Mailing List Archive: Wikipedia: Wikitech

Long term strategy for math on wikipedia

1 2

View All

Index | Next | Previous | View Threaded

mflaschen at wikimedia

Jul 29, 2013, 8:14 PM

Post #26 of 33 (46 views)
Permalink

Re: Long term strategy for math on wikipedia [In reply to]

On 07/21/2013 08:53 PM, Peter Krautzberger wrote:
> 2) TeX/LaTeX compatibility might be lost.
>
> "Native" content (e.g. <maction> or even subexpression links) has no
> counterpart in TeX. Conservative extensions of TeX can easily enable this
> kind of content but backward compatibility will be lost.

That only applies if MathML becomes the definitive format/source code
(which is currently TeX). If that happens, it will be well down the road.

Matt Flaschen

_______________________________________________
Wikitech-l mailing list
Wikitech-l [at] lists
https://lists.wikimedia.org/mailman/listinfo/wikitech-l

peter.krautzberger at mathjax

Jul 29, 2013, 9:02 PM

Post #27 of 33 (45 views)
Permalink

Re: Long term strategy for math on wikipedia [In reply to]

@Matthew Agreed, that's down the road (but I did call the thread "long
term" :) ).

There is the question if texvc could (should?) be replaced. From what I
understand it's a pain for people to set up (installing texlive, compiling
texvc etc), and leaving it behind could help several internationalization
bugs (like the one praveenp linked to previously).

Peter.

On Mon, Jul 29, 2013 at 8:14 PM, Matthew Flaschen
<mflaschen [at] wikimedia>wrote:

> On 07/21/2013 08:53 PM, Peter Krautzberger wrote:
> > 2) TeX/LaTeX compatibility might be lost.
> >
> > "Native" content (e.g. <maction> or even subexpression links) has no
> > counterpart in TeX. Conservative extensions of TeX can easily enable this
> > kind of content but backward compatibility will be lost.
>
> That only applies if MathML becomes the definitive format/source code
> (which is currently TeX). If that happens, it will be well down the road.
>
> Matt Flaschen
>
> _______________________________________________
> Wikitech-l mailing list
> Wikitech-l [at] lists
> https://lists.wikimedia.org/mailman/listinfo/wikitech-l
>
_______________________________________________
Wikitech-l mailing list
Wikitech-l [at] lists
https://lists.wikimedia.org/mailman/listinfo/wikitech-l

delirium at hackish

Aug 2, 2013, 8:36 AM

Post #28 of 33 (36 views)
Permalink

Re: Long term strategy for math on wikipedia [In reply to]

On 7/22/13 2:53 AM, Peter Krautzberger wrote:
> 2) TeX/LaTeX compatibility might be lost.
>
> "Native" content (e.g. <maction> or even subexpression links) has no
> counterpart in TeX. Conservative extensions of TeX can easily enable this
> kind of content but backward compatibility will be lost.
>

If this means MathML as the canonical format, i.e. people write MathML
into articles directly, rather than it just being an output/rendering
format, that gives me moderate worry:

1. From the perspective of being able to repurpose Wikipedia articles
outside of a web context, TeX-format equations are nice for articles in
the math/science sphere, since TeX-based publishing workflows are common
in math/science, and equations are particularly tedious and error-prone
to convert by hand, if that would end up necessary. Admittedly, in some
workflows there's no real difference: you can import both MathML and TeX
equations into MS Word with appropriate plugins (I haven't looked into
whether the two import paths differ on compatibility). Perhaps as
HTML-based print workflows improve this will drop off as an issue, but
right now only a smallish proportion of people are using workflows based
on something like PrinceXML, and the free-software alternatives to
PrinceXML are further behind.

2. From a wikitext-readability perspective, TeX-format equations are the
de-facto standard way of ASCII-fying equations, e.g. in plaintext
emails, while MathML isn't written in a syntax any humans normally
write. So using TeX as our underlying representation makes equations
possible to edit in text form, at least for people who already
professionally work in areas where that's common, while MathML equations
virtually require a visual editor (unless the idea is to use something
like ASCIIMathML?).

-Mark

_______________________________________________
Wikitech-l mailing list
Wikitech-l [at] lists
https://lists.wikimedia.org/mailman/listinfo/wikitech-l

me.praveen at gmail

Aug 2, 2013, 10:07 AM

Post #29 of 33 (36 views)
Permalink

Re: Long term strategy for math on wikipedia [In reply to]

On Friday 02 August 2013 09:06 PM, Delirium wrote:
> On 7/22/13 2:53 AM, Peter Krautzberger wrote:
>> 2) TeX/LaTeX compatibility might be lost.
>>
>> "Native" content (e.g. <maction> or even subexpression links) has no
>> counterpart in TeX. Conservative extensions of TeX can easily enable
>> this
>> kind of content but backward compatibility will be lost.
>>
>
> If this means MathML as the canonical format, i.e. people write MathML
> into articles directly, rather than it just being an output/rendering
> format, that gives me moderate worry:
>
> 1. From the perspective of being able to repurpose Wikipedia articles
> outside of a web context, TeX-format equations are nice for articles
> in the math/science sphere, since TeX-based publishing workflows are
> common in math/science, and equations are particularly tedious and
> error-prone to convert by hand, if that would end up necessary.
> Admittedly, in some workflows there's no real difference: you can
> import both MathML and TeX equations into MS Word with appropriate
> plugins (I haven't looked into whether the two import paths differ on
> compatibility). Perhaps as HTML-based print workflows improve this
> will drop off as an issue, but right now only a smallish proportion of
> people are using workflows based on something like PrinceXML, and the
> free-software alternatives to PrinceXML are further behind.
>
> 2. From a wikitext-readability perspective, TeX-format equations are
> the de-facto standard way of ASCII-fying equations, e.g. in plaintext
> emails, while MathML isn't written in a syntax any humans normally
> write. So using TeX as our underlying representation makes equations
> possible to edit in text form, at least for people who already
> professionally work in areas where that's common, while MathML
> equations virtually require a visual editor (unless the idea is to use
> something like ASCIIMathML?).
>
> -Mark

What??!!?? sorry I didn't get a thing from this. :-)

Current scenario is: In our current Math extension, textvc is simply
unable to generate equations in png except Latin languages. Also Mathjax
is heavily client dependent (Unsupportably dependent) and has its own
serious bugs.

>
>
> _______________________________________________
> Wikitech-l mailing list
> Wikitech-l [at] lists
> https://lists.wikimedia.org/mailman/listinfo/wikitech-l

_______________________________________________
Wikitech-l mailing list
Wikitech-l [at] lists
https://lists.wikimedia.org/mailman/listinfo/wikitech-l

delirium at hackish

Aug 2, 2013, 10:40 AM

Post #30 of 33 (34 views)
Permalink

Re: Long term strategy for math on wikipedia [In reply to]

On 8/2/13 7:07 PM, praveenp wrote:
>
> On Friday 02 August 2013 09:06 PM, Delirium wrote:
>> On 7/22/13 2:53 AM, Peter Krautzberger wrote:
>>> 2) TeX/LaTeX compatibility might be lost.
>>>
>>> "Native" content (e.g. <maction> or even subexpression links) has no
>>> counterpart in TeX. Conservative extensions of TeX can easily enable
>>> this
>>> kind of content but backward compatibility will be lost.
>>>
>>
>> If this means MathML as the canonical format, i.e. people write
>> MathML into articles directly, rather than it just being an
>> output/rendering format, that gives me moderate worry:
>>
>> 1. From the perspective of being able to repurpose Wikipedia articles
>> outside of a web context, TeX-format equations are nice for articles
>> in the math/science sphere, since TeX-based publishing workflows are
>> common in math/science, and equations are particularly tedious and
>> error-prone to convert by hand, if that would end up necessary.
>> Admittedly, in some workflows there's no real difference: you can
>> import both MathML and TeX equations into MS Word with appropriate
>> plugins (I haven't looked into whether the two import paths differ on
>> compatibility). Perhaps as HTML-based print workflows improve this
>> will drop off as an issue, but right now only a smallish proportion
>> of people are using workflows based on something like PrinceXML, and
>> the free-software alternatives to PrinceXML are further behind.
>>
>> 2. From a wikitext-readability perspective, TeX-format equations are
>> the de-facto standard way of ASCII-fying equations, e.g. in plaintext
>> emails, while MathML isn't written in a syntax any humans normally
>> write. So using TeX as our underlying representation makes equations
>> possible to edit in text form, at least for people who already
>> professionally work in areas where that's common, while MathML
>> equations virtually require a visual editor (unless the idea is to
>> use something like ASCIIMathML?).
> What??!!?? sorry I didn't get a thing from this. :-)
>
> Current scenario is: In our current Math extension, textvc is simply
> unable to generate equations in png except Latin languages. Also
> Mathjax is heavily client dependent (Unsupportably dependent) and has
> its own serious bugs.

I read Peter's point 2 as discussing the possible "native" use of MathML
tags, i.e. permitting people to write MathML into articles, rather than
only using MathML as an alternate rendering path for texvc/MathJax/etc.
If MathML is a render-only target, then "TeX/LaTeX compatibility might
be lost" doesn't seem like it could be an issue. So unless I'm totally
misreading, I took the discussion to be about allowing MathML in
articles, which could break TeX compatibility since not all MathML tags
can be rendered back into TeX equivalents. The two points above are my
two concerns w.r.t. that suggestion. Am I misreading the suggestion
entirely?

-Mark

_______________________________________________
Wikitech-l mailing list
Wikitech-l [at] lists
https://lists.wikimedia.org/mailman/listinfo/wikitech-l

peter.krautzberger at mathjax

Aug 2, 2013, 12:04 PM

Post #31 of 33 (34 views)
Permalink

Re: Long term strategy for math on wikipedia [In reply to]

@Mark Just to clarify. Personally, I don't think wikitext's math format
should move away from a TeX-like input language. The point I was trying
making was that a conservative extension would be useful if MathML becomes
a desired output. It seems to me that texvc was specifically designed to
prevent fully fledged TeX input, so I wonder if it wouldn't help everyone
if wasn't required on the backend anymore, only that the syntax stayed
backward compatible.

@paveenp I don't know what you mean by "unsupportably dependent". I am also
not aware of "serious bugs". Could you be more specific?

Peter.

On Fri, Aug 2, 2013 at 10:40 AM, Delirium <delirium [at] hackish> wrote:

> On 8/2/13 7:07 PM, praveenp wrote:
>
>>
>> On Friday 02 August 2013 09:06 PM, Delirium wrote:
>>
>>> On 7/22/13 2:53 AM, Peter Krautzberger wrote:
>>>
>>>> 2) TeX/LaTeX compatibility might be lost.
>>>>
>>>> "Native" content (e.g. <maction> or even subexpression links) has no
>>>> counterpart in TeX. Conservative extensions of TeX can easily enable
>>>> this
>>>> kind of content but backward compatibility will be lost.
>>>>
>>>>
>>> If this means MathML as the canonical format, i.e. people write MathML
>>> into articles directly, rather than it just being an output/rendering
>>> format, that gives me moderate worry:
>>>
>>> 1. From the perspective of being able to repurpose Wikipedia articles
>>> outside of a web context, TeX-format equations are nice for articles in the
>>> math/science sphere, since TeX-based publishing workflows are common in
>>> math/science, and equations are particularly tedious and error-prone to
>>> convert by hand, if that would end up necessary. Admittedly, in some
>>> workflows there's no real difference: you can import both MathML and TeX
>>> equations into MS Word with appropriate plugins (I haven't looked into
>>> whether the two import paths differ on compatibility). Perhaps as
>>> HTML-based print workflows improve this will drop off as an issue, but
>>> right now only a smallish proportion of people are using workflows based on
>>> something like PrinceXML, and the free-software alternatives to PrinceXML
>>> are further behind.
>>>
>>> 2. From a wikitext-readability perspective, TeX-format equations are the
>>> de-facto standard way of ASCII-fying equations, e.g. in plaintext emails,
>>> while MathML isn't written in a syntax any humans normally write. So using
>>> TeX as our underlying representation makes equations possible to edit in
>>> text form, at least for people who already professionally work in areas
>>> where that's common, while MathML equations virtually require a visual
>>> editor (unless the idea is to use something like ASCIIMathML?).
>>>
>> What??!!?? sorry I didn't get a thing from this. :-)
>>
>>
>> Current scenario is: In our current Math extension, textvc is simply
>> unable to generate equations in png except Latin languages. Also Mathjax is
>> heavily client dependent (Unsupportably dependent) and has its own serious
>> bugs.
>>
>
> I read Peter's point 2 as discussing the possible "native" use of MathML
> tags, i.e. permitting people to write MathML into articles, rather than
> only using MathML as an alternate rendering path for texvc/MathJax/etc. If
> MathML is a render-only target, then "TeX/LaTeX compatibility might be
> lost" doesn't seem like it could be an issue. So unless I'm totally
> misreading, I took the discussion to be about allowing MathML in articles,
> which could break TeX compatibility since not all MathML tags can be
> rendered back into TeX equivalents. The two points above are my two
> concerns w.r.t. that suggestion. Am I misreading the suggestion entirely?
>
> -Mark
>
>
>
> ______________________________**_________________
> Wikitech-l mailing list
> Wikitech-l [at] lists
> https://lists.wikimedia.org/**mailman/listinfo/wikitech-l<https://lists.wikimedia.org/mailman/listinfo/wikitech-l>
>
_______________________________________________
Wikitech-l mailing list
Wikitech-l [at] lists
https://lists.wikimedia.org/mailman/listinfo/wikitech-l

me.praveen at gmail

Aug 6, 2013, 10:53 AM

Post #32 of 33 (25 views)
Permalink

Re: Long term strategy for math on wikipedia [In reply to]

I've problems with browsers like IE (Mainly XP) and opera (ubuntu
12.04/Mint Maya), although I forgot exact version numbers. And also it
takes each code points independently so it converts rtl language to ltr
language, or breaks any ligatures etc. (Aren't they serious bugs?)

On Saturday 03 August 2013 12:34:56 AM IST, Peter Krautzberger wrote:
> @Mark Just to clarify. Personally, I don't think wikitext's math format
> should move away from a TeX-like input language. The point I was trying
> making was that a conservative extension would be useful if MathML becomes
> a desired output. It seems to me that texvc was specifically designed to
> prevent fully fledged TeX input, so I wonder if it wouldn't help everyone
> if wasn't required on the backend anymore, only that the syntax stayed
> backward compatible.
>
> @paveenp I don't know what you mean by "unsupportably dependent". I am also
> not aware of "serious bugs". Could you be more specific?
>
> Peter.
>
>
> On Fri, Aug 2, 2013 at 10:40 AM, Delirium <delirium [at] hackish> wrote:
>
>> On 8/2/13 7:07 PM, praveenp wrote:
>>
>>>
>>> On Friday 02 August 2013 09:06 PM, Delirium wrote:
>>>
>>>> On 7/22/13 2:53 AM, Peter Krautzberger wrote:
>>>>
>>>>> 2) TeX/LaTeX compatibility might be lost.
>>>>>
>>>>> "Native" content (e.g. <maction> or even subexpression links) has no
>>>>> counterpart in TeX. Conservative extensions of TeX can easily enable
>>>>> this
>>>>> kind of content but backward compatibility will be lost.
>>>>>
>>>>>
>>>> If this means MathML as the canonical format, i.e. people write MathML
>>>> into articles directly, rather than it just being an output/rendering
>>>> format, that gives me moderate worry:
>>>>
>>>> 1. From the perspective of being able to repurpose Wikipedia articles
>>>> outside of a web context, TeX-format equations are nice for articles in the
>>>> math/science sphere, since TeX-based publishing workflows are common in
>>>> math/science, and equations are particularly tedious and error-prone to
>>>> convert by hand, if that would end up necessary. Admittedly, in some
>>>> workflows there's no real difference: you can import both MathML and TeX
>>>> equations into MS Word with appropriate plugins (I haven't looked into
>>>> whether the two import paths differ on compatibility). Perhaps as
>>>> HTML-based print workflows improve this will drop off as an issue, but
>>>> right now only a smallish proportion of people are using workflows based on
>>>> something like PrinceXML, and the free-software alternatives to PrinceXML
>>>> are further behind.
>>>>
>>>> 2. From a wikitext-readability perspective, TeX-format equations are the
>>>> de-facto standard way of ASCII-fying equations, e.g. in plaintext emails,
>>>> while MathML isn't written in a syntax any humans normally write. So using
>>>> TeX as our underlying representation makes equations possible to edit in
>>>> text form, at least for people who already professionally work in areas
>>>> where that's common, while MathML equations virtually require a visual
>>>> editor (unless the idea is to use something like ASCIIMathML?).
>>>>
>>> What??!!?? sorry I didn't get a thing from this. :-)
>>>
>>>
>>> Current scenario is: In our current Math extension, textvc is simply
>>> unable to generate equations in png except Latin languages. Also Mathjax is
>>> heavily client dependent (Unsupportably dependent) and has its own serious
>>> bugs.
>>>
>>
>> I read Peter's point 2 as discussing the possible "native" use of MathML
>> tags, i.e. permitting people to write MathML into articles, rather than
>> only using MathML as an alternate rendering path for texvc/MathJax/etc. If
>> MathML is a render-only target, then "TeX/LaTeX compatibility might be
>> lost" doesn't seem like it could be an issue. So unless I'm totally
>> misreading, I took the discussion to be about allowing MathML in articles,
>> which could break TeX compatibility since not all MathML tags can be
>> rendered back into TeX equivalents. The two points above are my two
>> concerns w.r.t. that suggestion. Am I misreading the suggestion entirely?
>>
>> -Mark
>>
>>
>>
>> ______________________________**_________________
>> Wikitech-l mailing list
>> Wikitech-l [at] lists
>> https://lists.wikimedia.org/**mailman/listinfo/wikitech-l<https://lists.wikimedia.org/mailman/listinfo/wikitech-l>
>>
> _______________________________________________
> Wikitech-l mailing list
> Wikitech-l [at] lists
> https://lists.wikimedia.org/mailman/listinfo/wikitech-l

_______________________________________________
Wikitech-l mailing list
Wikitech-l [at] lists
https://lists.wikimedia.org/mailman/listinfo/wikitech-l

peter.krautzberger at mathjax

Aug 6, 2013, 11:56 AM

Post #33 of 33 (25 views)
Permalink

Re: Long term strategy for math on wikipedia [In reply to]

Thanks, praveenp.

Could you clarify if the problems you've seen are MediaWiki, texvc or
MathJax specific? I could only find
48032<https://bugzilla.wikimedia.org/show_bug.cgi?id=48032> (MathJax
should be fixed in the next release), and
48118<https://bugzilla.wikimedia.org/show_bug.cgi?id=48118>, from
which I understand, RTL is not supported by texvc. MathJax currently does
not support RTL but we plan to add it -- and, as I wrote, I'd be very
interested to hear if texvc is still being developed.

MathJax does not deal with ligatures directly since ligatures are really
text-mode, not math mode. So ligatures in text-blocks are passed through by
MathJax and should not be broken. Again, I don't know what texvc does.

Anyway, more bug reports would be great so that issues can be investigated.
I can't really comment if those are serious from a WMF pov.
Peter.

On Tue, Aug 6, 2013 at 10:53 AM, praveenp <me.praveen [at] gmail> wrote:

> I've problems with browsers like IE (Mainly XP) and opera (ubuntu
> 12.04/Mint Maya), although I forgot exact version numbers. And also it
> takes each code points independently so it converts rtl language to ltr
> language, or breaks any ligatures etc. (Aren't they serious bugs?)
>
>
> On Saturday 03 August 2013 12:34:56 AM IST, Peter Krautzberger wrote:
>
>> @Mark Just to clarify. Personally, I don't think wikitext's math format
>> should move away from a TeX-like input language. The point I was trying
>> making was that a conservative extension would be useful if MathML becomes
>> a desired output. It seems to me that texvc was specifically designed to
>> prevent fully fledged TeX input, so I wonder if it wouldn't help everyone
>> if wasn't required on the backend anymore, only that the syntax stayed
>> backward compatible.
>>
>> @paveenp I don't know what you mean by "unsupportably dependent". I am
>> also
>> not aware of "serious bugs". Could you be more specific?
>>
>> Peter.
>>
>>
>> On Fri, Aug 2, 2013 at 10:40 AM, Delirium <delirium [at] hackish> wrote:
>>
>> On 8/2/13 7:07 PM, praveenp wrote:
>>>
>>>
>>>> On Friday 02 August 2013 09:06 PM, Delirium wrote:
>>>>
>>>> On 7/22/13 2:53 AM, Peter Krautzberger wrote:
>>>>>
>>>>> 2) TeX/LaTeX compatibility might be lost.
>>>>>>
>>>>>> "Native" content (e.g. <maction> or even subexpression links) has no
>>>>>> counterpart in TeX. Conservative extensions of TeX can easily enable
>>>>>> this
>>>>>> kind of content but backward compatibility will be lost.
>>>>>>
>>>>>>
>>>>>> If this means MathML as the canonical format, i.e. people write
>>>>> MathML
>>>>> into articles directly, rather than it just being an output/rendering
>>>>> format, that gives me moderate worry:
>>>>>
>>>>> 1. From the perspective of being able to repurpose Wikipedia articles
>>>>> outside of a web context, TeX-format equations are nice for articles
>>>>> in the
>>>>> math/science sphere, since TeX-based publishing workflows are common in
>>>>> math/science, and equations are particularly tedious and error-prone to
>>>>> convert by hand, if that would end up necessary. Admittedly, in some
>>>>> workflows there's no real difference: you can import both MathML and
>>>>> TeX
>>>>> equations into MS Word with appropriate plugins (I haven't looked into
>>>>> whether the two import paths differ on compatibility). Perhaps as
>>>>> HTML-based print workflows improve this will drop off as an issue, but
>>>>> right now only a smallish proportion of people are using workflows
>>>>> based on
>>>>> something like PrinceXML, and the free-software alternatives to
>>>>> PrinceXML
>>>>> are further behind.
>>>>>
>>>>> 2. From a wikitext-readability perspective, TeX-format equations are
>>>>> the
>>>>> de-facto standard way of ASCII-fying equations, e.g. in plaintext
>>>>> emails,
>>>>> while MathML isn't written in a syntax any humans normally write. So
>>>>> using
>>>>> TeX as our underlying representation makes equations possible to edit
>>>>> in
>>>>> text form, at least for people who already professionally work in areas
>>>>> where that's common, while MathML equations virtually require a visual
>>>>> editor (unless the idea is to use something like ASCIIMathML?).
>>>>>
>>>>> What??!!?? sorry I didn't get a thing from this. :-)
>>>>
>>>>
>>>> Current scenario is: In our current Math extension, textvc is simply
>>>> unable to generate equations in png except Latin languages. Also
>>>> Mathjax is
>>>> heavily client dependent (Unsupportably dependent) and has its own
>>>> serious
>>>> bugs.
>>>>
>>>>
>>> I read Peter's point 2 as discussing the possible "native" use of MathML
>>> tags, i.e. permitting people to write MathML into articles, rather than
>>> only using MathML as an alternate rendering path for texvc/MathJax/etc.
>>> If
>>> MathML is a render-only target, then "TeX/LaTeX compatibility might be
>>> lost" doesn't seem like it could be an issue. So unless I'm totally
>>> misreading, I took the discussion to be about allowing MathML in
>>> articles,
>>> which could break TeX compatibility since not all MathML tags can be
>>> rendered back into TeX equivalents. The two points above are my two
>>> concerns w.r.t. that suggestion. Am I misreading the suggestion entirely?
>>>
>>> -Mark
>>>
>>>
>>>
>>> ______________________________****_________________
>>> Wikitech-l mailing list
>>> Wikitech-l [at] lists
>>> https://lists.wikimedia.org/****mailman/listinfo/wikitech-l<https://lists.wikimedia.org/**mailman/listinfo/wikitech-l>
>>> <ht**tps://lists.wikimedia.org/**mailman/listinfo/wikitech-l<https://lists.wikimedia.org/mailman/listinfo/wikitech-l>
>>> >
>>>
>>> ______________________________**_________________
>> Wikitech-l mailing list
>> Wikitech-l [at] lists
>> https://lists.wikimedia.org/**mailman/listinfo/wikitech-l<https://lists.wikimedia.org/mailman/listinfo/wikitech-l>
>>
>
> ______________________________**_________________
> Wikitech-l mailing list
> Wikitech-l [at] lists
> https://lists.wikimedia.org/**mailman/listinfo/wikitech-l<https://lists.wikimedia.org/mailman/listinfo/wikitech-l>
>
_______________________________________________
Wikitech-l mailing list
Wikitech-l [at] lists
https://lists.wikimedia.org/mailman/listinfo/wikitech-l

1 2

View All

Index | Next | Previous | View Threaded

Interested in having your list archived? Contact Gossamer Threads