mozilla :: #mathml

13 Mar 2014
01:22raniereteoli: thanks to invite Kaitlin
09:56fredwHi fscholz. About the MathML CSS fallback, I'm wondering if we could use the Kuma template system to indicate which wiki page requires this stylesheet (+ a warning). For example, the MathML demos and "sample rendering in your browsers" don't need that.
10:10fscholzI think we could do that. If I see it correctly, the first thing to figure out is, how to only apply the CSS to non MathML browsers, right?
10:11fscholzAnd I don't think we can do that differentiation using the template system
10:14fscholzThe template system could help to have always the same structure, though. For instance: {{mathml("yourmathmlcodehere")}}, could automatically add a note above the mathml.
10:16fscholzThe template system is only allowed to generate content that is also allowed if you would write it in the document. JavaScript gets stripped.
10:19fredwfscholz: OK. So I think this could not be done by the template system only.
10:20fredwfscholz: this is the UA string sniffing with Javascript:
10:21fredwJavascript can also be used to insert, say an mfrac, and check that the numerator is really above the denominator using getBoundingClientRect
10:21fredwFinally, there are "CSS hacks" to apply rules only for some browsers but I don't think it is very reliable
10:23fscholzOk, I am not so familiar with UA string sniffing
10:23fredw(BTW, while we are talking about Javascript, Kuma removes Javascript from MathML href attribute, right?)
10:24fscholzI hope so
10:25fscholzI dont think we want arbitrary JS to executed anywhere
10:25fscholzThere is live samples which run on a different domain
10:26fscholzYou have used that for the torture test or so, iirc
10:28fscholzYou are able to have CSS, JS and HTMLthere
10:32fredwfscholz: but does it help to load the mathml.css stylesheet? I don't think we want to use Live sample each time we have MathML equation...
10:33fscholzRight, I just wanted to mention that this is the only case where we would allow JS and CSS
10:37fredwfscholz: can we put a live sample inside a template?
10:38fredw(I tried to put <style> or <link> inside a <annotation-xml> to load mathml.css. But Gecko/WebKit does not ignore the style in the <annotation-xml>)
10:41fscholzfredw: I haven&#39;t tried that. But I think that might be doable
10:41fredwfscholz: ok. Otherwise there are CSS hacks at
11:23fredwfscholz: so here is a test page:
11:23fredwI don&#39;t think live sample will be enough since Javascript executes in another domain
11:23fredwSo I think that&#39;s something that should be implemented directly in Kuma code
11:24fscholzThe rollover is a bit flaky for me btw
11:25fredwfscholz: this can be easily removed. I don&#39;t really have a preference
11:25fredw(just drop the math:hover rules)
11:26fscholzyeah we can just drop it.
11:27fscholzBut very nice otherwise. good to have something better for our chrome readers
11:27fredwthe problem is that the TeX source line and the CSS-based rendering don&#39;t have the same size. So if the mouse is not at the right position, it&#39;s like if you were doing mouseout and mouseover
11:27fscholzyeah, I thought so
11:28fredwperhaps :focus would be better
11:30fscholzI have no strong preference
11:31ranierefredw: fscholz: good morning
11:32fredwraniere: Hi
11:33raniereI thought in only include the CSS for pages under and
11:34fredwraniere: I think it&#39;s needed in other places
11:34fscholzno, because there are JavaScript and SVG pages where we need it
11:34ranierebut first I have to discovery how to include the CSS only for Chrome.
11:34fredwe.g. for SVG/WebGL doc and Javascript
11:34fredwactually, I think we do *not* want to include the CSS for MathML demos and samples, since we want to show the actual MathML rendering of the browser
11:35fredw(see the discussion above)
11:36fredwbasically, I think we can use UA string sniffing or some detection with getBoundingClientRect
11:36fredwbut in any case the problem is to execute Javascript, so that must be done in the Kuma script itself
11:36raniereok. I see it.
11:37raniereI won&#39;t be able to work on it next week. I will be travelling. But will try to find some time when I came back.
11:38fredwraniere: ok, thanks no problem
11:38fredwfscholz: :focus does not seem to work very well
11:39fredwbut :active seems better. You can click on the expression to switch between TeX/CSS rendering
12:06fredwfscholz: do I need special privilege to create templates
12:07fscholzfredw: yes
12:08fscholzfredw: Permissions granted. Use your power wisely :)
12:12fredwfscholz: thanks.
12:13fredwfscholz teoli: here is the warning I propose
12:13fredw<strong>This page contains some mathematical formulas.</strong>
12:13fredw<p>For best mathematical rendering, be sure to use a [browser with MathML support]( and to have [appropriate fonts]( installed.</p>
12:13fredwso that will show as the harmony warning:
12:14teoliIf it appears only on browsers without MathML support, it looks fine for me.
12:14fscholzIs that then for non-MathML browsers only?
12:15fredwfscholz: teoli: I don&#39;t think I can do that with the template system.
12:15fscholzyep, you cant
12:15fredwI thought we wanted to propose math fonts for Firefox too...
12:15fredwIf that&#39;s for non-MathML browsers only, then this should be done by the script that inserts the mathml.css too
12:16fscholzyeah, good idea but I dont think that rather big banner should appear on pages where people just want to read about the topic (JS here)
12:22fredwteoli: fscholz: so what about just &quot;Your browser does not seem to support MathML. Some CSS fallback will be used instead, but the mathematics on this page might not render correctly.&quot; for non-MathML browsers only.
12:22fscholzSounds good to me
12:25fredwraniere: see the comments above
12:46ranierefredw: sounds good to me.
12:49ranierefredw: what branch of epiphany/webkitgtk did you use for test MathML?
12:49fredwraniere: I&#39;m not sure MathML is enabled in the stable release
12:49fredwit was disabled because of Chromium&#39;s comments
12:50ranierefredw: OK. I will try to compile it from source. =)
12:50fredwraniere: which platform are you developing on?
12:51ranierefredw: GNU/Linux - Arch
14:02fredwraniere: I have updated Authoring MathML with the MathML detection:
20:38raniereMozilla MathML Project | Meeting will start at 9pm UTC . Topics at;
20:45ranierefredw: fscholz: KhaledHosny: maxli: teoli: Our meeting will begin in 15-20min
20:45teoliI&#39;m here :-) Will listen to it mostly.
20:46fredwKarl said he is at a meeting all the week. So I guess he won&#39;t come.
20:59ranierejkitch: welcome. We will give ~5min to see if someone else join the meeting
21:04fredwraniere: I guess we can start
21:04ranierefredw: OK
21:05ranierefredw: fscholz: jkitch: KhaledHosny: maxli: teoli: welcome to our meeting.
21:05ranierethe topics are available at
21:05ranierethe first one in about the hour of our meeting.
21:06ranierecan we change it a little bit?
21:06ranierefredw: do you know if it will help Bill?
21:07fredwwe can change it earlier for me, but then that might be a bit too early for jkitch & karl. Although I don&#39;t know exactly.
21:07fredwI think teoli said he&#39;d prefer earlier too
21:07fredwand Michael Kohlhase (Germany) said that too
21:08teoliIf possible, yes, but I&#39;m fine with the current time. I was on CET the first time (now GMT)
21:08teoliMaybe a doodle to gather feedback?
21:09fredwteoli: yes, we can probably do that
21:09raniereteoli: I can send a doodle if we agree that it will help
21:10raniereany objection for the doodle?
21:10jkitchI&#39;ll only be able to attend until daylight saving ticks over, at which point the current meeting time will become too early. An earlier meeting time will make little difference to me.
21:11ranierejkitch: we can change the time when daylight saving end.
21:11fredwLet&#39;s send a doodle to gather feedback.
21:11ranierefredw: OK.
21:11fredwPeople for which the current time is problematic are not here to comment :-)
21:12ranierefredw: do you have any progress in WebGL + MathML demo?
21:12fredwI did one for my project. But it is private.
21:13fredwIt&#39;s the document coordinate-systems-20140131 for people who have access to it
21:13ranierewhen you make it public let us know.
21:14ranierePhysikerwelt isn&#39;t here. Anyone have news about Wikipedia?
21:14fredwI think he said he is only progressing slowly. He is still trying to get feedback/review from MediaWiki people
21:15raniereAny other topic from our previous meeting?
21:15fredwnot for me
21:16raniereNext topic is &quot;Web developments&quot;
21:16raniereI and fredw make some progress with TeXZilla
21:17ranieremy plugin is already available at MDN (feedbacks are welcome)
21:17raniereand fredw propose a CSS3 fallback for browsers thaat don&#39;t support MathML
21:18fredwIf you want to insert/edit MathML on MDN, there is now a &quot;square root of x&quot; button to open a TeXZilla dialog
21:19fredwthe CSS fallback was discussed in the previous MDN meeting, and people agree. So now one has to integrate it in Kume :-)
21:20fredwBTW, I also updated the MDN page about MathML fallback and browser detection:
21:21ranierefredw: and what is the <x-tex> tag?
21:21fredwIt&#39;s a Web Components Custom Element
21:21fredwBasically, it&#39;s a wrapper for LaTeX expression
21:22fredwone writes <x-tex>x^2 + y^2</x-tex> and this is automatically converted into the MathML output via TeXZilla
21:22fredwthere is a demo here:
21:22ranierefredw: what&#39;s the differenc from MathJax when I use $x^2 + y^2$?
21:23fredwIt&#39;s essentially the same, except that it handles better dynamic change
21:24fredwalso, at the moment it uses the x-tag polyfill, but I think when shadow DOM is available, it will be cleaner
21:24teoliAnd <tex> will be allowed by the spec. (Fx 30 should support Shadow DOM, btw)
21:25fredwSo concretely, you can do var y = document.createElement(&quot;x-tex&quot;); y.source = &quot;x^2 + y^2&quot;; document.body.appendChild(y);
21:25raniereShadow DOM:
21:26ranierefredw: looks nice.
21:26raniereanyone has any other topic about &quot;Web developments&quot;?
21:26fredwwith MathJax, you have to use the insert some LaTeX source somewhere, then use the MathJax API to explicitly ask a reparse of LaTeX-to-MathML and perhaps a rerendering of the MathML output
21:29raniereNext topic is &quot;Gecko developments&quot; (line 33)
21:29fredwSo the idea is just to create a &quot;<tex>&quot; tag for people complaining about the fact that MathML is too verbose, while still being compatible with usual the Web API.
21:30ranierefredw: sorry. let me know when we finish the previous topic.
21:30fredwI haven&#39;t done much on Gecko for the past month
21:30fredwjkitch has done some work, though
21:30fredw(I&#39;ve been focusing on WebKit while waiting karl&#39;s review)
21:31raniereWe have close 6 bugs, open 4 and update 5.
21:31fredwAlso Anuj and Branko, two new contributors have done some work.
21:31raniereanyone who had work in any of this bug and want to talk a little, just raise your hand
21:32fredw&quot;[Bug 975747] New: Convert gOperatorTable to a modern hashtable&quot; might be a &quot;good first bug&quot; for newcomers
21:32raniereAnuj is from India (this isn&#39;t a good time for him)
21:35fredwSo Branko has been implemeting the new MathML3 &quot;phasorangle&quot; notation
21:35fredwand Anuj is working on new MathML reftests since many of the features are not tested
21:36fredwjkitch is working on various bugs.
21:36jkitchMostly waiting on review, but I&#39;m still working through the details of Bug 877563
21:37fredwWhich review?
21:38fredwjkitch: and do you have comments/questions on bug 877563?
21:38jkitchthe rowspacing bug primarily.
21:39jkitchI&#39;m still working through the details of 877563. I take it we are meant to ignore all user set values of lspace and rpsace in such situations.
21:39fredwjkitch: I think roc was not really fond of with this change, so we will see.
21:40fredwjkitch: yes, I think that&#39;s what LaTeX & MathJax do. And what WebKit used to do until recently.
21:41jkitchIn a case of <msub><mo>x</mo><mn>1</mn>, should the lspace/rspace also be removed? the spacing belongs to the <msub> element rather than the <mo>
21:41fredwit&#39;s an embellished op
21:42fredwso it depends on the context of the <msub>
21:42jkitchso remove it if the msub is not in an inferred mrow.
21:43fredwthat&#39;s why the line I indicate is to remove the space for the &quot;outermost embellished container&quot;, not in the core <mo>
21:45jkitchRight. now I understand.
21:46fredwraniere: can we move to the next topic?
21:46ranierefredw: yes
21:46raniereNext topic is &quot;Documentation&quot; (line 87)
21:47raniereTwo bugs were closed.
21:47fredwI think we already mentioned that
21:47fredwthat&#39;s basically the integration of TeXZilla in MDN
21:47fredwand I updated some MathML pages during the MDN Weekend
21:47ranierefredw: that&#39;s true
21:47fredwSo for example
21:47raniereNext week I will try to translate some pages
21:48fredwteoli: fscholz: any comment on that?
21:48raniereteoli: I can test the tools ^
21:48fredwI translated a few MathML pages into French during the MDN Work weekend
21:49fscholzI already discussed this enough I think. Next step to improve the non_mathml browser experience and we are good to go with mathml on mdn I think
21:49ranierefscholz: I believe that we what to know about the translation process
21:50fscholzoh sure
21:50fscholzThis is an overview about the translation status of MathML docs
21:51fscholzIf you click on the locale, e.g. &quot;fr&quot; you get a status table
21:51fscholzWhich then tells you if a translation is availabe and if it is in sync with the english one
21:52ranierefscholz: this documentation status rocks!
21:53fscholzthanks :)
21:53teoliraniere: yes, it definitively rocks
21:53ranierefscholz: teoli: what type of feedback do you need?
21:53fscholzIf you are using this system and notice any bugs or things you need, please let me / teoli know
21:54teoliraniere: if all the needed info is available for example
21:54teoliraniere: or if you notice things that are not quite pratical (like going up and down to get some info)
21:55raniereLast topics
21:55raniereI will apply to gsoc for extend Firefox OS keyboard to support math
21:56ranierethe draft are available at
21:56raniereany comments are welcome
21:56ranierefscholz: need some info after this meeting.
21:57fscholzI am not that familar with GSoC to be honest
21:57fscholzteoli knows more than me
21:57teoliraniere: do you have a mentor?
21:58teoliThat&#39;s the most important point: finding a mentor, else the project has very few chances to be accepted.
21:58raniereteoli: some folks say that they can be the mentor
21:58raniereI need to know what they should do.
21:59fredwraniere: they should put their names on the proposal page of the Mozilla wiki
21:59fredwand then the proposal should move to the official wiki page IIRC
21:59ranierefredw: thanks
22:00ranierefscholz: teoli: fredw: any other topic about the MDN work weekend?
22:00teoliyes, they also need to announce themselves as mentor to the GSoC coordinator at Mozilla. He will help them.
22:01teoliI can find his name again (he changed, he is French now). It is quite easy, but they will need an account on the submission tool so that can formerly accept the project.
22:02fscholzNothing more from the work weekend other that I finally met fredw :-)
22:02fredwand I finally met fscholz :-)
22:02ranieremaybe I go to europe in July. Hope to meet fredw and fscholz.
22:03raniereteoli: are you in Europe?
22:03teoliraniere: yes, London.
22:03raniereteoli: =)
22:04raniereThe last topic of this meeting is LaTeXML
22:04ranierelook like that it now support TikZ
22:04teoliNote that there will be 2 l10n community events next month (French + German) and fscholz and I will go there. So maybe will find some translators for MathML (but not sure)
22:04fredw(raniere: you need to have your mentor name here: Then the GSoC admin will review it and move to
22:04* raniere need to test LaTeXML TikZ support ASAP
22:05fscholzoh, we need to invite physikerwelt in Berlin
22:05fredwteoli: yes, you told me that before I leave. What are the dates/places?
22:06fscholzBerlin for German is April 5 and 6
22:06fscholzSprint on translating mozilla products, SUMO, MDN and more
22:07fredwSo I guess people can participate remotely
22:07fscholzlike always, yes
22:07teoliParis, 26-27 April
22:09ranierefredw: fscholz: jkitch: KhaledHosny: maxli: teoli: Any other comments?
22:09jkitchNot from me
22:09fredwSo on LaTeXML
22:10fredwif someone is interested to try to build and packaging it on Windows, that could be helpful
22:10fredwAlso we missed a comment about TeXZilla
22:11fredwPeople can try to write plugins for other CMS (WordPress, Drupal, Joomla etc)
22:11fredwOr try to use the TeXZilla server-side conversion
22:11fredwSo if some people want to get involved, that would be a good way to help
22:12fredwAt the moment, we have raniere&#39;s
22:13raniereWas an pleasure to meet you all.
22:13fredwAlso, as usual raniere&#39;s virtual keyboard project is a web app, so will not only work on FirefoxOS but also can be reused on any Web platform. See
22:14teoliraniere: same for me
22:14raniereand thanks for your time.
22:15fredwthanks for organizing the meeting
22:15raniereany feedbacks on how I&#39;m organizing this meeting are very welcome
22:17teoliGood night all (or good evening)
22:17fredwSo good bye everybody
22:19* raniere going out!
14 Mar 2014
No messages
Last message: 131 days and 8 hours ago