- From: Peter Krautzberger <peter.krautzberger@mathjax.org>
- Date: Mon, 15 May 2017 10:08:41 +0200
- To: Daniel Marques <dani@wiris.com>
- Cc: "Swenson, Kyrce" <kyrce.swenson@pearson.com>, mathonwebCG <public-mathonwebpages@w3.org>
- Message-ID: <CABqxo82P9XOT6SNpWbN7+vO_C7-o_2_C2E2Uz39Wui4813Sv0A@mail.gmail.com>
Thanks so much, Kyrce! Best, Peter. On Fri, May 12, 2017 at 2:53 PM, Daniel Marques <dani@wiris.com> wrote: > Thank you very much Kyrce! > > > > Dani > > > > *From:* Swenson, Kyrce [mailto:kyrce.swenson@pearson.com] > *Sent:* jueves, 11 de mayo de 2017 19:19 > *To:* public-mathonwebpages@w3.org > *Subject:* Minutes 2017/05/11- DRAFT > > > > *Topic:* Layout > > *Attendees: *Peter Krauzberger, Daniel Marques, Jean Kaplansky, Volker > Sorge, Sam Dooley, Charles LaPierre, Kyrce Swenson > > *Scribe:* Kyrce Swenson > > > > ----- > > > > *Peter: *What can improve and give feedback on layout requirements so > that other web standards can be influenced by the WG? > > > > Collect use cases. > > > > Some contact already with the CSS working group, Including actual issue > that MathJAX needs to respond to. > > > > Link in issue to codepen example to sketch out the basic problem: [1] Two > dimensional content in inline content. Must be aligned with the baseline. > In CSS only the top or the bottom can be aligned to the baseline.. No tools > for something like a simple fraction, etc. This example was trying to > explain this problem. Current solutions all require manual baseline fixing > and shifts. > > > > This takes time and adds complexity. We need a way to tell CSS what should > be at the baseline level. > > > > *Dani**: *The limits of the technology are MathJAX and we can use MathML > directly in FF, and there are solutions from Wiris. Be able to do > mathematics using only html5 + css. We need to be able to tell the CSS > people to convince them. > > > > *Peter: * > > there's also KaTeX, MathQuill, and a bunch of hobby projects. > > > > *Peter* identify use cases that are broader than mathematical. Example: > aligned SVGs inline because they had some textual content. Wanted to align > the text not the image [2] > > > > Little chance that this will be something only math needs. These > challenges are interesting in a general sense as well. > > > > *Volker: *Right now we are talking about vertical growth not ascenders > and descenders growing. More serious problem: Left to Right layout you have > to change the layout while you are rendering a stretchy character. Hard to > describe but obvious when i see it. > > > > Fraction as a vertical shift. Potentiallyl two directions. To render a > line you go from the left to the right. Once the fraction grows then you > have to go back and start stretching your brackets, or similar. Not > supported by CSS. > > > > *Jean: *re: CSS and the need to go forward and back... You're talking > about multiple runs through to create one rendition. > > > > Browsers don't do this stuff. Formatting engines do. > > > > I could see mathJax or a web service doing this, but think the browser > vendors will kick and scream over the request. > > > > *Volker: * So meaning that there's no reason to have this in the CSS > spec? Yes they will but there might be other ways of doing this. How do I > style my parent element based on whether or not it has a particular child? > > > > *Jean: *Choose your battles? Do you think you can get vendors on board > with it in the spec? > > > > Right. Conditional stuff. But again - its the ability to look forward and > look back. > > > > *Volker: *goal to get things in the spec? or figure out what is needed? > > > > Browsers aren't good at looking forward and look back in declarative > languages. > > > > *Peter: *First step to collect what problems we see. But next steps are > to identify the right pieces to tackle. Might be more sensible to solve > with SVG group. > > > > *Dani: *Better to discuss how to discuss the examples? Wiki? > > > > *Peter: *just use the github issue tracker with a label -- individual > issues labeled as layout. > > > > *Jean: *Thumbs up for a wiki - maybe a github page? > > Why not a github wiki page so we can see everything in context? > > > > *Peter: *does not allow to inject CSS. Not enough control. > > > > *Dani: *what is the idea to work collaboratively? people just fetch the > file and edit it? > > > > *Peter:* just go to edit file and make a pull request? Lots of things > available via the web interface. > > > > *Dani: * better to have each of its problems on its own page. > > > > *Peter: *if we split it on the issue tracker we can actually help people > give feedback if they think it is a priority to them. Not expecting > hundreds of people however. > > > > *Dani*: at the point that something is written, more people who are > interested will join. > > > > [Peter had to leave. ] > > > > *Dani: *do we want to continue or just speak in two weeks? > > > > Decision was made to speak in two weeks. Dani will reformat and send > examples. > > > > [1] https://github.com/w3c/csswg-drafts/issues/1339 > > [2] https://w3c.github.io/mathonwebpages/examples/display/html5.html > > > > -- > > *Kyrce Swenson* > > Solutions Manager > C4 Semantic Markup: Tooling & Players > C5 Structured Authoring: Proofing > > > > Pearson PLC > > 221 River Street, Second Floor > Hoboken, NJ 07030 USA > > VM: +1 (201) 236 5611 <(201)%20236-5611> > M: +1 (973) 744-0741 <(973)%20744-0741> > kyrce.swenson@pearson.com > > > > *Learn more at pearson.com <http://pearson.com>* > > > [image: Pearson] > > [image: Pearson Tagline] > > > > ᐧ >
Received on Monday, 15 May 2017 08:09:16 UTC