Want to take part in these discussions? Sign in if you have an account, or apply for one below
Vanilla 1.1.10 is a product of Lussumo. More Information: Documentation, Community Support.
Yes, "\mathcal{A}" does the trick.
Here is a list of all math commands that the software here understands: itex Commands.
(This is linked to from the HowTo page.)
Thank you both. It is working very well, but only for certain letters. The commands \mathcal{A} and \mathcal{C} are producing a strange Maya hieroglyph.Why is that? I have observed the same hieroglyph a few times in the nLab page. Is this a problem with my computer?
That's a somewhat annoying issue with fonts.
The trouble is that the MathML software operating in the background is still a bit ahead of its time, in a way. Browsers are supposed to be able to display it all correctly, but in practice it turns out that some versions of some browserers lack certain fonts to display some characters.
If you use Firefox, you can install the Stix fonts by hand, to remedy this.
Or possibly just updating to a new version of Firefox might do, too. I never know exactly how these things work. On the Firefox that I am using the \mathcal{A} does display correctly, and I never installed any Fonts by hand.
I am presently developing the theory of factorisation systems in the CatLab:
http://ncatlab.org/joyalscatlab/show/Factorisation+systems
Can I create a link to it in the nLab page on orthogonal systems?
http://ncatlab.org/nlab/show/orthogonal+factorization+system
Sure, and I see that you did. I fixed the links.
By the way, you can create links here in the same way as on the nLab.
[[joyalscatlab:Factorisation systems]] → ‘Factorisation systems (joyalscatlab)’[[orthogonal factorization system]] → ‘orthogonal factorization system’Nice page!
Makes me think that we should also use the CodeCogs-diagrams more on the nLab. That does certainly yield a better look of the page.
As far as stix fonts I installed them once (it is easy and quick!), and it seemed that all font problems disappeared on firefox of my laptop. But then, at some point i noticed that some of the mathcal letters do not work while some do, the problem which I had before on another computer. Is this possible at all, or I messed the system with some other action in the meantime ?
To Andre: I saw earlier in your Barcelona notes these iterated localizations, and I like it, but in familiar setups like with abelian categories one has just localizations in similar factorization theorems which I know. I intuitively saw the kind of phenomena which would in general lead to examples of iterated localizations which are not localizations at the time, but did not fully do it. Do you have some handy convincing example of an iterated localization which is not just a single localization ?
Urs: CodeCogs is nice but the source quoting of images at internet site makes the execution of sourcecode depending on internet. So we need to save html and not source in our own copies of nlab for viewing. Now the sourcecode of codecogs is also opensource, just like itex2MML (am i right), so maybe we can have it running as well as on our own computers. These are just potentially relevant remarks, never mind.
@Toby: thank you for the link.
@Urs: I am glad you like the diagrams. CodeCogs is working well.
@Zoran: I had an example, but I forgot it. I will have to reconstruct it.
@Harry: I suppose you are thinking about the example 4. The connectedness condition is not on the pullback square but on the category b\A defined by the pullback square. The example is rather obscure if you are not familiar with the notion of final functor (the example is due to Street and I should add a reference). I will eventually link the example with the theory of final functors which I intend to write in the CatLab.
Ah, so direction of arrows doesn't matter?
Yes, you may imagine that we replace the category by its "groupoidificaton", its Kan fibrant replacement and then look at the connected components of that.
I am presently developing the theory of weak factorisation systems
http://ncatlab.org/joyalscatlab/show/Weak+factorisation+systems
I sometime need to refer to a proposition situated in a different page of the CatLab. For example, in the proof of Corollary 7, I would like to create a link to Theorem 11 and Proposition 16 of the Factorisation systems page:
http://ncatlab.org/joyalscatlab/show/Factorisation+systems
I wonder if it can be done and how?
I seem to remember that we had some discussion about this once, but I forget what the answer was.
What is supported and does work is automatic numbering and referencing of propositions and theorems on one single page. I am afraid across different pages there is not software support for this. But others here might know more.
@Urs:I hope it can be done. Otherwise I may have to abandon my project of writing proofs in the CatLab.
@Harry: I am not sure I know what you mean by doing it manually. Could you give me an example?
I am a computer illiterate, I can only learn by examples. I wonder if someone can give me an example.
Oh, now I understand: the question is not so much about how to get the numbering and links be generated automatically as about how to link to a specific point in a page?
Yes, this is easily done with anchors, as Harry indicates. I'll prepare a demonstration example in the Sandbox, which you can copy then.
okay, if you look at Sandbox now, you'll see a test example.
At the beginning there are two links called "jump to the first anchor point" and "jump to the second anchor point".
Click on these links to see what happens. Then look at the source code. It's probably self-explanatory.
In this example the links go within one page, but if you give the full URLs as you can see in the source code at the Sandbox, then this also works across different pages.
So typing this is a bit more inconvenient than just producing ordinary links. But at least it works.
Thank you both! The anchor works perfectly. I can continue to work. Very nice!
I've expanded the example a little since if you want to refer to definitions and theorems then the syntax is slightly simpler (and more LaTeX-like).
Thanks, Andrew. I should have mentioned this, probably. But this doesn't work across two different pages, does it?
What doesn't work is mixing Wikilinks with anchors, so HowTo#svgedit (with double square brackets) just links to the HowTo, not specifically to the section on the new SVG editor. However, using Markdown links does work, so putting HowTo#svgedit as the URL will work (sample in the Sandbox). That is:
[this works](HowTo#svgedit)
i see. Thanks, Andrew. I did's know that this was possible. That's of course much simpler than what I suggested.
You should use the \ref syntax whenever you can (i.e. with definitions and theorems within the same page) because they work properly when the page is exported to LaTeX. Links to other pages should be hyperlinks anyway in the exported document.
Update: Jacques says that [this works](/nlab/show/HowTo#svgedit) is better than [this works](HowTo#svgedit) since it's more robust for the history - when viewing an old version of a page then the latter will not work correctly but the former will.
@Andrew's last comment: that's weird, is that a bug or a feature? I don't see a reason why those should be treated differently.
I added some info about this to the HowTo.
Feature, but one of HTML not Instiki. If you put a link: <a href="page.html">Look at my page!</a> in an (X)HTML document then what the browser does is to strip the last part off the current URL and substitute in "page.html". So if your current page is "http://ncatlab.org/nlab/show/HomePage.html" then the browser goes and fetches "http://ncatlab.org/nlab/show/page.html". If, however, you are at "http://ncatlab.org/nlab/revision/HomePage/37" then it goes off and tries to fetch "http://ncatlab.org/nlab/revision/HomePage/page.html".
I should have remembered this and not made that mistake. It's also an issue when one page can be referred to in two different ways; for example, "http://www.math.ntnu.no/~stacey" and "http://www.math.ntnu.no/~stacey/index.html" refer to the same page, but a relative link <a href="mypage.html"> would do different things on each version. So relative links should be handled carefully to avoid this sort of thing.
(NB Some browsers may be aware of the issue and check first to see if they are in a directory or at a file, I don't know. Regardless, they won't know the subtleties of the different ways that an nLab page can be rendered.)
Anyway, as this all happens at the browser end, the only thing that we can do is Avoid The Issue by using absolute paths. Leaving off the hostname is okay, though.
PS Thanks for updating the HowTo! We should all be careful to do things like that when we discover little nuggets of information as to how things work around here.
To Urs, Andrew and Mike: I only have a vague understanding of the technical issues involved. Are you saying that one may link-refer to a dynamically numbered proposition in a different page, as in Latex? Would you please illustrated the method with a concrete example in the Sandbox. That's the only way I can learn. Thanks.
Harry Gindi wrote:
Quick question. What's your policy on typos? If we see typos, would you like us to fix them?
You are welcome to correct a typos in the CatLab when you find one, but you should not loose your time chasing them around. Everything I wrote is work in progress and there is a pretty good chance that I will introduce new typos in a near future.
Yes, one may link-refer to a dynamically numbered proposition, but the number will not be shown on the referring page; all you can do is make a link. I showed it at the Sandbox and also mentioned it at the HowTo.
@Andrew: Ah, I see. I didn't realize that instiki was just passing whatever is in the (...) of a [link](...) verbatim to the html anchor tag.
@Joyal: Please let us know which examples are good to learn from and which aren't - we'll transfer the good ones to the HowTo as the Sandbox can get wiped clean from time to time. And if there's something you can't figure out, drop a line here and one of the lab elves can have a look at the page in question and sort it out for you.
@Mike: well, it doesn't quite pass it blindfold - it checks for nasty stuff, but the point of the [text](link) syntax is to form links to other pages anywhere in the web. Instiki doesn't check to see if they happen to be actually on the nLab.
The informations in the Sandbox are good enough for me, except those about the WYSIWYG SVG editor, but it does not matter at the moment. I keep worrying about link-references. The anchor technique is unprecise. For example, the link here in the Proof 2 of Proposition 3 of
http://ncatlab.org/joyalscatlab/show/Weak+factorisation+systems
is taking me at the top of the page and not to Lemma 4 as intended. This can be very disorienting. Am I doing something wrong?
The anchors need to be attached to something, they can't hang in midair. So either add a (sub)section header where you want an anchor, or link directly to the lemma. That particular lemma has reference 'cancelquasicartesian' so link to that instead:
[here](http://ncatlab.org/joyalscatlab/show/Epicartesian+squares#cancelquasicartesian)
(You're editing the page at the moment, otherwise I would have corrected it for you.)
@Stacey: I understand my error. I corrected it and it now works. Thanks.
Another technical problem: I sometime would like to refer to a diagram in a proof but I do not know how to label the diagram (see for example the end of proof 1 of Proposition 3 in
http://ncatlab.org/joyalscatlab/show/Weak+factorisation+systems)
where are refer to an unlabelled square. Can you tell me how to label diagrams (and equations)?
Equations can be given anchors by using \[ and \] instead of double dollars, and then put \label{foo} inside the delimiters to create a label. Refer back to this either via (eq:foo) or \eqref{foo}.
(If you want to refer to an equation on a different page then it'll be something like the [text](link) syntax but I'll need to do a modicum of investigating to figure out exactly what to put where and I've not time to do that right now.)
If you put your diagram inside an equation, then you can refer to it like this. If your diagram is by itself, then there's a few different ways depending on exactly how the diagram is placed. Again, to be absolutely sure of the syntax I'll need to do a couple of experiments to see what is correct. If it's alright by you, I'll try to fix the references that you mention above and that'll tell me what the correct general syntax is - but it'll be tomorrow at the earliest. Rest assured that it can be done, and fairly easily, but I want to be sure that I've got it right.
Dear Andrew,
I have followed your instruction without success. When I compile
\eqref{abc} or (eq:abc)
the system reacts by returning a blank page with an error message. I must be doing something wrong. I will be grateful if you could fix the problem for me. --André
I'll take a look tomorrow.
I've had a look and I have a solution for you. It perhaps is not the most elegant method, but it does work.
When using square brackets for displayed equations then the opening bracket has to appear at the start of a line (you can have a little whitespace in front, but not too much). That's the first thing to note. The second thing is that if the mathematics is empty, then the label is still produced but gets attached to the next line (in particular, you don't get a blank line with a label at the tend). This is good because it means that we can attach the label to the next line, which consists of the diagram - I tried embedding the diagram in the equation environment but it didn't like that.
So what I propose is that you do:
\[\label{foo}\]
<center>
<img src="whatever your picture is">
</center>
Then the equation number appears next to the diagram. Admittedly, it is aligned with the top of the diagram rather than the usual LaTeX standard of the middle. However, that is actually beneficial since it means that someone clicking on the link to jump to that diagram can see the whole diagram straight away without having to scroll.
Refer back to it via \eqref{foo} or (eq:foo) as I said above. I've put this in on Weak+factorization+systems (joyalscatlab) at the end of proof 1 of proposition 3 - I hope I got the right diagram! - so you can see from that example how to use it.
Hope that helps!
@Andrew: This was very helpful. It works. Thanks!
I found something quite bizarre: I have reproduced two squares of objects and arrows in my Sandbox:
http://ncatlab.org/joyalscatlab/edit/Sandbox
The squares were first created in my page
http://ncatlab.org/joyalscatlab/edit/Weak+factorisation+systems
The system is refusing to compile the second, while it is accepting the first. But they look syntaxically identical to me! Beware that the second square was not obtained by copying the first. They differ by their history, their life line. The second guy was a problem from birth. A copy of the bad guy is bad and a copy of the good guy is good. Why is that? Equality is evil?
The closing 'center' tag wasn't right. In the second square you had '<center>' but it should have been '</center>'. Changing it made it work (see Sandbox (joyalscatlab)).
(Incidentally, you can link to pages in your web from here using the double-square-bracket notation. Type 'joyalscatlab:Sandbox' surrounded by double square brackets to make the link. If pasting in a URL as you do above, you should paste in the "show" URL, not the "edit" one - then there's no risk of people accidentally editing your page.)
This was my stupid mistake! My eyes are not good anymore. Thanks.
Is there a problem with the computer system? I was not able to compile new diagrams today. Also, compiling is very sluguish.
it seems to me that the CodeCogs server that generates the diagrams is down.
I just checked explicitly on their own equation editor webpage (here). While that page itself is available, it does not disyplay any equations/diagrams at the moment.
I think at the moment all we can do is wait a bit until they have restarted their equation server. In the long run we might maybe want to contact the people who run this software. Maybe there is a way that we buy the software or someting and install it on the nLab server itself.
Maybe there is a way that we buy the software or someting and install it on the nLab server itself.
I'm letting people get away with using codecogs for the moment, but I should warn you that on the nLab, I will be converting all the codecogs diagrams to SVGs using the SVG-Editor that Jacques has installed. The only real reason I haven't done so so far is time (and wanting to get a few bugs ironed out first). It really is impressive and it has the Killer Feature in that one can edit an existing diagram.
But depending on another site for serving part of the nLab is something I'd really like to get rid of. This just explains why.
I will be converting all the codecogs diagrams to SVGs using the SVG-Editor that Jacques has installed. The only real reason I haven't done so so far is time
Any chance that the process could be automated?
@Urs: I really really doubt it. I think codecogs takes the LaTeX, generates the diagram then "takes a picture". To somehow write that into SVG would essentially be "convert xypic stuff to SVG". It could be done (maybe even should be), but I can't see it being easy, and it's probably beyond our abilities to do it (unless any of you has secret programming skills that I'm not aware of [I know I don't!]).
Well, a crude way of automating it would be to pass the xy code through PHPLaTeX to generate a reasonable approximation of the SVG and then edit that in the editor to make it actually look nice. Or I could take a little time to clean up the SVG produced by PHPLaTeX (there's a few things that I know now that I didn't know last year when I wrote it) and then it would probably work okay on most xy diagrams without modification.
But Jacques has a few more ideas about the SVG editor so I'm holding back on the mass conversion until those settle down. This was more in the way of advance warning!
It would be nice if he could add a grid to it so we could arrange things more easily.
Edit: And apparently Andrew does have the abilities I doubted anyone here would have. Oh me of little faith. =)
It's worth making clear that the SVG editor does not originate with Jacques (though he is now allowed to commit code to the project) and he's not (unlike Instiki) intending to take it over (as far as I know). He's mainly focussing on making it work well with Instiki, and so his main focus is getting it to play nicely with itex. Stuff like having a grid has more of a feel of a suggestion that should be made to the main developers (who, I'm sure, would love to hear from people using their software!).
There is a crude way of doing what you say: it's possible to edit the co-ordinates for stuff directly, so you can put the stuff roughly in the right places and then line them up manually.
I am presently writing a page in the Catlab on model structures on the category Cat but I have messed up everything by pressing the wrong keys. I would like to return to verson 24 of the page. How can I do that ? Thanks.
I have found the answer to my question: I just rool back. Sorry for disturbing you.
Here is how to do it:
click on the link "History" at the very bottom of the page.
you'll be led to a page that lists all the different versions of the entry. Click on the one that you want to get back to.
Then when you see the page with the version you want, go again to the bottom and click on "Rollback". That will open the edit window as usual, but now with that version's source code.
Hit "submit" to save this as the latest version.
@joyal
In the list of contributors on the CatLab you have one 'Robert, David'. I hazard a guess that this is me, but there may be a David Robert out there that I haven't heard of. If this is me, shall I change it, and can I make a rudimentary page which just points to my nLab page?