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.
I've switched us over to the all new, all singing, all dancing MathMLised nForum. With a bit of luck, no-one's noticed so far![1] I've tried to make it so that the only changes are good changes and that they only work positively, in that if I didn't tell you about the upgrade then you wouldn't notice. That's probably not quite possible, though, due to the fact that as we're serving up MathML now then we have to be a bit stricter on our formatting.
Anyway, the major change is that there's a new text filter, 'Markdown+Itex'. This is pretty close to the nLab format. There's a few things on the nLab that don't work here, but those are things like theorem environments. Basically, if it uses curly braces, it won't work here. Hopefully, though, everything that works here will work there.
The other one worth mentioning is the "source" view. If you look up to the top right, there's a word "Source" (next to "PermaLink"). If you click on that, you see the original source of the comment. Just right (I hope) for cutting-and-pasting.
Minor changes:
I've upgraded the old way of doing maths so that it now ought to work with all text filters (except Markdown+Itex for obvious reasons) and the preview mode. If it doesn't, let me know.
As we're serving up MathML, we have to be sure that it's properly valid otherwise browsers won't show it. So comments get validated. After strong protests (you know who you are), you can submit invalid comments, but then you just get the source displayed. If you preview your post first then you can check whether or not it is valid. If it is your comment (and in preview mode) then you also see what the errors are (so that you can correct them). The most likely errors are: ampersands should be entities (so & instead of &), and (for those using the Html formatter), <blockquote> needs a <p> just inside it (I may modify the Html formatter to check for this as so many old comments have this problem).
The stylesheets might have gotten a bit mangled in the crossing. If it suddenly looks weird, please let me know and I'll try to fix it for you.
[1] Yes, I know. Hopefully you're the only one.
Let me see, I remember that, unlike in the lab many of the symbols with diacritics did not work before here. Let me try ?š??ž
You know, it really would help if I knew what I was doing ...
Apologies! For some reason, I had the wrong setting for the old-style latex converter so it wasn't seeing the old-style mathematics and then the XHTML sanitiser was stripping out the "illegal" mathematics. Fixed now (I hope).
(Presumably, Harry, this is what you were referring to as well; am I right?)
Yes they still do not work – as you see š and ž work but the other 3 Croatian symbols (including those corresponding to LaTeX \v c and ' c and d with the horizontal line across (dj, similar to h bar for Planck constant). All these do work in lab.
Even when I put those symbols they not only do not appear but also when I go back to reedit the comment they are in the meantime replaced by question marks so the work is lost. Of course, that problem was earlier as well in nForum.
Testing characters: ? ?
Okay, they looked fine in the preview but got mangled after submitting them. This suggests that it's something to do with the database and how it stores unusual characters. I didn't realise that this was a problem - I'll look in to it.
What is the correct way to do ordered and unordered lists?
@Zoran: A bit of mucking around in the database shows that I'm right - so I just need to learn how to safely convert a database from swedish (?? why swedish as default ??) to utf8 and then all will be fine. Luckily, my kids go to school with the son of a MySQL developer!
@Eric: same as on the nLab: * for unordered, 1. for ordered. However, your stylesheet may be mucking up the difference. Here's an ordered list:
Here's an unordered:
(If you look at "source" you'll see the code!). If they don't look like an ordered and unordered list, then the stylesheet is being silly and I need to adjust it.
Hmmm… what is wrong with this comment?
Nevermind. The error messages were actually helpful. Imagine that! :)
That reminds me, if you use Markdown+Itex then the old-style latex->pictures is turned off. So to get the maths to display correctly in that comment you should either use just the Markdown filter or use itex.
(Did you write the mathematics using <latex> tags or using double-dollars?)
(Ignore that: I didn't realise that it was an old comment that you were editing, so how you entered the mathematics is irrelevant. But if you wanted to convert it to MathML, you need to convert all those <latex> tags back to single dollars.)
By the way, I like the “Source” button. It gives clues on how to write stuff the right way by peeking at other’s comments. All this time I was using html for ordered and unordered lists here :)
I don't know about anyone else, but I've stopped getting emails informing me of new comments as of today. Since this MathML malarkey was a big change, one could surmise that it's the cause, but seems unlikely to me.
Actually, it's very likely to be the cause since what I didn't tell you is that I haven't just modified the forum to be MathML-capable, I switched the whole forum under your noses! (There was an upgrade to the forum software due as well, and I've been making loads of little tweaks over time and I figured it was time to make them a little more systematic) So I'm not surprised that other things have broken. I'll have a play and see if I can track it down.
Thanks, Andrew, once again, this is very nice.
It would be useful if, when the dust has settled, we’d have a concise HowTo somewhere. Right now from the discussion I am not 100% sure that I understand what works when, but I gather I’ll quickly see it in practice.
Let me test something:
That's what the nForum FAQ page is meant to be for. I guess I need to update it.
Notification is back. It was managed by a plugin which has vanished from the main repository so didn't get installed in the new version. Fortunately, I still have a copy from the old forum and simply copied it across.
(I don't know if all the old notifications will simply pick up where they left off. If not, you may need to re-enable whatever notifications you had set up. Please make a note here of what happens in either case so I know for future. Thanks)
Thanks, Andrew. I've got some coming in nicely.
Now if only the nCafe had a “Source” button :)
That is so awesome :)
Now if only the nCafe had a “Source” button
It does; it’s called “View PGP Sig.” The only catch is that the person who wrote the comment in question needs to have PGP-signed it. I usually PGP sign my comments that contain significant math for this reason, but no one else seems to have picked it up.
no one else seems to have picked it up.
I did this years ago, but stopped doing it because I felt that the software should help me save time and not I should help the software get its work done.
It’s hard, because it’s sort of an altruistic action. Signing my comments doesn’t save me any time, but if other people sign their comments, it does sometimes save me time.
Signing comments is also really fast and easy with FireGPG, were you using something like that?
@Urs
lab is draining pretty much energy from me, I will end in hospital if lab appears soon…
I like the transition diagram :)
Andrew, has anybody mentioned lately that you rock? (^_^)
Toby, I think someone did say that somewhere ...
I managed to break Domain theory.
I wrote a comment, previewed it and nothing appeared. Just an empty “Preview” as if I had hit “Preview Post” without typing anything. This has happened to me before.
I decided to go ahead and post it to see what would happen. Now that discussion seems inaccessible. I hope we can recover the comment, but not the end of the world if we can’t.
Signing comments is also really fast and easy with FireGPG, were you using something like that?
No, back then I used a really complicated procudure that quickly drove me mad. That may be part of the reason.
I don’t know how you did it, but you’ve managed to insert a strange unicode character in one of the words in your comment. It was in the bit you cut-and-pasted from earlier so it really oughtn’t to have happened. Unfortunately, I don’t seem able to access the database from home so can’t easily go in and remove the character!
Clearly, there’s more about character sets and encodings that I need to learn.
Me? In which word? I can’t see anything strange on my system.
Sorry! That was a reply to Eric’s comment. He’s managed to insert some strange non-printing character in a certain word in the discussion at domain theory with the result that the discussion is refusing to display.
Andrew, do you happen to know anything about why in Markdown+Itex mode it seems the boldface math is not displayed in boldface?
My source code for
is
\mathbf{\Pi}(X)
but, at least on my system, I don’t see it displayed in boldface.
I can't see it in bold using Firefox 3.0.10 (ugh, work internet - nothing I can do to update)
The XHTML code looks right (the Pi gets a ’mathvariant=bold’ attribute) so I’m guessing that it’s a CSS thing. Presumably what the browser does with bold text is completely determined by CSS (rather than there being a sensible default which can be overridden by CSS).
I haven’t done much (anything) with the CSS here with regard to the upgrade. I guess I should copy over the maths bits from the nlab. There’s a few other bits and bobs of CSS that also need updating - and then there’s still the outstanding issue of CSS on the nlab itself - so I was going to have a “CSS week” in a couple of weeks to look at all of these (our semester ends in a couple of weeks time so I’ll have a bit more time then). I can hack something for this particular one if you think it needs solving right now, if you like.
Note that the itex here and on the nlab should be 100% identical since we’re using the nlab to generate the mathml from the itex. So any differences like this are almost certainly down to CSS. But definitely worth pointing out since we want the nForum to mimic the nLab as much as possible.
@David: you can always install a local copy of the latest firefox. That’s what I did on my work machine. What system are you using?
@David: bleugh. Still, is there a reason why you can’t install a local version of the latest firefox?
Update on the continuing saga of character encodings: apparently, not all unicode characters are valid XHTML. Eric seems to have found one; what’s particularly interesting about Eric’s is that it is valid HTML so I suspect his browser is working in HTML mode (are you using IE, Eric?) and so thought it was acceptable to send the horrible character to the server, which then had kittens as it wasn’t valid XHTML. This has actually led to me asking a question on stackoverflow, so Eric gets this month’s “Bizarre Bug” prize!
Right. A quick hack means that Eric’s comment is now visible. Eric, to make it valid XHTML then you need to go in and do the following: in the quoted section delete the word “first” and then rewrite it. The illegal character is invisible and is somewhere around the ’r’ so just deleting and rewriting the whole word will get rid of it.
(NB a more permanent solution will remove such illegal characters automatically so you could leave it as it is and wait for me to implement that solution. However, that will depend on the answer to this stackoverflow question.)
Thanks Andrew. In pdf’s the two letters “fi” are often combined into one character when you copy and paste. That is probably what happened.
Ack. The evil character won’t let me edit. Your fix let me at least see the discussion, but when I click “edit”, I get that evil dead page again.
This has actually led to me asking a question on stackoverflow, so Eric gets this month’s “Bizarre Bug” prize!
It’s an honour! :D
Edit: Corrected a shameful typo :)
Actually, it's an honoUr since I'm doing the awarding and I'm unashamedly a Brit.
Okay, try editing your comment now. I've added the "remove evil unicode" function in a couple more places so that it should work now.
(It wasn't the ligatured 'fi', that's perfectly valid. What it was was U+000C which is the unicode character for a form-feed. According to the w3 docs, it's perfectly valid HTML but invalid XHTML. What makes me wonder a little is why your browser sent it in the first place, which is why I asked about IE: if your browser thought it was sending HTML then it would have thought it was okay to send that character, but the server sends XHTML, so that character is invalid. However, I don't know enough about how different browsers work to make more than dumb guesses, nor do I particularly want to know how they work - I just want to know how to make sure that any output is valid, without any particular assumptions on the input.)
Still, is there a reason why you can't install a local version of the latest firefox?
I tried it, but it doesn't work. Actually, it downloaded, installed (to 'C:/' but that doesn't seem to exist for me), but then didn't launch when it should have, and now I can't find it. For all intents and purposes, there is no 'local': everything runs off some server or other, and I don't have permissions to do anything remotely like installing programs. We don't even have some standard programs (EndNote for example, and we are a research organisation), because they don't run under Citrix. And it's company policy to not have any software that doesn't come with vendor technical support, so no open source anything. >:-(
Edit: Deleted a sincere but off-topic comment :)
@David: Ouch! So at home you have problems with editing the nLab due to the overzealous spam filter, and at work you are restricted to using Bad Software.
(Any chance of installing firefox on a USB stick and running it from there? Or even installing Linux on a USB stick and running that!)
I can edit at work, but of course not too much - I've toyed with editing stuff offline and just uploading it all at once, but for some reason it hasn't caught on yet.
I could try the USB stick - but the system being what it is, I can see the stick as soon as I plug it in, but need to log out and in so that other software that is running can see it. We've been promised an upgrade eventually (at leas, they're running a pilot study to have virtual desktops and with that, USB hot plugging - yay!)
Anyway, enough gripes. I can get by - the forum works ok.
Actually, a thought just struck me: with our new 'source' button I can type (at home) big chunks of potential lab material here and it can be copied, according to my detailed instructions, into a lab page by a friendly lab elf, or indeed myself (at work). ^-^
Actually, a thought just struck me: with our new ’source’ button I can type (at home) big chunks of potential lab material here and it can be copied, according to my detailed instructions, into a lab page by a friendly lab elf, or indeed myself (at work). ^-^
I was going to suggest this. It’s a sad thought that you are sitting there eager to contribute to the Lab, and the software won’t let you. Post it here, and somebody can paste it in for you.
1 to 45 of 45