Bug 24207 requests switching the math rendering preference default from its current setting (which usually produces a nice PNG and occasionally produces some kinda ugly HTML) to the "always render PNG" setting.
I'd actually propose dropping the rendering options entirely...
* "HTML if simple" and "if possible" produce *horrible* ugly output that nobody likes, so people use hacks to force PNG rendering. Why not just render to PNG? * "MathML" mode is even *MORE* limited than "HTML if simple", making it entirely useless. * nobody even knows what "Recommended for modern browsers" means, but it seems to be somewhere in that "occasionally crappy HTML, usually PNG" continuum.
So we're left with only two sane choices:
* Always render PNG * Leave it as TeX (for text browsers)
Text browsers will show the alt text on the images, which is... the TeX code. So even this isn't actually needed for its stated purpose. (Hi Jidanni! :) lynx should show the tex source when using the PNG mode.)
It's conceivable that a few folks really honestly prefer to see the latex source in their graphical browsers (should at least do a quick stat check to see if anybody uses it on purpose), but I wouldn't mind removing that either.
Fancier rendering like MathJax etc should be considered as a separate thing (and implemented a bit differently to avoid parser cache fragmentation!), so don't let future mode concerns worry y'all. Any thoughts on whether this makes sense to do for 1.18 or 1.19?
https://bugzilla.wikimedia.org/show_bug.cgi?id=24207#c9
-- brion
On Mon, Jul 18, 2011 at 5:38 PM, Brion Vibber brion@pobox.com wrote:
I'd actually propose dropping the rendering options entirely...
+10
It's conceivable that a few folks really honestly prefer to see the latex source in their graphical browsers (should at least do a quick stat check to see if anybody uses it on purpose), but I wouldn't mind removing that either.
It *should* be ok to drop this, but checking first is nice too :)
Fancier rendering like MathJax etc should be considered as a separate thing (and implemented a bit differently to avoid parser cache fragmentation!), so don't let future mode concerns worry y'all. Any thoughts on whether this makes sense to do for 1.18 or 1.19?
Definitely not 1.18. 1.19 or beyond sounds like a great idea to me though.
-Chad
On Mon, Jul 18, 2011 at 8:38 PM, Brion Vibber brion@pobox.com wrote:
Bug 24207 requests switching the math rendering preference default from its current setting (which usually produces a nice PNG and occasionally produces some kinda ugly HTML) to the "always render PNG" setting.
This would also close bugs: * 25645: Improve decision when to render a formula as HTML (this reporter complains about our PNG rendering including fonts) * 26546: Option "MathML if possible" doesn't work * 19177: XHTML parse error when using some math formulas (if we convert to PNG-only you won't have XHTML garbage)
And may be related to: * 15057: problems with forced PNG rendering (possibly a caching issue, but does not seem to have been fixed) * 17465: Port texvc to PHP, reducing external dependencies and development impedence for Math extension (it would be easier to mandate a particular solution if it was /our/ solution and perhaps slightly less bad) * 7208: Change font for texvc-generated images (if people think texvc is ugly because of inconsistent fonts, that's easy enough to fix before we make it the only option) * 16719: Math hashes should include versioning to allow sensible updates (this requires a schema change, so should probably be out for a version or so to iron out the bugs before we really rely on it, but it would be excellent if this was working before PNG rendering became the only option, as it would make PNG rendering a much 'cleaner' only option to have)
I'd actually propose dropping the rendering options entirely...
Drop all of the preferences.
It's conceivable that a few folks really honestly prefer to see the latex source in their graphical browsers (should at least do a quick stat check to see if anybody uses it on purpose), but I wouldn't mind removing that either.
Not likely, latex is hardly pretty, but it's easy enough to check whether it's in use on WMF wikis. I'm not really up to date on the release cycle but is seems really late to do this for 1.18, especially given that we'll likely see a few more complaints about texvc's flow, and it would be nice if there was time for someone to check on bugs 15057, 24445, and 16719.
--Dan
2011/7/19 Brion Vibber brion@pobox.com:
I'd actually propose dropping the rendering options entirely...
I'm mostly OK with it, but see below.
- "MathML" mode is even *MORE* limited than "HTML if simple", making it
entirely useless.
So that's why it doesn't work.
It's conceivable that a few folks really honestly prefer to see the latex source in their graphical browsers (should at least do a quick stat check to see if anybody uses it on purpose), but I wouldn't mind removing that either.
I know one such person, a respected university professor. Leave this option and kill the rest.
-- Amir Elisha Aharoni · אָמִיר אֱלִישָׁע אַהֲרוֹנִי http://aharoni.wordpress.com "We're living in pieces, I want to live in peace." - T. Moore
Why not consider building in SVG rendering natively while we are at it, instead of relying on extensions.
Many browsers have joined the SVG bandwagon anyway, and it would cut down on the "pixelated look" which we see when zooming in on tablets, as well as on PCs with very large displays. I'm sure we can use PNG as a fall-back, like the extension "LaTeXSVG" does.
On Tue, Jul 19, 2011 at 9:46 AM, Amir E. Aharoni < amir.aharoni@mail.huji.ac.il> wrote:
2011/7/19 Brion Vibber brion@pobox.com:
I'd actually propose dropping the rendering options entirely...
I'm mostly OK with it, but see below.
- "MathML" mode is even *MORE* limited than "HTML if simple", making it
entirely useless.
So that's why it doesn't work.
It's conceivable that a few folks really honestly prefer to see the latex source in their graphical browsers (should at least do a quick stat check
to
see if anybody uses it on purpose), but I wouldn't mind removing that either.
I know one such person, a respected university professor. Leave this option and kill the rest.
-- Amir Elisha Aharoni · אָמִיר אֱלִישָׁע אַהֲרוֹנִי http://aharoni.wordpress.com "We're living in pieces, I want to live in peace." - T. Moore
Wikitech-l mailing list Wikitech-l@lists.wikimedia.org https://lists.wikimedia.org/mailman/listinfo/wikitech-l
On Mon, Jul 18, 2011 at 8:38 PM, Brion Vibber brion@pobox.com wrote:
It's conceivable that a few folks really honestly prefer to see the latex source in their graphical browsers (should at least do a quick stat check to see if anybody uses it on purpose), but I wouldn't mind removing that either.
Some people are using this method to have MathJax render the math on Wikipedia pages, by installing the MathJax javascript in their user javascript. Disabling the text-only option would also break this.
It would really be worthwhile for us to look into making MathJax available as an option for all users without forcing them to modify their user javascript. It's basically just a few javascript files and some static files on a webserver. They claim very broad browser compatibility ( http://www.mathjax.org/resources/browser-compatibility/ ) , and everything is released under the Apache license IIRC.
- Carl
On Tue, Jul 19, 2011 at 12:58, Carl (CBM) cbm.wikipedia@gmail.com wrote:
On Mon, Jul 18, 2011 at 8:38 PM, Brion Vibber brion@pobox.com wrote:
It's conceivable that a few folks really honestly prefer to see the latex source in their graphical browsers (should at least do a quick stat check
to
see if anybody uses it on purpose), but I wouldn't mind removing that either.
Some people are using this method to have MathJax render the math on Wikipedia pages, by installing the MathJax javascript in their user javascript. Disabling the text-only option would also break this.
For more info, see: http://en.wikipedia.org/wiki/Special:PrefixIndex/User:Nageh/mathJax
It would really be worthwhile for us to look into making MathJax available as an option for all users without forcing them to modify their user javascript.
+1
Helder
Sounds like someone should propose MathJax as a gadget: http://en.wikipedia.org/wiki/Wikipedia:Gadget/proposals
Ryan Kaldari
On 7/19/11 8:58 AM, Carl (CBM) wrote:
On Mon, Jul 18, 2011 at 8:38 PM, Brion Vibberbrion@pobox.com wrote:
It's conceivable that a few folks really honestly prefer to see the latex source in their graphical browsers (should at least do a quick stat check to see if anybody uses it on purpose), but I wouldn't mind removing that either.
Some people are using this method to have MathJax render the math on Wikipedia pages, by installing the MathJax javascript in their user javascript. Disabling the text-only option would also break this.
It would really be worthwhile for us to look into making MathJax available as an option for all users without forcing them to modify their user javascript. It's basically just a few javascript files and some static files on a webserver. They claim very broad browser compatibility ( http://www.mathjax.org/resources/browser-compatibility/ ) , and everything is released under the Apache license IIRC.
- Carl
Wikitech-l mailing list Wikitech-l@lists.wikimedia.org https://lists.wikimedia.org/mailman/listinfo/wikitech-l
On Tue, Jul 19, 2011 at 4:03 PM, Ryan Kaldari rkaldari@wikimedia.org wrote:
Sounds like someone should propose MathJax as a gadget: http://en.wikipedia.org/wiki/Wikipedia:Gadget/proposals
That solves some of the problems, but making it a gadget isn't technically very different from just putting it in user javascript.
Not all the MathJax files are javascript: there are also web fonts and fallback image fonts. Ideally these would be served from Wikimedia servers. The system they are using right now loads them from off-site somewhere, and that would not be fixable by the gadget system. Users can also install the fonts locally but that is not a good solution for a widespread roll-out.
The ideal implementation would be a system that detects whether the user has javascript enabled and a decent browser, and if so rolls them over to MathJax automatically. Users who fail that test or who intentionally opt out would receive images only. This would completely replace the current system (but re-use texvc to make the images).
This system would not be hard to implement, but there is little incentive for anyone to spend time on it without some pre-commitment from the site admins to make it live. If we just want something that users can opt in to get, we already have that.
- Carl
Ok, I've set up an RfC page on MediaWiki.org which'll be easier to comment on for folks not on the lists: http://www.mediawiki.org/wiki/Requests_for_comment/Reduce_math_rendering_pre...
and added some links, comments & requests for further feedback at: http://en.wikipedia.org/wiki/Wikipedia_talk:WikiProject_Mathematics#RFC:_Sho...
I'd also appreciate if folks who know some of the math-heavy editors who work on other projects -- other languages, or wikibooks/wikiversity etc -- can poke around and get more feedback.
In addition to the immediate issue of tweaking the rendering modes, there's been some good feedback about related issues such as the lack of proper baseline alignment of the math PNG images, which should be taken into account for the next set of improvements on Math -- improving or replacing texvc with another tool, maybe dusting off Blahtex and making sure it does everything we need, etc.
-- brion
Hey folks, remember this discussion from a few months ago?
I've gone ahead and implemented the initial part for 1.19 (disabled the math rendering options other than PNG and source), and opened another bug entry with some details on how to add the baseline shifting that should make PNGs nicer inline:
https://bugzilla.wikimedia.org/show_bug.cgi?id=32694
While I'm mucking about in there, I'd like to note that I'm also interested in making higher-resolution images, suitable both for printing and for high-density screens (iPhone 4 being the best-known example in use in the wild). Generating another PNG or two should be easy; getting them to display reliably might be slightly funkier. :)
https://bugzilla.wikimedia.org/showdependencytree.cgi?id=32696
When I've got something working for that I'll check around to make sure it works consistently for people and doesn't break any tools.
-- brion
On Thu, Jul 21, 2011 at 11:38 AM, Brion Vibber brion@pobox.com wrote:
Ok, I've set up an RfC page on MediaWiki.org which'll be easier to comment on for folks not on the lists:
http://www.mediawiki.org/wiki/Requests_for_comment/Reduce_math_rendering_pre...
and added some links, comments & requests for further feedback at:
http://en.wikipedia.org/wiki/Wikipedia_talk:WikiProject_Mathematics#RFC:_Sho...
I'd also appreciate if folks who know some of the math-heavy editors who work on other projects -- other languages, or wikibooks/wikiversity etc -- can poke around and get more feedback.
In addition to the immediate issue of tweaking the rendering modes, there's been some good feedback about related issues such as the lack of proper baseline alignment of the math PNG images, which should be taken into account for the next set of improvements on Math -- improving or replacing texvc with another tool, maybe dusting off Blahtex and making sure it does everything we need, etc.
-- brion
Brion Vibber wrote:
I've gone ahead and implemented the initial part for 1.19 (disabled the math rendering options other than PNG and source), and opened another bug entry with some details on how to add the baseline shifting that should make PNGs nicer inline:
https://bugzilla.wikimedia.org/show_bug.cgi?id=32694
While I'm mucking about in there, I'd like to note that I'm also interested in making higher-resolution images, suitable both for printing and for high-density screens (iPhone 4 being the best-known example in use in the wild). Generating another PNG or two should be easy; getting them to display reliably might be slightly funkier. :)
https://bugzilla.wikimedia.org/showdependencytree.cgi?id=32696
When I've got something working for that I'll check around to make sure it works consistently for people and doesn't break any tools.
Did you have a chance to evaluate MathJax? http://www.mathjax.org/ I know it's come up in past math discussions and that a lot of math folks think it looks promising. A technical analysis of its feasibility on Wikimedia wikis would be great. Killing the less-used, ancient math options is great, but perhaps adding one wouldn't be too bad to do too. :-)
MZMcBride
On Mon, Nov 28, 2011 at 3:05 PM, MZMcBride z@mzmcbride.com wrote:
Brion Vibber wrote:
[snip my notes about removing the non-PNG non-source options, wanting higher-resolution renderings]
Did you have a chance to evaluate MathJax? http://www.mathjax.org/ I know it's come up in past math discussions and that a lot of math folks think it looks promising. A technical analysis of its feasibility on Wikimedia wikis would be great. Killing the less-used, ancient math options is great, but perhaps adding one wouldn't be too bad to do too. :-)
That's an excellent thing to bring up -- MathJAX *does* look very promising, and things seem to render pretty nicely. Need to make sure that we can either do that type of rendering cleanly with the PNG fallback (older browsers will still need the PNGs, so it may still be worth spending the time to fix baselines).
Size of the library, and compatibility, could be an issue for mobile but is worth checking on. (Looks like it *does* work in Android and iPhone browsers, so has a good chance of being something we could use as a progressive enhancement.)
It's actually probably a better idea to go ahead in that direction than to worry about high-resolution renderings from texvc for now. (Other image types, including icons in the UI, will still need high-resolution versions though.)
-- brion
wikitech-l@lists.wikimedia.org