Why are MathJax images not displaying on my webpage?

[Greg Bernhardt]

While correlation $\not =$ causation, I stopped having those problems after I installed the fonts. It's possible that there is some problem with the way FF4 treats the fontdata.js or the fonts when loaded from the server.

Let me try with FF4 in my W7 VM (which doesn't have the STIX fonts) and get back to you.

hmm I don't have STIX locally loaded and FF4 on W7 looks fine to me.

 

[jhae2.718]

Works fine for me in FF4.0.1 in Win 7 SP1 x86_64 under VirtualBox 4.0.8. No local STIX fonts, so it seems my guess was wrong.

 

[cepheid]

Full version.

Right, so I mean, that's exactly what I was reporting in the first place: that it doesn't work on the mobile version of the site.

 

[jhae2.718]

Right, so I mean, that's exactly what I was reporting in the first place: that it doesn't work on the mobile version of the site.

Sorry, didn't catch that.

 

[Greg Bernhardt]

Right, so I mean, that's exactly what I was reporting in the first place: that it doesn't work on the mobile version of the site.

I haven't added the code to the mobile skin.

 

[cepheid]

I haven't added the code to the mobile skin.

Got it, thanks.

 

[Fredrik]

I'm using Chrome 12 and all the math on that page appears to render fine. I'll test with FF 4.0.1

hmm using Win7 FF 4.0.1 fred's post looks fine except for the one equation he mentioned

Maybe I didn't make it clear enough that everything looks fine until I move my mouse pointer over a math expression.

 

[George Jones]

I just printed the thread

https://www.physicsforums.com/showthread.php?t=491078&highlight=speed,

and every "+", "-", "=", and integral sign (and possible other stuff) is missing

 

[Greg Bernhardt]

Maybe I didn't make it clear enough that everything looks fine until I move my mouse pointer over a math expression.

Yes I see what you are saying now. It seems the equation disappears, depending on your zoom trigger, on anything above Theorem 2 in your post. I will send this to the devs and see what they say. thanks!

 

[Greg Bernhardt]

I just printed the thread

https://www.physicsforums.com/showthread.php?t=491078&highlight=speed,

and every "+", "-", "=", and integral sign (and possible other stuff) is missing

printed to a physical piece of paper? do they display fine in the thread?

 

[Fredrik]

By the way, the math looks much better in IE than in FF. I'd say that in IE, it looks the way I want it to look, while in FF it's too small, and there are weird spaces inserted in simple expressions like $\|\phi\|=\|x_0\|$.

 

[George Jones]

printed to a physical piece of paper?

Yes.

do they display fine in the thread?

Yes.

I used printer at work. When I get home, I will try printing on my home printer.

 

[Greg Bernhardt]

By the way, the math looks much better in IE than in FF. I'd say that in IE, it looks the way I want it to look, while in FF it's too small, and there are weird spaces inserted in simple expressions like $\|\phi\|=\|x_0\|$.

This is a well known problem because of the differences in how IE and FF handle web fonts.

 

[micromass]

I just saw this point https://www.physicsforums.com/showthread.php?p=3309596#post3309596 and it's quite difficult to read because each equation starts on a new line.

 

[Greg Bernhardt]

I just saw this point https://www.physicsforums.com/showthread.php?p=3309596#post3309596 and it's quite difficult to read because each equation starts on a new line.

because the user didn't use itex

 

[micromass]

because the user didn't use itex

So you always need to use itex from now on? Before, just using tex didn't give you a new line...

 

[Greg Bernhardt]

So you always need to use itex from now on? Before, just using tex didn't give you a new line...

hmmm then what was the function of itex when using images?

 

[jhae2.718]

So you always need to use itex from now on? Before, just using tex didn't give you a new line...

I think in MathJax [itex] behaves more closely like $...$ and [tex] like \begin{equation}...\end{equation} than they did in imgtex.

 

[cepheid]

hmmm then what was the function of itex when using images?

Well itex created images with a symbol size that was closer to the plain text characters (for the default font size). It also didn't mess up the line spacing nearly as much as inline images generated using tex tags did. On the downside, I always found itex equations kind of small to read.

Edit: also, the inline tex images were never in the right vertical position wrt the rest of the line. They were too high. The itex ones were aligned properly.

 

[Fredrik]

Sometimes it was necessary to use tex instead of itex even for inline text, because itex cut off the top of the taller symbols, like $\vec A$. So I suspect that a lot of posts will look weird if tex tags start new lines now. I'm not sure how much effort I'd put into solving that problem though. Maybe we will just have to accept that old posts will look ugly.

 

[mishrashubham]

because the user didn't use itex

I wish I had posted only a day later.

BTW, how old do posts have to be in order to become un-editable by the poster?

 

[FtlIsAwesome]

We currently host 1,114,554 latex graphics.

I am having no difficulties seeing the equations, other than a brief loading time. I'm using Chrome on Windows 7.

What is the difference between tex and itex tags?

 

[Greg Bernhardt]

i'm going to keep it up all night because the devs haven't gotten back to me yet

 

[Lacy33]

Let me know when you get this fixed because I still can't do math.

 

[coelho]

For me the greatest problem with this new implementation of latex is it only works when the javascript is enabled.
I usually let it disabled as it makes the pages load way faster (or rather less slowly... dial-up speed here ), and i would prefer keep this way.

 

[Redbelly98]

I just printed the thread

https://www.physicsforums.com/showthread.php?t=491078&highlight=speed,

and every "+", "-", "=", and integral sign (and possible other stuff) is missing

printed to a physical piece of paper? do they display fine in the thread?

I notice that if you select equations in the post George linked to, the "+", "-", "=", and "/" symbols are not highlighted like the others. This would be a faster way to see the problem than printing out hardcopies, if you are trying to debug this issue.

 

[jhae2.718]

It looks like operators don't highlight in any post. It seems to work fine on the MathJax demo page, but not on PF.

Edit: it works in FF4 if you switch the renderer to HTML-CSS. (Right click on an equation, go to Settings > Math Renderer > HTML-CSS).

Math also looks better IMO.

 

[Greg Bernhardt]

It looks like operators don't highlight in any post. It seems to work fine on the MathJax demo page, but not on PF.

Edit: it works in FF4 if you switch the renderer to HTML-CSS. (Right click on an equation, go to Settings > Math Renderer > HTML-CSS).

Math also looks better IMO.

hmm it should have always been HTML-CSS. Was your default set to MathML? FF has known MathML problems still.

On second thought it does appear at least FF is defaulting to NativeMML. I manually changed FF to use HTML-CSS and the equations look greatly improved.

 

[jhae2.718]

On PF at least Firefox defaulted to MathML rendering. Not sure if you can force HTML-CSS rendering from the server side.

Also, when using HTML rendering it seems to me that the 120% scaling is overkill. (FF seemed to render MathML unusually small, but for me at least HTML/CSS rendering is fine at 100%.)

 

[Fredrik]

BTW, how old do posts have to be in order to become un-editable by the poster?

I think the current limit is 11 hours and 40 minutes (700 minutes).

 

[mishrashubham]

I think the current limit is 11 hours and 40 minutes (700 minutes).

thanks

 

[Greg Bernhardt]

If there aren't any serious problems I think I'll continue to leave the MathJax up for the time being.

 

[Fredrik]

If there aren't any serious problems I think I'll continue to leave the MathJax up for the time being.

It looks like the worst problems have been solved, so I don't see any reason to go back to the old system. Thanks for all the work you've done on this.

The most serious problem right now is that we sometimes have to insert spaces in correct LaTeX code. For example, I tried this earlier today:

|\psi\rangle\langle\psi|=\frac{1}{2}\big(|\text{here}\rangle\langle\text{here}|+|\text{here}\rangle\langle\text{there}|+|\text{there}\rangle\langle\text{here}|+|\text{there}\rangle\langle\text{there}|\big)

The result:
$$|\psi\rangle\langle\psi|=\frac{1}{2}\big(|\text{here}\rangle\langle\text{here}|+|\text{here}\rangle\langle\text{there}|+|\text{there}\rangle\langle\text{here}|+|\text{there}\rangle\langle\text{there}|\big)$$

The first "here" is displayed as "he re". The second time \rangle is followed by \langle, the parser seems to miss the backslash. And then the \text before the last "here" seems to be read as \te. It can all be fixed by inserting spaces before the keywords that cause problems:

$$|\psi\rangle\langle\psi|=\frac{1}{2}\big(| \text{here}\rangle\langle\text{here}|+|\text{here}\rangle \langle\text{there}|+|\text{there}\rangle\langle \text{here}|+|\text{there}\rangle\langle\text{there}|\big)$$

But then the last "there" is displayed as "ther e", so I have to insert another space before the last \text.

$$|\psi\rangle\langle\psi|=\frac{1}{2}\big(| \text{here}\rangle\langle\text{here}|+|\text{here}\rangle \langle\text{there}|+|\text{there}\rangle\langle \text{here}|+|\text{there}\rangle\langle \text{there}|\big)$$

Isn't there something you can do to force Firefox to display math that looks as pretty as in Internet Explorer? Is there something we can do, other than switch to IE?

 

[Greg Bernhardt]

Isn't there something you can do to force Firefox to display math as well as Internet Explorer? Is there something I can do, other than switch to IE?

I recently forced FF to use HTML instead of MathML. It looks much better to me in FF now. Maybe try clearing your cache.

 

[Borek]

The first "here" is displayed as "he re".

As far as I can tell this is vBulletin problem (feature). It splits long text inserting spaces after every 50 characters:

01234567890123456789012345678901234567890123456789012345678901234567890123456789012345678901234567890123456789012345678901234567890123456789012345678901234567890123456789012345678901234567890123456789012345678901234567890123456789012345678901234567890123456789012345678901234567890123456789012345678901234567890123456789

If you try to quote my post, there are no spaces between digits, they are added to the displayed text only. Unfortunately, MathJax works not with the original post text, but with whatever vBulletin displays.

This way forum formatting doesn't break, but sometimes it means other problems (like in this case).