source. We also can do the MathML for a screen
reader. We want -for accessibilty - is to have an accessible rendering
of the content. The existing way is to split the accessible rendering
from the regular - which isn't very accessible."
Ivan: "You noted you wanted to address the standards side - what should be added/improved/changed from the MathML side to make things smoother. I think it would be really important to have those things clearly described somewhere. If they are, we can try to move ahead and make things happen - but we need that start."
Peter: "When i said standards, I was thinking more about CSS and ARIA, not MathML - as they are moving..."
Ivan: "MathML can be moved if there is a direction to move it. But we should have a clear documentation as to how you want those orgs to move."
Peter: "There won't be much detail in the white paper - as it's too high detail."
Tzviya: "We've talked ARIA a few times here, but if we're moving forward with CSS - which we hope to - I would really love to see you and Houdini work on this together."
<astearns> +1
<clapierre> +1
<pkra> +1
...: "We should work on the object model together. If they could all
talk to assistive technology at the same time, that would be best. Not
sure if we could make a task force for Math across the whole W3C, but
that'd be great."
<clapierre> http://w3c.github.io/dpub-accessibility/
CharleS: "Link to accessibility note above. Have done a WCAG - all seem to be relevant - and did a GAP analysis and came up with the following: Page numbering, drop caps, position/location of text, indication of text, nouns, layouts, influences, deeply nested headings, semantic list-heads, skipability, escapability, diagram models, appendix, and also needed but being addressed: notes & footnotes (aria), and annotations"
...: "We were hoping to have the meat of the note done by end of year, then have it wordsmithed and ready by early spring next year"
Ivan: "Practical thing, the best way of doing it is that we issue a draft of a note, which we can use to issue the draft export for the working group of WCAG, then we can have the final note."
Charles: "When would we do that draft?"
Ivan: "When you think it's good enough for publishing. Publishing a draft is very easy."
Deborah: "The specifics of how we were planning on doing it is to get to a point where we have something ready, then go to ivan and ask what to do. It's on our todo list. 'we have this draft, now what do we do?'"
Tzviya: "There has been a request to a time change for meeting.
<astearns> I'm OK with most any time
change
...: "What about a time-change for the meeting?"
Ivan: "It's unpleasant, but the co-chair is in Europe."
Nick: "Crazy idea - staggering meetings - have alternative weeks late"
Tzviya: "Alternating would be difficult."
Tzivya: "Would people be interested in switching to a different day of the week? Different hour?"
Tzviya: "We'll poll people via e-mail to
figure out a good time"
...: "December meetings..."
<dauwhe> +1 for January 4
+1 for Jan 4
<laudrain> -1 28
<clapierre> +1 4
<pkra> -1 Dec 28, Jan 4 (traveling)
<laudrain> +1 for Jan 4
Tzivya: "All else on, but the 28th of December is a no"
<ayla_stein> +1 Deborah
Deborah: "I understand how difficult it is to schedule - we've tried to solve this on accessibility. It is an ongoing problem of diversity in the W3C and it would be great if we could do something to be part of the solution."
<pkra> I heard alternate timeline.
Ivan: "Deborah, there is an experiment - that wasn't terribly successful - we had india colleagues who tried to set up an alternate timezone call where an expert of that region would try to discuss things with that interest group - so the results of that might be good"