Dear all, I've updated my repository of OpenType math WOFF fonts to add some CSS rules for old-style numbers and calligraphic letters: http://fred-wang.github.io/MathFonts/ So the current status from my testing is - Asana Math and XITS have calligraphic variants. - Asana Math, Neo Euler and XITS have old style numbers. - Other fonts do not seem to have such variants (or they are proprietary and I can test them). I listed all the possible syntaxes, either with font-variant-numeric/font-variant-alternates not yet enabled/implemented in browsers or font-feature-settings with/without browser-specific prefix. -- Frédéric Wang maths-informatique-jeux.com/blog/frederic

Le 25/07/2014 15:35, Frédéric WANG a écrit : > - Asana Math and XITS have calligraphic variants. > - Asana Math, Neo Euler and XITS have old style numbers. > - Other fonts do not seem to have such variants (or they are > proprietary and I can test them). > Khaled Hosny pointed out that old style numbers are available for the text font and don't really make sense for the math font. So I updated my GitHub page to reflect that. Now, as I see, all the free fonts have old style numbers, but only Asana Math and XITS Math have calligraphic scripts. -- Frédéric Wang maths-informatique-jeux.com/blog/frederic

(apologies for possible multiple copies ; feel free to forward this message) Dear all, Starting with Gecko 31 (to be released the week of July 22), the recommended fonts for the MathML rendering are mathematical fonts with an OpenType MATH table [1]. The old math fonts (e.g. MathJax TeX, STIX General) are now deprecated in favor of the modern OpenType math fonts (e.g. Latin Modern Math, XITS Math, Cambria Math). There are still some known bugs/limitations in Gecko, math fonts and systems so that might take some time before the migration can be fully done. However, note that we plan to remove support for the old fonts in a near future, so please upgrade your setup as soon as possible and report any issues you find. In particular: * For Web users: If you installed MathML fonts in the past, please consider cleaning up your previous installation. Read [2] for the latest documentation of the recommended math fonts. On desktop, if you use Windows >= 7 or if you have a TeX distribution installed, it's now really likely that you don't need to install additional math fonts. On mobile devices, the MathML fonts add-on [3] is still the only option for now but we have work-in-progress patches for Android and FirefoxOS. Currently, there is not any native feature to configure your preferred math fonts, but Firefox users can try the MathML-font-settings add-on [4]. * For Web authors: If you use Web font versions of deprecated math fonts (e.g. MathJax TeX or STIX General), perhaps with the help of the ::-moz-math-stretchy pseudo-element, you should update your pages. The rule is now very simple: just apply the desired font-family on the <math> elements. You might also provide WOFF fallback or additional font-family for the text. Download [5] to get WOFF fonts together with CSS stylesheets to use on your Website. * For software maintainers: Please ensure that an OpenType MATH font is available on your system or that the appropriate dependencies for Gecko-based programs are set. You might also want to remove obsolete fonts or at least drop the dependencies for Gecko-based programs. * For font designers: You are invited to create your own math fonts and test them with Gecko. Note that fontforge has menus and a Python interface to do that [6] [7]. Existing math fonts have issues with ascent/descent, please be sure that the OS/2 metrics are consistent, so that they work the same in all platforms/browsers. Related news: * The STIX fonts (STIX-Word set) have an OpenType MATH table, but there are known to have many bugs. The STIX Consortium announced version 2.0.0 for early 2015 but for now, it is recommended to use the XITS fork instead [8]. * Partial support for the OpenType MATH table is also available in WebKit Nightly and will improve the rendering of MathML operators [9]. Requests to make OpenType MATH font installed by default on iOS & Mac systems have been submitted to Apple. [1] Section 6.3.6, http://mpeg.chiariglione.org/standards/mpeg-4/open-font-format/text-isoiec-cd-14496-22-3rd-edition [2] https://developer.mozilla.org/en-US/docs/Mozilla/MathML_Project/Fonts [3] https://addons.mozilla.org/firefox/addon/mathml-fonts/ [4] https://addons.mozilla.org/en-US/firefox/addon/mathml-font-settings/ [5] https://github.com/fred-wang/MathFonts/archive/gh-pages.zip [6] http://www.fontforge.org/math.html [7] http://www.fontforge.org/python.html [8] http://www.stixfonts.org/ [9] https://bug-133569-attachments.webkit.org/attachment.cgi?id=233418 -- Frédéric Wang maths-informatique-jeux.com/blog/frederic

Frédéric WANG <fred.wang@free.fr> writes: > ... The old > math fonts (e.g. MathJax TeX, STIX General) are now > deprecated in favor of the modern OpenType math fonts > (e.g. Latin Modern Math, XITS Math, Cambria Math). There are > still some known bugs/limitations in Gecko, math fonts and > systems so that might take some time before the migration > can be fully done. However, note that we plan to remove > support for the old fonts in a near future, so please > upgrade your setup as soon as possible and report any issues > you find. ... Why should it be necessary to remove support for old fonts just because new things will be supported? -- Bill

Le 22/07/2014 08:51, William F Hammond a écrit : > Why should it be necessary to remove support for old fonts just > because new things will be supported? -- Bill I supposed you understood that it is not "just because new things will be supported" ;-) First the Unicode-only constructions in arbitrary fonts will still be supported, what we want to remove is the old font tables for Asana Math, "STIX General" and MathJax. Asana Math and the "STIX Word" set have an OpenType MATH table so can be used with the new code. Additionally, Latin Modern Math is a "modernized implementation of the Computer Modern fonts" (with reshaping and other stuff that are out of my area of expertise, see the papers on the GUST website for details) and so can now replace the "MathJax TeX" fonts (which is just generated via an autotracer, targeted to MathJax's needs and was a temporary solution to replace the old BaKoMa TeX support). So first, nobody should need these old fonts in the long term since you have equivalent OpenType MATH version... why do you want that?? The technical reason why we want to remove support for these old fonts is that they need special handlings which make the stretchy operator code "complicated" (to use an euphemism). Since we want to optimize it in the future, we will need to cleanup the code before refactoring it. Moving to OpenType MATH has actually been the plan for 7 years, but that finally could only be realized recently, thanks to the Ulule crowdfunding... Moreover, from the user's point of view, you will get a much more consistent way to use the various OpenType MATH fonts in WebKit/Gecko. Basically, you now have only one single Math font to deal with and only need to set the font-family on the <math> (and perhaps the companion fonts for the text). All the mathvariant selection, correction of prime size, math layout parameters, stretchy op constructions etc are stored in the font and Gecko&WebKit only need to read what has been specified by the font authors. One can even create his own math fonts with fontforge and they will be immediately supported by Gecko&WebKit without having to hardcode new data. For example, if some Mathematica people read that, they could just make an OpenType MATH version of their fonts and restore the original 15-years old support in Gecko :-) -- Frédéric Wang maths-informatique-jeux.com/blog/frederic

Dear all, Firefox 31 will be released in a couple of days, please find some release notes below. As usual, you can check https://wiki.mozilla.org/MathML:Home_Page#Last_bugs_fixed and https://developer.mozilla.org/en-US/Firefox/Releases/31#MathML for details. * Gecko 31: - Remove obsolete/non-standard features: font.mathfont-family* preferences, ::-moz-math-stretchy pseudo-element, STIX 1.0 beta support. - The scale correction is no longer always applied to stretchy operators, to follow the TeX / Open Type MATH behavior. - Fix some bugs with dotless i's and j's, one-argument mrow and invisible operators. - lspace/rspace now only have effect on embellished operators inside (inferred) mrow. - Fix a bug with the columnspacing of rtl mtable. - Use Unicode Mathematical alphanumeric characters for bold/italic/bold-italic mathvariants - Initial implementation of the MathVariants Table of the Open Type MATH table. See https://developer.mozilla.org/en-US/docs/Mozilla/MathML_Project/Fonts * Gecko 32: - Implement menclose notation "phasorangle" - Other improvements to the MathML stretchy code. * Gecko 33: - Add support for mtable@rowspacing/columnspacing/framespacing attributes - Use Open Type MATH constants for fractions, stacks, radicals and scripts -- Frédéric Wang maths-informatique-jeux.com/blog/frederic

I asked around and looked it up in some source (including Cajori), but to no avail. Any clues about where else to look? Thanks. Shervin On Thu, Jun 26, 2014 at 6:00 PM, Murray Sargent < murrays@exchange.microsoft.com> wrote: > Anyone know how "tie over infinity" ⧝ (29DD) is used in mathematics? > > Thanks, > Murray > >

Thank you for the information. Do you of any publications that I can see the symbol in context? ↪ Shervin On Thu, Jul 17, 2014 at 7:02 PM, George Weilenmann < george.weilenmann@insightsoftware.com> wrote: > Tie over infinity is a variant of the incomplete Infinity, and rarely > used to represent big infinity (uncountable infinity) > > > > *George Weilenmann* > > IT Help Desk | CRM Customization | Portal Developer > > *[image: Description: cid:82D8368C-56BA-48C5-ADFE-8F6F70761408] * > *Office +1 303.800.5027 <%2B1%20303.800.5027>| Mobile +1 720.940.0974 > <%2B1%20720.940.0974> *5613 DTC Parkway, Suite 570, Greenwood Village, CO > 80111 > > > > [image: Fans of InsightSoftware.com] > <http://www.insightsoftware.com/fans/?utm_source=Employees&utm_medium=Email&utm_content=Fans&utm_campaign=Signature> > > > > *From:* Shervin Afshar [mailto:shervinafshar@gmail.com] > *Sent:* Thursday, July 17, 2014 6:26 PM > *To:* Murray Sargent > *Cc:* www-math@w3.org; Barbara Beeton > *Subject:* Re: Tie over infinity > > > > > > I asked around and looked it up in some source (including Cajori), but to > no avail. Any clues about where else to look? Thanks. > > > > Shervin > > > > On Thu, Jun 26, 2014 at 6:00 PM, Murray Sargent < > murrays@exchange.microsoft.com> wrote: > > Anyone know how "tie over infinity" ⧝ (29DD) is used in mathematics? > > Thanks, > Murray > > > > Disclaimer: This message contains confidential information and is intended > only for the individual named. If you are not the named addressee you > should not disseminate, distribute or copy this email. Please notify the > sender immediately by email if you have received this email by mistake and > delete this email from your system. > > Our company accepts no liability for the content of this email, or for the > consequences of any actions taken on the basis of the information provided. > Any views or opinions presented in this email are solely those of the > author and do not necessarily represent those of the company. Finally, the > recipient should check this email and any attachments for the presence of > viruses. The company accepts no liability for any damage caused by any > virus transmitted by this email. > > InsightSoftware.com is a Trading Name of 'InsightSoftware.com > Limited',Registered in England No. 2860790 at 53-55 Uxbridge Road, Ealing, > London, W5 5SA, United Kingdom, VAT Number: GB 766 8160 95 >

The only time I have seen them in Context is in some mid 1800’s treatises for the incomplete infinity and in a few documents concerning uncountable infinity from the early 1900’s, I have not seen their usage in any mathematical texts after the 1950’s that I can recall. It would seem their usage fell out of favor. Instead Aleph and Beth Cardinal Numbers, א U+05D0 ב U+05D1 replaced Uncountable Infinities, and (INCOMPLETE INFINITY) U+29DC ⧜ or (PROPORTIONAL TO) U+221D ∝ or (INFINITY NEGATED WITH VERTICAL BAR) U+29DE for incomplete infinity. George Weilenmann IT Help Desk | CRM Customization | Portal Developer [Description: cid:82D8368C-56BA-48C5-ADFE-8F6F70761408] Office +1 303.800.5027| Mobile +1 720.940.0974 5613 DTC Parkway, Suite 570, Greenwood Village, CO 80111 [Fans of InsightSoftware.com]<http://www.insightsoftware.com/fans/?utm_source=Employees&utm_medium=Email&utm_content=Fans&utm_campaign=Signature> From: Shervin Afshar [mailto:shervinafshar@gmail.com] Sent: Sunday, July 20, 2014 5:49 PM To: George Weilenmann; www-math@w3.org Subject: Re: Tie over infinity Thank you for the information. Do you of any publications that I can see the symbol in context? ↪ Shervin On Thu, Jul 17, 2014 at 7:02 PM, George Weilenmann <george.weilenmann@insightsoftware.com<mailto:george.weilenmann@insightsoftware.com>> wrote: Tie over infinity is a variant of the incomplete Infinity, and rarely used to represent big infinity (uncountable infinity) George Weilenmann IT Help Desk | CRM Customization | Portal Developer Office +1 303.800.5027<tel:%2B1%20303.800.5027>| Mobile +1 720.940.0974<tel:%2B1%20720.940.0974> 5613 DTC Parkway, Suite 570, Greenwood Village, CO 80111 <http://www.insightsoftware.com/fans/?utm_source=Employees&utm_medium=Email&utm_content=Fans&utm_campaign=Signature> <http://www.insightsoftware.com/fans/?utm_source=Employees&utm_medium=Email&utm_content=Fans&utm_campaign=Signature> From: Shervin Afshar [mailto:shervinafshar@gmail.com] Sent: Thursday, July 17, 2014 6:26 PM To: Murray Sargent Cc: www-math@w3.org; Barbara Beeton Subject: Re: Tie over infinity<http://www.insightsoftware.com/fans/?utm_source=Employees&utm_medium=Email&utm_content=Fans&utm_campaign=Signature> <http://www.insightsoftware.com/fans/?utm_source=Employees&utm_medium=Email&utm_content=Fans&utm_campaign=Signature> <http://www.insightsoftware.com/fans/?utm_source=Employees&utm_medium=Email&utm_content=Fans&utm_campaign=Signature> I asked around and looked it up in some source (including Cajori), but to no avail. Any clues about where else to look? Thanks.<http://www.insightsoftware.com/fans/?utm_source=Employees&utm_medium=Email&utm_content=Fans&utm_campaign=Signature> <http://www.insightsoftware.com/fans/?utm_source=Employees&utm_medium=Email&utm_content=Fans&utm_campaign=Signature> Shervin<http://www.insightsoftware.com/fans/?utm_source=Employees&utm_medium=Email&utm_content=Fans&utm_campaign=Signature> <http://www.insightsoftware.com/fans/?utm_source=Employees&utm_medium=Email&utm_content=Fans&utm_campaign=Signature> On Thu, Jun 26, 2014 at 6:00 PM, Murray Sargent <murrays@exchange.microsoft.com> wrote:<http://www.insightsoftware.com/fans/?utm_source=Employees&utm_medium=Email&utm_content=Fans&utm_campaign=Signature> Anyone know how "tie over infinity" ⧝ (29DD) is used in mathematics? Thanks, Murray<http://www.insightsoftware.com/fans/?utm_source=Employees&utm_medium=Email&utm_content=Fans&utm_campaign=Signature> <http://www.insightsoftware.com/fans/?utm_source=Employees&utm_medium=Email&utm_content=Fans&utm_campaign=Signature> Disclaimer: This message contains confidential information and is intended only for the individual named. If you are not the named addressee you should not disseminate, distribute or copy this email. Please notify the sender immediately by email if you have received this email by mistake and delete this email from your system. Our company accepts no liability for the content of this email, or for the consequences of any actions taken on the basis of the information provided. Any views or opinions presented in this email are solely those of the author and do not necessarily represent those of the company. Finally, the recipient should check this email and any attachments for the presence of viruses. The company accepts no liability for any damage caused by any virus transmitted by this email. InsightSoftware.com is a Trading Name of 'InsightSoftware.com Limited',Registered in England No. 2860790 at 53-55 Uxbridge Road, Ealing, London, W5 5SA, United Kingdom, VAT Number: GB 766 8160 95<http://www.insightsoftware.com/fans/?utm_source=Employees&utm_medium=Email&utm_content=Fans&utm_campaign=Signature> <http://www.insightsoftware.com/fans/?utm_source=Employees&utm_medium=Email&utm_content=Fans&utm_campaign=Signature> Disclaimer: This message contains confidential information and is intended only for the individual named. If you are not the named addressee you should not disseminate, distribute or copy this email. Please notify the sender immediately by email if you have received this email by mistake and delete this email from your system. Our company accepts no liability for the content of this email, or for the consequences of any actions taken on the basis of the information provided. Any views or opinions presented in this email are solely those of the author and do not necessarily represent those of the company. Finally, the recipient should check this email and any attachments for the presence of viruses. The company accepts no liability for any damage caused by any virus transmitted by this email. InsightSoftware.com is a Trading Name of 'InsightSoftware.com Limited',Registered in England No. 2860790 at 53-55 Uxbridge Road, Ealing, London, W5 5SA, United Kingdom, VAT Number: GB 766 8160 95

Thank you for the background info, but none of these documents (including proposals to UTC; [1], [2]) have any background about this single symbol (which doesn't surprise me since it was proposed in a larger repertoire). So I don't think digging in the ISO archives would make us wiser about the background and usage. [1]: http://www.unicode.org/L2/L2000/00002-math.pdf [2]: http://www.unicode.org/L2/L2000/00119-math.pdf ↪Shervin On Fri, Jul 18, 2014 at 7:12 AM, <bnb@ams.org> wrote: > I asked around and looked it up in some source (including Cajori), but > to > no avail. Any clues about where else to look? Thanks. > > thanks for checking. > > Shervin > > my records show that it was included in > the stix requests from the isotech entity > set associated with sgml support facilties, > as listed in > > draft technical report ISO/IEC DTR 9573-13-1990(E) > ISO/IEC JTC 1 > Information technology -- SGML support facilities -- > Techniques for using SGML -- > Part 13 > Public entity sets for mathematics and science > > which was used as the starting set for the > stix collection. > > i'm not sure what the fate of this tech > report might be; i believe it was rearranged, > but whether it was ever voted into final > status, i don't know. what i *do* know, > though, is that it was adopted into the > w3c entity definitions for characters; see > > XML Entity Definitions for Characters (2nd Edition) > W3C Recommendation 10 April 2014 > http://www.w3.org/TR/xml-entity-names/ > > and in particular > > ISOTECH > http://www.w3.org/TR/xml-entity-names/isotech.html > > an update is in the works: > > XML Entity Definitions for Characters (3rd Edition) > W3C Working Draft 16 April 2014 > http://www.w3.org/2003/entities/2007doc/Overview.html > > with no change in the isotech complement. > > i correspond fairly regularly with both > the editors of this recommendation (both > of them are long-time friends), so could > ask them if they know anything more. but > this has been asked before, so i'm sure > they don't. > > i obtained the initial copy of 9573-13 > on paper, as a representative of ams and > ansi to what was then iso-iec jtc1/sc18/wg8, > the working group responsible for sgml. > anders berglund, at the time employed in > the iso headquarters in geneva, was also > a member of that wg, and when the stix > project was established, he provided me > with electronic copies of relevant > material, including the entity sets in > 8879 (sgml) and 9573-13. however, i'm > pretty sure he didn't/doesn't know the > source of the isotech material. (i think > i asked at the time, and remember being > told they were odds and ends from no > single source.) > > i don't know anyone currently in a > position to investigate old (very old!) > records at iso, but i'm still in contact > with a few people who might be able to > tell me how to find someone, if it is > worth the effort. let me know, please. > -- bb > > > On Thu, Jun 26, 2014 at 6:00 PM, Murray Sargent < > murrays@exchange.microsoft.com> wrote: > > > Anyone know how "tie over infinity" ? (29DD) is used in mathematics? > > > > Thanks, > > Murray >

Michael, this is an interesting question and I would love it if you manage to publish back answers when you have them. As far as I understand, you have no other way than really going native in this process since MathType equations are embedded objects. I would give a last chance from the flash side of the world, which has a tendency to be a tick less protective or to a trusted Java applet which does have full access to the clipboard (we were able to implement the MathML clipboard operations for example within the Wiris OpenMath Input Editor using that, for example). I would assume that the fine team at Design Science, two of them are on this list at least, would be able to dig out solutions, since they are the makers of MathType. I would expect solutions to be dependent on just about any combination of versions (OS, Word, MathType). paul On 12 juil. 2014, at 20:25, Michael <spdrmnky@pacbell.net> wrote: > Hello, Group - > > Professional Publications, Inc. (ppi2pass.com) publishes content in print, web, and e-book forms containing significant mathematical content. Originally, we were a TEX-based publisher, but we have been steeped in XML/MathML for about a decade. We are trying to implement a process whereby authors who have developed content in Microsoft Word with a MathType plug-in can cut and paste that content into a web form running on a Linux server. The goal is to retain all/most of the XML/MathML coding of equations, or to have post-processing where we end up with that coding after the submission. That's the bottom line. > > Retaining the coding after cut-and-paste on a Linux-platform has stumped us. (This is specifically a MathType issue, not a Word Equation Editor issue.) Can anyone in this group (a) offer a ray of hope, (b) point to something, or (c) make themselves available for consultation? > > I've been a member of this group for years, so I apologize if this is an off-topic post. We'll appreciate any guidance for moving forwards. > > Thank you. > > Michael R. Lindeburg, PE > President > Professional Publications, Inc. > mlindeburg@ppi2pass.com > (650) 593-9119 ext. 123 >

Neil, Thanks for the information. MathFlow tools are not downloadable for evaluation. That said, MathPlayer fails to render the following MathML, see screenshot attached. <math xmlns="http://www.w3.org/1998/Math/MathML"> <mstyle displaystyle="true"> <mstack stackalign="right"> <mscarries> <none/> <mn> 1 </mn> <none/> </mscarries> <msrow> <mn> 241 </mn> </msrow> <msrow> <mo> + </mo> <none/> <mn> 29 </mn> </msrow> <msline position="0" length="3"/> <msrow> <mn> 270 </mn> </msrow> </mstack> </mstyle> </math> Gregory ________________________________ From: neil.soiffer@gmail.com <neil.soiffer@gmail.com> on behalf of Neil Soiffer <NeilS@dessci.com> Sent: Thursday, June 26, 2014 7:03 PM To: David Carlisle Cc: Grégory Pakosz; www-math@w3.org Subject: Re: Questions about Elementary Math As noted, MathPlayer (which died at IE9 unless MS fixes a bug in enterprise mode -- enterprise mode was introduced a few months back), supports it. Along with MathPlayer, the MathFlow SDK tools (EquationComposer and DocumentComposer) also support it. That's where the rec's images came from. Neil On Wed, Jun 25, 2014 at 8:10 AM, David Carlisle <davidc@nag.co.uk<mailto:davidc@nag.co.uk>> wrote: On 24/06/2014 15:56, Grégory Pakosz wrote: Hello, I have two questions regarding elementary math as specified by MathML 3.0: 1) Is there a renderer out there that supports rendering additions, substractions, multiplications, and divisions with <mstack>, <mscarries>, and <mlongdiv> ? I failed to find one so far (downgrading IE to IE9 + installing a plugin isn't really future proof). Possibly currently only MathPlayer supports it natively, and as you indicate that is not available in current IE however it's possible to transform the markup to mathml2 for rendering in other clients. The MathJax "content mathml" extension and the firefox mathml-mml3ff addon both work by using some XSLT of mine to translate the markup to mathml2 mtable. https://code.google.com/p/web-xslt/source/browse/trunk/ctop Most of that content mathml to presentation transformation has also been re-encoded in javascript to avoid the XSLT stage (which is very slow in chrome) although not currently the elementary math part, that shouldn't be hard to add, given some time. 2) Despite being XML, <mstack> relies on children order instead of named elements like <dividend>, <divisor>, <quotient>. What's the rationale behind this choice? Positional children are used quite a lot in the mathml design: mfrac msub etc also do not have named arguments. Thanks you, Gregory David

Nevermind, I didn't reboot my Windows 7 + IE9 VM after having installed MathPlayer since instructions said I just had to relaunch IE. In any case, is there a way to adjust the position of the divisor and the quotient? My use case is the following: I recognize handwritten elementary math operations and I would like to position everything with respect to what's been written. In the screenshot attached, to achieve that, I would need to shift ":891 = 12" a bit to the left. Thanks. Gregory ________________________________ From: Grégory Pakosz Sent: Thursday, July 03, 2014 5:54 PM To: Neil Soiffer; David Carlisle Cc: www-math@w3.org Subject: RE: Questions about Elementary Math Neil, Thanks for the information. MathFlow tools are not downloadable for evaluation. That said, MathPlayer fails to render the following MathML, see screenshot attached. <math xmlns="http://www.w3.org/1998/Math/MathML"> <mstyle displaystyle="true"> <mstack stackalign="right"> <mscarries> <none/> <mn> 1 </mn> <none/> </mscarries> <msrow> <mn> 241 </mn> </msrow> <msrow> <mo> + </mo> <none/> <mn> 29 </mn> </msrow> <msline position="0" length="3"/> <msrow> <mn> 270 </mn> </msrow> </mstack> </mstyle> </math> Gregory ________________________________ From: neil.soiffer@gmail.com <neil.soiffer@gmail.com> on behalf of Neil Soiffer <NeilS@dessci.com> Sent: Thursday, June 26, 2014 7:03 PM To: David Carlisle Cc: Grégory Pakosz; www-math@w3.org Subject: Re: Questions about Elementary Math As noted, MathPlayer (which died at IE9 unless MS fixes a bug in enterprise mode -- enterprise mode was introduced a few months back), supports it. Along with MathPlayer, the MathFlow SDK tools (EquationComposer and DocumentComposer) also support it. That's where the rec's images came from. Neil On Wed, Jun 25, 2014 at 8:10 AM, David Carlisle <davidc@nag.co.uk<mailto:davidc@nag.co.uk>> wrote: On 24/06/2014 15:56, Grégory Pakosz wrote: Hello, I have two questions regarding elementary math as specified by MathML 3.0: 1) Is there a renderer out there that supports rendering additions, substractions, multiplications, and divisions with <mstack>, <mscarries>, and <mlongdiv> ? I failed to find one so far (downgrading IE to IE9 + installing a plugin isn't really future proof). Possibly currently only MathPlayer supports it natively, and as you indicate that is not available in current IE however it's possible to transform the markup to mathml2 for rendering in other clients. The MathJax "content mathml" extension and the firefox mathml-mml3ff addon both work by using some XSLT of mine to translate the markup to mathml2 mtable. https://code.google.com/p/web-xslt/source/browse/trunk/ctop Most of that content mathml to presentation transformation has also been re-encoded in javascript to avoid the XSLT stage (which is very slow in chrome) although not currently the elementary math part, that shouldn't be hard to add, given some time. 2) Despite being XML, <mstack> relies on children order instead of named elements like <dividend>, <divisor>, <quotient>. What's the rationale behind this choice? Positional children are used quite a lot in the mathml design: mfrac msub etc also do not have named arguments. Thanks you, Gregory David

You shouldn't have needed to reboot, but who knows what else was going on... A reboot never hurts :-) There's no way to adjust the position using mlongdiv attrs. I tried playing with mpadded, but because MathPlayer places the : and =, those tricks don't work. So I'm sorry to say there doesn't seem to be a way to change that. If you have suggestions for improving mlongdiv such as adding a tight=true/false for making the divisor tight against the dividend vs what MathPlayer currently does where it pushes it to the right of the "division stack" (my name -- maybe there's an official name for that). Of course, tight would only apply to those longdivstyles where it is sensible. I'm not saying that "tight" is the right attribute to add, but just throwing it out as a possibility. Neil On Thu, Jul 3, 2014 at 9:31 AM, Grégory Pakosz <gpakosz@myscript.com> wrote: > Nevermind, > > > I didn't reboot my Windows 7 + IE9 VM after having installed MathPlayer > since instructions said I just had to relaunch IE. > > > In any case, is there a way to adjust the position of the divisor and > the quotient? > > > My use case is the following: I recognize handwritten elementary math > operations and I would like to position everything with respect to what's > been written. In the screenshot attached, to achieve that, I would need to > shift ":891 = 12" a bit to the left. > > > Thanks. > > Gregory > > > ------------------------------ > *From:* Grégory Pakosz > *Sent:* Thursday, July 03, 2014 5:54 PM > *To:* Neil Soiffer; David Carlisle > *Cc:* www-math@w3.org > > *Subject:* RE: Questions about Elementary Math > > > Neil, > > > Thanks for the information. MathFlow tools are not downloadable for > evaluation. > > > That said, MathPlayer fails to render the following MathML, see > screenshot attached. > > > <math xmlns="http://www.w3.org/1998/Math/MathML"> > <mstyle displaystyle="true"> > <mstack stackalign="right"> > <mscarries> > <none/> > <mn> 1 </mn> > <none/> > </mscarries> > <msrow> > <mn> 241 </mn> > </msrow> > <msrow> > <mo> + </mo> > <none/> > <mn> 29 </mn> > </msrow> > <msline position="0" length="3"/> > <msrow> > <mn> 270 </mn> > </msrow> > </mstack> > </mstyle> > </math> > > > Gregory > > > ------------------------------ > *From:* neil.soiffer@gmail.com <neil.soiffer@gmail.com> on behalf of Neil > Soiffer <NeilS@dessci.com> > *Sent:* Thursday, June 26, 2014 7:03 PM > *To:* David Carlisle > *Cc:* Grégory Pakosz; www-math@w3.org > *Subject:* Re: Questions about Elementary Math > > As noted, MathPlayer (which died at IE9 unless MS fixes a bug in > enterprise mode -- enterprise mode was introduced a few months back), > supports it. Along with MathPlayer, the MathFlow SDK tools > (EquationComposer and DocumentComposer) also support it. That's where the > rec's images came from. > > Neil > > > > On Wed, Jun 25, 2014 at 8:10 AM, David Carlisle <davidc@nag.co.uk> wrote: > >> On 24/06/2014 15:56, Grégory Pakosz wrote: >> >> Hello, >> >> >> I have two questions regarding elementary math as specified by MathML >> 3.0: >> >> >> 1) Is there a renderer out there that supports rendering additions, >> substractions, multiplications, and divisions with <mstack>, <mscarries>, >> and <mlongdiv> ? I failed to find one so far (downgrading IE to IE9 + >> installing a plugin isn't really future proof). >> >> >> >> Possibly currently only MathPlayer supports it natively, and as you >> indicate that is not available in current IE >> however it's possible to transform the markup to mathml2 for rendering >> in other clients. >> >> The MathJax "content mathml" extension and the firefox mathml-mml3ff >> addon both work by using some XSLT of mine >> to translate the markup to mathml2 mtable. >> >> https://code.google.com/p/web-xslt/source/browse/trunk/ctop >> >> Most of that content mathml to presentation transformation has also been >> re-encoded in javascript to avoid the XSLT stage >> (which is very slow in chrome) although not currently the elementary math >> part, that shouldn't be hard to add, given some time. >> >> >> >> >> >> 2) Despite being XML, <mstack> relies on children order instead of >> named elements like <dividend>, <divisor>, <quotient>. What's the rationale >> behind this choice? >> >> >> >> Positional children are used quite a lot in the mathml design: mfrac >> msub etc also do not have named arguments. >> >> Thanks you, >> >> Gregory >> >> >> David >> >> >

Neil, In the screen shot attached of a Dutch division, maybe a position attribute on the quotient <msrow> would do the job. Reminder: this is a handwritten division recognized with our tech and I serialize the interpretation into MathML. I was able to generate a <msgroup> to properly align 345 with -42 (as written on paper). But the original ink shows that the \ sign is aligned with the 3 of 23 and the 3 of 345. Then, the 91 of 291 is aligned with the 45 of 345. Adding a <none/> before 291 AND having a position to offset the quotient would enable to achieve that alignement. But that's just reasoning on a single example... Not sure how well it generalizes. Regards, Gregory ________________________________ From: neil.soiffer@gmail.com <neil.soiffer@gmail.com> on behalf of Neil Soiffer <NeilS@dessci.com> Sent: Thursday, July 03, 2014 7:50 PM To: Grégory Pakosz Cc: Neil Soiffer; David Carlisle; www-math@w3.org Subject: Re: Questions about Elementary Math You shouldn't have needed to reboot, but who knows what else was going on... A reboot never hurts :-) There's no way to adjust the position using mlongdiv attrs. I tried playing with mpadded, but because MathPlayer places the : and =, those tricks don't work. So I'm sorry to say there doesn't seem to be a way to change that. If you have suggestions for improving mlongdiv such as adding a tight=true/false for making the divisor tight against the dividend vs what MathPlayer currently does where it pushes it to the right of the "division stack" (my name -- maybe there's an official name for that). Of course, tight would only apply to those longdivstyles where it is sensible. I'm not saying that "tight" is the right attribute to add, but just throwing it out as a possibility. Neil On Thu, Jul 3, 2014 at 9:31 AM, Grégory Pakosz <gpakosz@myscript.com<mailto:gpakosz@myscript.com>> wrote: Nevermind, I didn't reboot my Windows 7 + IE9 VM after having installed MathPlayer since instructions said I just had to relaunch IE. In any case, is there a way to adjust the position of the divisor and the quotient? My use case is the following: I recognize handwritten elementary math operations and I would like to position everything with respect to what's been written. In the screenshot attached, to achieve that, I would need to shift ":891 = 12" a bit to the left. Thanks. Gregory ________________________________ From: Grégory Pakosz Sent: Thursday, July 03, 2014 5:54 PM To: Neil Soiffer; David Carlisle Cc: www-math@w3.org<mailto:www-math@w3.org> Subject: RE: Questions about Elementary Math Neil, Thanks for the information. MathFlow tools are not downloadable for evaluation. That said, MathPlayer fails to render the following MathML, see screenshot attached. <math xmlns="http://www.w3.org/1998/Math/MathML"> <mstyle displaystyle="true"> <mstack stackalign="right"> <mscarries> <none/> <mn> 1 </mn> <none/> </mscarries> <msrow> <mn> 241 </mn> </msrow> <msrow> <mo> + </mo> <none/> <mn> 29 </mn> </msrow> <msline position="0" length="3"/> <msrow> <mn> 270 </mn> </msrow> </mstack> </mstyle> </math> Gregory ________________________________ From: neil.soiffer@gmail.com<mailto:neil.soiffer@gmail.com> <neil.soiffer@gmail.com<mailto:neil.soiffer@gmail.com>> on behalf of Neil Soiffer <NeilS@dessci.com<mailto:NeilS@dessci.com>> Sent: Thursday, June 26, 2014 7:03 PM To: David Carlisle Cc: Grégory Pakosz; www-math@w3.org<mailto:www-math@w3.org> Subject: Re: Questions about Elementary Math As noted, MathPlayer (which died at IE9 unless MS fixes a bug in enterprise mode -- enterprise mode was introduced a few months back), supports it. Along with MathPlayer, the MathFlow SDK tools (EquationComposer and DocumentComposer) also support it. That's where the rec's images came from. Neil On Wed, Jun 25, 2014 at 8:10 AM, David Carlisle <davidc@nag.co.uk<mailto:davidc@nag.co.uk>> wrote: On 24/06/2014 15:56, Grégory Pakosz wrote: Hello, I have two questions regarding elementary math as specified by MathML 3.0: 1) Is there a renderer out there that supports rendering additions, substractions, multiplications, and divisions with <mstack>, <mscarries>, and <mlongdiv> ? I failed to find one so far (downgrading IE to IE9 + installing a plugin isn't really future proof). Possibly currently only MathPlayer supports it natively, and as you indicate that is not available in current IE however it's possible to transform the markup to mathml2 for rendering in other clients. The MathJax "content mathml" extension and the firefox mathml-mml3ff addon both work by using some XSLT of mine to translate the markup to mathml2 mtable. https://code.google.com/p/web-xslt/source/browse/trunk/ctop Most of that content mathml to presentation transformation has also been re-encoded in javascript to avoid the XSLT stage (which is very slow in chrome) although not currently the elementary math part, that shouldn't be hard to add, given some time. 2) Despite being XML, <mstack> relies on children order instead of named elements like <dividend>, <divisor>, <quotient>. What's the rationale behind this choice? Positional children are used quite a lot in the mathml design: mfrac msub etc also do not have named arguments. Thanks you, Gregory David

On 05/07/2014 22:45, William F Hammond wrote: > > Not that you asked, but your original addition example > can be rendered using only CSS (no scripting): > > http://www.albany.edu/~hammond/demos/purecss/mstackfs.html > > > -- Bill > > Sure, similar CSS rules are given in http://www.w3.org/TR/mathml-for-css/#mstack but the mathml css profile has to have so many restrictions on the input for it to work it really doesn't scale well. Using newer css notably flexbox you could probably do a bit more in browsers that support it but still the preferred result is that browsers implement mathml natively, we are not there yet for all browsers as you know, but these things take time... David

I appreciate the input. But I'm not really interested in rendering MathML per se, be it with the help of CSS or any other "hack". In the context of handwriting recognition, what I want is to get a sense of which MathML markup I should produce that could be consumed by any renderer and achieves alignment similar to what's been written. I attached yet another screenshot in which you can see a handwritten long division in the center, on the left is a LaTeX render achieved using array and on the right is the MathML markup I'm generating so far. As you can see, the fact that the decimal dot takes a column makes me insert <none/> elements to keep things aligned. Gregory ________________________________________ From: William F Hammond <hammond@csc.albany.edu> Sent: Saturday, July 05, 2014 11:45 PM To: Grégory Pakosz Cc: Neil Soiffer; www-math@w3.org Subject: Re: Questions about Elementary Math Not that you asked, but your original addition example can be rendered using only CSS (no scripting): http://www.albany.edu/~hammond/demos/purecss/mstackfs.html -- Bill

Le 22/06/2014 19:31, David Carlisle a écrit : > > If you don't follow updates to the unicode.xml file used as the source > for entity definitions in mathml and html > skip this message:-) > > Christian's recent questions about unicode.xml (and some recent bug > reports about unicode-math latex package) > prompted me to look again at the tex mappings in unicode.xml. > > The existing ones were mostly speculative assignments dating from the > 1990's some years before the bulk of > math characters were added to Unicode. > > I have extended the schema to allow multiple <latex> and <mathlatex> > elements so the file can track different mappings, > and added a set attribute do distinguish these. So that now for > example there are entries such as > > <mathlatex set="unicode-math">\rightarrow</mathlatex> > > for U+"2192" . > > This mathlatex set="unicode-math" set is mechanically extracted from > the source of the unicode-math package > (the principle method for using unicode math fonts with xelatex and > lualatex) > https://github.com/wspr/unicode-math/blob/master/unicode-math-table.tex > > So, while I'm not sure I like all the mappings here they correspond to > running TeX code which is a definite improvement > over the previous ones. > > Frédéric Wang reported some problems with the TeX mappings a while > back I haven't fixed those yet, I may just remove them > in favour of this new set, or perhaps this set and a set derived from > a package for classic TeX (amssymb or stix-latex) > comments welcome on the best plan of action here.... As I recall, the last time I checked unicode.xml, there were several sets with different mappings (AMS, IIEE etc) and some of them clearly had mistakes, so that was a bit messy. Personally, I don't mind dropping the old TeX mappings as long as a clear list of TeX mapping for math char commands is provided, on which one can rely on. -- Frédéric Wang maths-informatique-jeux.com/blog/frederic

On 02/07/2014 16:22, Frédéric WANG wrote: > As I recall, the last time I checked unicode.xml, there were several > sets with different mappings (AMS, IIEE etc) and some of them clearly > had mistakes, so that was a bit messy. Yes but (unless I broke them) the "publisher" entries in the file are historical data relating to publishers internal character tables as input data to forming the stix submission to unicode so I'm not sure I can change them now (but on the other hand they are not really that useful to anyone apart from historians:-) > Personally, I don't mind dropping the old TeX mappings as long as a > clear list of TeX mapping for math char commands is provided, on which > one can rely on. Yes I think I'm going to drop the "nameless" <latex> entries and try to maintain entries like the new ones that specifically reference a latex package such as unicode-math so that if you load that package (or say you are emulating that package) the names should work (even if one can argue about the actual names) David

David, On 7/2/14 11:38 AM, David Carlisle wrote: > On 02/07/2014 16:22, Frédéric WANG wrote: >> As I recall, the last time I checked unicode.xml, there were several >> sets with different mappings (AMS, IIEE etc) and some of them clearly >> had mistakes, so that was a bit messy. > > Yes but (unless I broke them) the "publisher" entries in the file are > historical data relating to publishers internal character tables as > input data to forming > the stix submission to unicode so I'm not sure I can change them now > (but on the other hand they are not really that useful to anyone apart > from historians:-) > It's a fine contribution to get labelled definite mappings into unicode.xml, and of course to check them. We must all be very grateful for your doing that. > >> Personally, I don't mind dropping the old TeX mappings as long as a >> clear list of TeX mapping for math char commands is provided, on >> which one can rely on. > > Yes I think I'm going to drop the "nameless" <latex> entries and try > to maintain entries like the new ones that specifically reference a > latex package such as > unicode-math so that if you load that package (or say you are > emulating that package) the names should work (even if one can argue > about the actual names) > I would urge you to change the "nameless" <latex> entries to a 'set="latex-historical">' at least so as not to lose the record of where people may have got their authoritative info. That seems to me no particular overhead and may clarify bugs reported later. Those only concerned to be up-to-date can ignore that info as they will all sorts of other material. Presumably names can be reported as occurring in several sets (packages). All the best, Patrick > David > > -- Associate Editor Emeritus, Mathematical Reviews [1] 734-769-2015 <ion@ams.org> Mathematical Reviews (MR) <http://www.ams.org/mr-database/> <http://www.ams.org/publications/60ann/AnniversaryYear.html> University of Michigan [mailto:p@u.edu where p=pion and u=umich] <http://www-personal.umich.edu/~pion/> Geometry and DFT <http://www.ams.org/samplings/feature-column/fcarc-geo-dft> W3C Math Working Group Co-Chair <http://www.w3.org/Math/> MathML3 <http://www.w3.org/TR/MathML3/> XML Entity Definitions for Characters <http://www.w3.org/TR/xml-entity-names/> MathML for CSS Profile <http://www.w3.org/TR/mathml-for-css/> Math on the Web pages <http://www.mathontheweb.org/> (replacing http://www.ams.org/mathweb/) MSC2010 Revision <http://msc2010.org/> MSC2010 as SKOS <http://msc2010.org/resources/MSC/2010/info/> MSC2010 as a TiddlyWiki <http://msc2010.org/MSC-2010-server.html> GLC NARGS <http://glcnargs.com>

On 03/07/2014 13:49, Patrick Ion wrote: > I would urge you to change the "nameless" <latex> entries to a > 'set="latex-historical">' at least so as not to lose the record of > where people may have got their authoritative info yes but then I should fix them to be right (not with obvious typos such as Frédéric reported:-) Ok will do something like that... David

On 7/3/14 9:31 AM, David Carlisle wrote: > On 03/07/2014 13:49, Patrick Ion wrote: >> I would urge you to change the "nameless" <latex> entries to a >> 'set="latex-historical">' at least so as not to lose the record of >> where people may have got their authoritative info > > > yes but then I should fix them to be right (not with obvious typos > such as > Frédéric reported:-) > > Ok will do something like that... > > David The trouble is that, in some sense, it's useful to know what the long persistent typo was---at least sometimes. Essentially that's why I'm advocating keeping the historical, and maybe superseded, information about, but clearly labelled. The correction presumably follows implicitly if one uses a good 'current' set as basis in an application. That's what 'good current' is supposed to mean and a motivation for developers keeping up-to-date. Patrick P.S. You may recollect that it was clear that the early SGML names from a report by American publishers, that went on into ISO12083, could be seen to be lifted without attribution from casual advertising of TeX names put out by the AMS because they reproduced the typos in that list.

On 03/07/2014 14:49, Patrick Ion wrote: > P.S. You may recollect that it was clear that the early SGML > names from a report by American publishers, that went on > into ISO12083, could be seen to be lifted without attribution > from casual advertising of TeX names put out by the AMS > because they reproduced the typos in that list. Oh Ok that's simpler: don't fix the names in the file just mark then as set="historical" OK good plan! David