LaTeX Why are MathJax images not displaying on my webpage?

AI Thread Summary
MathJax has been enabled on the site, but users are experiencing issues with LaTeX images not displaying correctly, leading to frustration over the appearance and functionality. Some users report that certain LaTeX environments, like align and tabular, are not rendering properly, and there are inconsistencies in how equations appear across different browsers. The transition from image-based LaTeX to MathJax is intended to reduce server load, but users are concerned about the visual quality and usability of the new format. Feedback indicates that while MathJax can handle more equations without server strain, it may introduce rendering issues that affect the user experience. The discussion reflects a need for further adjustments and testing to optimize MathJax's implementation.
  • #251
Fredrik said:
££math££, ##math##, §§math§§ and ··math·· for inline (i.e. itex)
$$math$$ and ¢¢math¢¢ for displaystyle (i.e. tex)

almost every keyboard has $, and either £ € or ¢ next to it,

so how about $$…$$ for tex, and alternatives of ¢¢…¢¢ or ££…££ or €€…€€ for itex ? :smile:

though I'd prefer it the other way round, with the default ($$) choice being for itex, and people having to think a bit harder to choose tex … many posts are getting very tall and broken up by repeatedly starting a new line for just one symbol :redface:
 
Physics news on Phys.org
  • #252
tiny-tim said:
almost every keyboard has $, and either £ € or ¢ next to it,

so how about $$…$$ for tex, and alternatives of ¢¢…¢¢ or ££…££ or €€…€€ for itex ? :smile:
That's certainly acceptable to me. The reasoning behind my suggestion in #250 is that it allows almost everyone to use either Shift+3 for itex and Shift+4 for tex, or Alt Gr+3 for itex and Alt Gr+4 for tex. the Swiss and a few others would have to do it in a different way, but it looks like it would be just as easy for them.

tiny-tim said:
I'd prefer it the other way round, with the default ($$) choice being for itex
The main reason why that idea bothers me a little is that $$math$$ has the same effect as \begin{equation*}math\end{equation*} in a LaTeX document, i.e. it has the same effect as tex tags here.

Uhh...I don't know why MathJax turns that into an image. Click quote to see what I typed.
 
Last edited:
  • #253
tiny-tim said:
almost every keyboard has $, and either £ € or ¢ next to it...
In U.S., we don't have £ € or ¢. Used to have ¢ on typewriters long ago, but it is not present on computer keyboards here.

Greg Bernhardt said:
What about a double pound sign? Is that ever used in anything?
Sounds like that would work. The # character appears to be common to all keyboards using latin alphabets (N. & S. America and Europe) I have just found there's a wiki page showing standard keyboard layouts for different countries:

http://en.wikipedia.org/wiki/Keyboard_layout#United_Kingdom"
http://en.wikipedia.org/wiki/Keyboard_layout#United_States"
etc.
 
Last edited by a moderator:
  • #254
You guys call # a pound sign? Then what do you call £?
 
  • #255
Fredrik said:
The main reason why that idea bothers me a little is that $$math$$ has the same effect as \begin{equation*}math\end{equation*} in a LaTeX document, i.e. it has the same effect as tex tags here.

but would it really matter if newcomers thought they were typing tex but got itex instead ?
Redbelly98 said:
In U.S., we don't have £ € or ¢. Used to have ¢ on typewriters long ago, but it is not present on computer keyboards here.

oops! i assumed ¢ was essential in the US! :redface:

ok, in that case ¢ is pointless, so instead …

how about $$…$$ for tex, and alternatives of ##…## or ££…££ or €€…€€ for itex ? :smile:

(or t'other way round)
 
  • #256
tiny-tim said:
but would it really matter if newcomers thought they were typing tex but got itex instead ?
It wouldn't matter much to anyone. It shouldn't matter at all to newcomers. I would find it a little bit annoying to use $$...$$ in a way that's different from what that code means in actual LaTeX, but I could get used to it.

tiny-tim said:
ok, in that case ¢ is pointless, so instead …
The (admittedly very small) point is to let French Canadians use Alt Gr+3 and Alt Gr+4 instead of Alt Gr+3 and Shift+4. So I don't think they will be too disappointed if we drop that symbol. Similar things can be said about that dot symbol and Spaniards, and the paragraph symbol and Germans.

tiny-tim said:
how about $$…$$ for tex, and alternatives of ##…## or ££…££ or €€…€€ for itex ? :smile:

(or t'other way round)

I don't see a use for the Euro symbol. Almost everyone has # and/or £ on the 3 key, and $ on the 4 key. Most (all?) of the ones who don't can still type # and £ as easily as €.
 
Last edited:
  • #257
Question is - if something else that tex tags is used, what if one day Greg will decide/need to change MathJax to something else? It can potentially make future migration much more difficult (now adding MathJax doesn't require any changes to existing posts).

I guess what I actually mean is "if it ain't broken, don't fix it".
 
  • #258
Borek said:
I guess what I actually mean is "if it ain't broken, don't fix it".

Hi Borek! :smile:

If you use the "bad" Enlish "ain't", then you should also use the "bad" English "broke" (not "broken")! :wink:

(although wikipedia claims that this expression dates from 1977, a google book search show plenty of prior usage, including The American school board journal, Volume 166, 1891, at http://books.google.co.uk/books?id=H_YqAAAAMAAJ&q=%22If+it+ain't+broke,+don't+fix+it%22…&sa=X&oi=book_result&ct=result&resnum=3&ved=0CDwQ6AEwAg)
 
  • #259
I see. I am afraid this is one of these things that may not stick :frown:
 
  • #260
Borek said:
Question is - if something else that tex tags is used, what if one day Greg will decide/need to change MathJax to something else? It can potentially make future migration much more difficult (now adding MathJax doesn't require any changes to existing posts).
It would almost certainly be possible to specify what to use as delimiters in that software too, and I don't think we would be completely screwed even if it isn't. We could e.g. run a script that replaces the delimiters that will cause problems with the new software. This would be a bit of a pain, but this pain should be multiplied with the probability that we (actually Greg) will experience it.
 
  • #261
Fredrik said:
Uhh...I don't know why MathJax turns that into an image. Click quote to see what I typed.
Looks like MathJax processes stuff between \begin and \end even though it's not between delimiters. It worked with equation and array at least.
 
  • #262
yenchin said:
I have the same problem with Chrome with Vista [on two computers]. Everything just hanged for a minute or so and I couldn't do anything.

This still irritates me :cry:
 
  • #263
Am I the only person for who the math looks rough around the edges?
I just found, after some searching, a method of producing antialiased typography: it can be done through generated Flash images (these are temporary, and only generate on top of the current equations if Flash is available). This is in my opinion an elegant solution for creating PDF-quality Latex.

Pleae see the link I found: http://www.mikeindustries.com/blog/sifr
 
  • #264
Oh no, that site has a picture from when I met Stan Lee:

[PLAIN]http://28.media.tumblr.com/tumblr_ljfhd0uyTH1qbocrho1_500.jpg
 
Last edited by a moderator:
  • #265
Here is a working example of a serious implementation of Latex in Flash
http://www.fmath.info/latexImplemented.jsp

I would like to know your opinion about possibly implementing such a rendering method on physicsforums, since flash can produce scalable vector graphics, producing fast and high-quality renderings of Latex.
 
  • #267
Sybren said:
Am I the only person for who the math looks rough around the edges?

It looks just as smooth as normal text to me. This is on Chrome 13, Windows Vista.
 
  • #268
Sybren said:
Am I the only person for who the math looks rough around the edges?

I am afraid so.
 
  • #269
Borek said:
I am afraid so.

Problem solved!
I had to turn on my Windows Cleartype.. Maybe helpful information for other people around here, because the rendering is now done by the browser, so quality can variate.

Attached is a comparison with cleartype on/off, it really makes a difference:
 

Attachments

  • cleartype.jpg
    cleartype.jpg
    7 KB · Views: 376
  • #270
Borek said:
I see. I am afraid this is one of these things that may not stick :frown:
Remember the rule by remembering the horrible pun:
If it ain't baroque, don't fix it!​
 
  • #271
Fredrik said:
Oh no, that site has a picture from when I met Stan Lee:

[PLAIN]http://28.media.tumblr.com/tumblr_ljfhd0uyTH1qbocrho1_500.jpg[/QUOTE]

:smile: :smile: :smile: :smile:

Can't... stop... laughing... help!

:smile: :smile: :smile: :smile:
 
Last edited by a moderator:
  • #272
Sybren said:
Problem solved!
I had to turn on my Windows Cleartype.. Maybe helpful information for other people around here, because the rendering is now done by the browser, so quality can variate.

Attached is a comparison with cleartype on/off, it really makes a difference:
Looks like the problem was there with normal text as well, but the thinner strokes in the math characters made it more pronounced.
 
  • #273
FYI to IE users! Just received this response from a developer:

There are a number of security settings that affect the operation of
MathJax. I did some testing in IE7 (for a different report of the
same issue), and found that there are at least four important settings
that can cause MathJax to not process the page. They are accessible
in the Internet Options menu item (in the Tools Menu) under the
Security tab. If the security setting is set to "High" it sets
disables all four. You can use the Custom Level button to change the
following:

First, you need to make sure "Active Scripting" under the Scripting
section is enabled (near the bottom of the list). This allows
Javascript to run (or not). If Javascript is disabled, MathJax will
not be loaded or run, so this one is critical. Since you are seeing a
javascript error message, I assume this one is enabled for your
limited account.

Second, you need to make sure "Run ActiveX controls and Plugins" is
enabled (or prompted) in the "ActiveX Controls and Plugins" section
(second to last option). This is the setting that is causing the
error that you are reporting. The line you indicate is one where
MathJax is trying to set up MathPlayer (which is an ActiveX control),
and if this setting is disabled, IE throws an error.

Third, you need to make sure "Script ActiveX controls marked safe for
scripting" is enabled (or prompted) again in the "ActiveX Controls and
Plugins" section (last option). It requires a restart of IE if you
change this setting. This is needed in order to allow MathJax to
create an XML parser, which is used in the MathML input jax.

Fourth, you need to make sure "Font Download" is enabled (or prompted)
in the "Downloads" section. This is required for MathJax to use the
web-based fonts. Without this, IE will time out waiting for the first
font, and then fail over to image fonts, so it is not strictly
required, but makes for a better experience.

Currently, whenever the NativeMML output jax is configured, it tries
to register MathPlayer in IE, and whenever the MathML input jax main
code is loaded, it instantiates the XML parser. Since you are using
the TeX-AMS-MML_HTMLorMML configuration, you get the NativeMML output
jax as part of that regardless of the renderer for the page. That is
why you are seeing this error. You can avoid the error for now by
switching to the TeX-AMS_HTML configuration file instead, but this
means you users would have to switch to the MathML renderer manually
in order to use their native MathML support (if any).
I will look into changing the timing of the setup for MathPlayer so
that it will be attempted only when the renderer is explicitly set to
MathML, which should avoid the problem in most cases. It is probably
possible to trap the errors for the second and third settings as
well. But as long as those settings are in place, MathJax will not be
able to use MathPlayer to render mathematics, and it will not be able
to parse MathML input, so some MathJax functionality will be
unavailable.
 
  • #274
micromass said:
:smile: :smile: :smile: :smile:

Can't... stop... laughing... help!

:smile: :smile: :smile: :smile:
That's good to hear :smile: I almost thought I was the only one laughing when no one showed any appreciation for more than 4 hours. :biggrin:
 
  • #275
FYI the devs have told me that the "Loading ..." messages problem will be fixed in the next release
 
  • #276
for those having odd/serious mathjax problems run this script and then click the email results button
https://www.physicsforums.com/mathjax/mjp.html
 
Last edited:
  • #277
I have sent two emails two whomsoever, immediately before and after pressing the run button on the linked page.

I could see the quadratic formula on the mathjax page, no problem.

Since there has been no improvement to my viewing of formulae I wondered if there had been any progress on my discovery that the problem only arises if the user has limited privileges (as mine and nearly all corporate ones do)?

go well
 
  • #278
Studiot said:
I wondered if there had been any progress on my discovery that the problem only arises if the user has limited privileges (as mine and nearly all corporate ones do)?

go well

The devs are aware of this problem and say an update will be released in the next week that should resolve it.
 
  • #279
The devs are aware of this problem and say an update will be released in the next week that should resolve it.

Looking forward to it.

All the 4 items in your post #273 have been enabled for a while, to no avail.
 
  • #280
Sorry to be a pest but all I ended up with on emailing (clicking the button on the webpage) the mathjax test was two emails in my outlook express box, without recipient.

Now my OE does not work.

I can still transfer these to webmail, but do not know the recipient, please?
 
  • #281
Edit

I think OE is working it just gives an error message because there is not recipient.
 
  • #283
Greg Bernhardt said:
for those having odd/serious mathjax problems run this script and then click the email results button
https://www.physicsforums.com/mathjax/mjp.html

ok you can submit now
 
Last edited:
  • #284
Yes boss done.
 
  • #285
Sometimes there is not much to report...

And when I have problems, it always looks identical - 15 seconds waiting for downloads, I guess that means timeout.
 

Attachments

  • Untitled-1.png
    Untitled-1.png
    12.3 KB · Views: 413
  • #286
No, looks like it can get different - now LaTeX was processed, but Opera still waits for servers.

I have three reasons to use Opera - first, I have zillions of usernames and passwords stored in Opera configuration files. Second, I paid to register Opera in 2004, when it was ad-sponsored. Third - nobody uses Opera, so it is rarely (if ever) targeted by hackers. Unfortunately lately using Opera to access PF is becoming a PITA (it works OK on other sites I visit, but I spend most of the time at PF), so I may have to think it over again
 

Attachments

  • Untitled-2.png
    Untitled-2.png
    28 KB · Views: 436
  • #287
Borek said:
No, looks like it can get different - now LaTeX was processed, but Opera still waits for servers.

Are the results in the image relatively constant?
 
  • #288
Usual pattern looks more like this.

(And when I want to show this 15 sec timeout everything works perfectly, why should it not?)
 

Attachments

  • Untitled-3.jpg
    Untitled-3.jpg
    37.9 KB · Views: 382
  • #289
Borek said:
Usual pattern looks more like this.

(And when I want to show this 15 sec timeout everything works perfectly, why should it not?)

I gtg for now. Keep posting interesting results and I will compile them and send them to the devs later today. thanks!
 
  • #290
OK, that's the timeout. Note that on the first image it shows timeline, but Opera still waits for mathjax server.
 

Attachments

  • Untitled-1.jpg
    Untitled-1.jpg
    36.3 KB · Views: 386
  • Untitled-2.jpg
    Untitled-2.jpg
    34.3 KB · Views: 406
  • #291
Greg Bernhardt said:
The devs are aware of this problem and say an update will be released in the next week that should resolve it.
You can fix it now by using TeX-AMS_HTML instead of TeX-AMS-MML_HTMLorMML.
Fredrik said:
The main reason why that idea bothers me a little is that $$math$$ has the same effect as
math
in a LaTeX document, i.e. it has the same effect as tex tags here.

Uhh...I don't know why MathJax turns that into an image. Click quote to see what I typed.
You could also change this behavior Fredrik saw by setting processEnvironments to false in the tex2jax parameters. On the other hand, perhaps you'd prefer MathJax, like TeX, to process anything between \begin and \end as math without needing to type in delimiters.
 
Last edited:
  • #292
Can we make it so clicking the [PLAIN]https://www.physicsforums.com/products/latexreference/images/Icon3.gif icon inserts itex tags instead of tex? I'm finding a lot of new members using tex for LaTeX code that is supposed to be inline with their text, and it is breaking up expressions into multiple lines.

Also, I have updated the LaTeX stickied thread in Forum Feedback.
 
Last edited by a moderator:
  • #293
Redbelly98 said:
Can we make it so clicking the [PLAIN]https://www.physicsforums.com/products/latexreference/images/Icon3.gif icon inserts itex tags instead of tex? I'm finding a lot of new members using tex for LaTeX code that is supposed to be inline with their text, and it is breaking up expressions into multiple lines.
I never use that button, but I would guess that it's both possible and preferable to replace it with two buttons, one for itex and one for tex.
 
Last edited by a moderator:
  • #294
Fredrik said:
I never use that button, but I would guess that it's both possible and preferable to replace it with two buttons, one for itex and one for tex.

for the meantime I switched it to itex
 
  • #295
Thanks Greg!

I don't know if that is the root of the problem, but we'll see if we stop getting posts from new members like this:

***** said:

Homework Statement


Three sinusoidal waves of the same frequency travel along a string in the positive direction of an x axis. Their amplitudes are y1, y1/3.0, and y1/4.0, and their phase constants are 0, \pi/4.0, and \pi, respectively. What are (a) the ratio of the amplitude to y1 and (b) the phase constant of the resultant wave?

Even if you're aware of the issue, it's easy to miss that one of the phase constants is \pi/4.0
 
  • #297
Greg, is there anything I can do to help?

- Warren
 
  • #298
chroot said:
Greg, is there anything I can do to help?

- Warren

Thanks Warren! I think we're in pretty good shape.

I believe 1.1a is now out. Those who were having problems please report any differences. Remember to clear your browser cache as the JS is cached for two weeks.
 
  • #299
I tried visiting the forums using a limited account in XP, and the access-denied problem no longer occurs.

Any decision on new delimiters?
 
  • #300
Greetings from the land of XP limited privilege users:

Whatever you did 10 minutes prior to the timestamp of this post suddenly worked properly.

The yellow error triangle was no longer evident and I thought tex worked again.

However now the issue has returned.
 
Last edited:

Similar threads

Replies
3
Views
4K
Replies
15
Views
25K
Replies
4
Views
4K
Replies
8
Views
2K
Replies
3
Views
2K
Replies
4
Views
4K
Replies
3
Views
2K
Replies
3
Views
13K
Back
Top