LaTeX for quizbowl, again

[jonah]

I've read an old thread or two about this, but there seemed to be one thing missing: an actual template and class file that looked nice for producing packets. Putting aside the question of getting people to use it, I think there should be a good class file available for people who do want to use it. Is there already such a thing publicly available?

Believing the answer to the above question to be no, I attempted to create one. I'm not very experienced with LaTeX, and this could probably use a lot of improvement. Here is the class file as well as the code and PDF output for a sample packet (round 1 of 2010 Prison Bowl). Feel free to use the class file as is, or even better, suggest changes/improvements/better ways to do things. Or if I'm reinventing a wheel without improving it, let me know where the extant wheel is.

Two features not demonstrated in sample.tex: put

Code: Select all

\power{portion of the question in power}

in a \question{...} section to indicate powers, and you can add an optional argument to \part{...} of a bonus to indicate the point value like so:

Code: Select all

\part[point value]{part text}{answer}

(This defaults to 10 if not specified.)

[lchen]

PDF output for a sample packet (round 1 of the most recent Prison Bowl).

You nearly gave me a heart attack there....

This looks really nice. There's a weird character in word "Jörmungandr" in the first part of bonus 10 that I don't think is supposed to be there. It might be easier to use XeLaTeX instead of having to manually insert diacritics every time.

Also, this is 14 pages long whereas the original .doc is 8 pages long. This probably isn't a big deal if you're reading off a laptop screen, but some people might want to print out hard copies. Is there any reason for the font to be 12pt? I think font sizes in LaTeX are a bit larger than the corresponding font sizes in MS Word. (I usually use 11pt Computer Modern if I want to approximate "12pt Times New Roman.") Aren't packets usually typeset in 10pt font anyway?

[jonah]

PDF output for a sample packet (round 1 of the most recent Prison Bowl).

You nearly gave me a heart attack there....

Oops, sorry. As you have obviously discovered, it's 2010; I fixed that above.

There's a weird character in word "Jörmungandr" in the first part of bonus 10 that I don't think is supposed to be there.

Fixed. That was an artifact from the pronunciation guide style used in the original document (inserting middle dots between syllables of hard-to-pronounce words); incidentally, I like this method of doing things and would've replicated it except I can't find a way to do a \cdot outside of math mode.

It might be easier to use XeLaTeX instead of having to manually insert diacritics every time.

If someone can coherently explain to me the differences between LaTeX and XeLaTeX and how to use them (and tell me the packages I'm using will work, etc.), that would be great and I'll consider it. I'm not sure the diacritic issue is a huge deal, but this is something I've been wondering about for a while but not been interested enough to really research.

Also, this is 14 pages long whereas the original .doc is 8 pages long. This probably isn't a big deal if you're reading off a laptop screen, but some people might want to print out hard copies. Is there any reason for the font to be 12pt? I think font sizes in LaTeX are a bit larger than the corresponding font sizes in MS Word. (I usually use 11pt Computer Modern if I want to approximate "12pt Times New Roman.") Aren't packets usually typeset in 10pt font anyway?

I re-did it in 10- (11 pages) and 11-point (13 pages) Computer Modern in addition to 12-point. Right now it's three separate class files, one for each size, but at some point soon I'll figure out how to do it "right".

[Mike Bentley]

This just makes packets harder to parse for their post-tournament life. Don't ever, ever, ever do this kids!

[jonah]

This just makes packets harder to parse for their post-tournament life. Don't ever, ever, ever do this kids!

This seems like the exact opposite of what I think. Can you explain?

[Mike Bentley]

This just makes packets harder to parse for their post-tournament life. Don't ever, ever, ever do this kids!

This seems like the exact opposite of what I think. Can you explain?

I don't ever see someone uploading the LaTeX files to quizbowlpackets.com or something. All you're going to get are the PDFs, which are a pain in the ass to process.

[jonah]

This just makes packets harder to parse for their post-tournament life. Don't ever, ever, ever do this kids!

This seems like the exact opposite of what I think. Can you explain?

I don't ever see someone uploading the LaTeX files to quizbowlpackets.com or something. All you're going to get are the PDFs, which are a pain in the ass to process.

With that assumption, sure, but people could (and certainly should) still make the .tex source files available. I don't see this being widely adopted in the near future anyway, so I don't think we need to worry about it too much. I'm planning on using it for some personal projects and wanted to get the community's input, and also make it available for everyone else's use also.

[AKKOLADE]

With that assumption, sure, but people could (and certainly should) still make the .tex source files available.

Here is what a typical person will think seeing a .tex file.

"Oh God I don't want to download a virus. *Clicks X*"

[jonah]

With that assumption, sure, but people could (and certainly should) still make the .tex source files available.

Here is what a typical person will think seeing a .tex file.

"Oh God I don't want to download a virus. *Clicks X*"

People should make both the source files and the generated PDFs available, to avoid the problem you described and also to save time for the majority of users who don't want to compile the .tex file even if they know how.

[Schweizerkas]

There's a weird character in word "Jörmungandr" in the first part of bonus 10 that I don't think is supposed to be there.

Fixed. That was an artifact from the pronunciation guide style used in the original document (inserting middle dots between syllables of hard-to-pronounce words); incidentally, I like this method of doing things and would've replicated it except I can't find a way to do a \cdot outside of math mode.

Maybe I'm missing something, but can't you just insert $\cdot$?

[jonah]

There's a weird character in word "Jörmungandr" in the first part of bonus 10 that I don't think is supposed to be there.

Fixed. That was an artifact from the pronunciation guide style used in the original document (inserting middle dots between syllables of hard-to-pronounce words); incidentally, I like this method of doing things and would've replicated it except I can't find a way to do a \cdot outside of math mode.

Maybe I'm missing something, but can't you just insert $\cdot$?

I don't love the spacing, but it does work. I guess I could fiddle with the spacing manually.

[Cody]

With that assumption, sure, but people could (and certainly should) still make the .tex source files available.

Here is what a typical person will think seeing a .tex file.

"Oh God I don't want to download a virus. *Clicks X*"

People should make both the source files and the generated PDFs available, to avoid the problem you described and also to save time for the majority of users who don't want to compile the .tex file even if they know how.

Uh, who is this mythical reader who wants to transform a tournament from .tex -> .dvi -> .pdf instead of just getting the end product?

In addition, you should reread one of those old threads, which came to the conclusion that using LaTeX would be a dumb move because it is needlessly complicated for quizbowl's needs. I believe the conclusion they reached was to use Almost Free Text due to its simplicity and its ability to provide everything needed.

[grapesmoker]

I don't think we came to the conclusion that it would be a "dumb move." After all, LaTeX is a consistent markup and thus fairly straightforward to parse. The limitation is in how useful that would be to people.

Here's some code I used to use for typesetting tossups and bonuses:

Code: Select all

\newcounter{TossupCounter}
\newcounter{BonusCounter}
\setcounter{TossupCounter}{1}
\setcounter{BonusCounter}{1}

\newcommand{\Tossup}[2]{{\noindent\arabic{TossupCounter}. \addtocounter{TossupCounter}{1}} #1\\ \indent Answer: #2 \\}
\newcommand{\Ans}[1]{\underline{\bf #1}}
\newcommand{\BPN}[3]{[#1 points] #2 \\ \indent Answer: #3\\}
\newcommand{\BPE}[3]{[#1 points each] #2 \\ \indent Answer: #3\\}

\newenvironment{Bonus}[1]{\noindent\arabic{BonusCounter}. \addtocounter{BonusCounter}{1} #1 \\}

[AKKOLADE]

You guys do realize that most people don't have nor want any background in coding, right?

[Mechanical Beasts]

You guys do realize that most people don't have nor want any background in coding, right?

In terms of complexity, this is nothing more than, like, formatting text. If you can click bold, you can type \textbf. It takes no background in programming to understand. But that's not much the point--the only people who would ever need the .tex files (which should be made available, sure) are the people who are writing packet parsers for their archives, and if they want to do that but are scared of a .tex file, why on earth are they doing that?

[Skepticism and Animal Feed]

Since the last time the LaTeX cult emerged on HSQB, there have been great advances in the ability of non MS-Office software on many platforms to read MS Office formatting.

There is now almost no moderator or quizbowl player who can't read a packet formatted in MS Word 2003 or 2007. I'm also not aware of any quizbowl packet archive that chokes on either of these formats. And if DOC or DOCX fails, it is very easy to convert either one of these to a PDF, which, again, almost everyone can read.

[Mechanical Beasts]

Since the last time the LaTeX cult emerged on HSQB, there have been great advances in the ability of non MS-Office software on many platforms to read MS Office formatting.

There is now almost no moderator or quizbowl player who can't read a packet formatted in MS Word 2003 or 2007. I'm also not aware of any quizbowl packet archive that chokes on either of these formats. And if DOC or DOCX fails, it is very easy to convert either one of these to a PDF, which, again, almost everyone can read.

Less a cult and more "we see certain benefits to LaTeX, including but not limited to the ability of databases to parse questions incredibly easily, and few drawbacks." Bruce, as a longtime proponent of the way the canon helps new players, why not support a position that would make it far easier for a new player to determine exactly how many times an answer has come up before?

[grapesmoker]

I know this is dangerously close to "telling people how to post" territory, but why is it that people who don't know anything about LaTeX feel irresistibly compelled to post in these threads and tell us how stupid we are?

We get it, dudes. No one wants to do background coding, people can read Word, etc. For SOME INSCRUTABLE REASON we thing this might be a thing with some benefits to a small subset of the community. If that reason is indeed inscrutable to you, that's fine, you're not the intended audience for this particular geekout. You are perfectly free to continue on your merry way using Word just like everyone (even me!) in quizbowl continues to do. No one is going to make you switch to some horrible system you hate, we're just trying to think of ways that people who like to work with LaTeX can improve packet formatting for others without inconveniencing anyone.

[Skepticism and Animal Feed]

If the proposal is to have LaTeX become the language in which packets are written and submitted, I think that would make quizbowl so intimidating and confusing for new players that it would outweigh the ability to quickly generate frequency lists. I would vociferously oppose such a move by posting (repeatedly) on the internet against it.

If the proposal is to have a bunch of LaTeX people re-format packets into LaTeX in their spare time, and thereby create tools that other players of varying levels of LaTeX proficiency can use, then I can't say I object.

[grapesmoker]

If the proposal is to have LaTeX become the language in which packets are written and submitted, I think that would make quizbowl so intimidating and confusing for new players that it would outweigh the ability to quickly generate frequency lists. I would vociferously oppose such a move by posting (repeatedly) on the internet against it.

Unless I mis-read what Andy is saying, no one is actually proposing this.

If the proposal is to have a bunch of LaTeX people re-format packets into LaTeX in their spare time, and thereby create tools that other players of varying levels of LaTeX proficiency can use, then I can't say I object.

Thanks for your magnanimous leniency, we sure do appreciate it.

[Mechanical Beasts]

Thanks for your magnanimous leniency, we sure do appreciate it.

jerry can we have a picture of bruce in the cult facility? the peons might like it.

[AKKOLADE]

I know this is dangerously close to "telling people how to post" territory, but why is it that people who don't know anything about LaTeX feel irresistibly compelled to post in these threads and tell us how stupid we are?

Because I do know something about LaTeX?!?!//1/1/1//11oneforwardslash

[grapesmoker]

I know this is dangerously close to "telling people how to post" territory, but why is it that people who don't know anything about LaTeX feel irresistibly compelled to post in these threads and tell us how stupid we are?

Because I do know something about LaTeX?!?!//1/1/1//11oneforwardslash

Ok, look: I've been working with LaTeX since I was an undergraduate. I've written articles, presentations, and a thesis in it. While I don't claim to be an expert in it by any means, I feel like I have a reasonably high level of proficiency in it, which, no offense, but I'm guessing is probably higher than yours. As such, I think I have some reasonable claim to authority on the usefulness of LaTeX. Not, mind you, its usefulness to the majority of the quizbowl community in its raw form, but rather to people who can come up with a good way of using it as a backend for producing something the community will find valuable, like pretty-looking PDFs of packets.

[Skepticism and Animal Feed]

Unless I mis-read what Andy is saying, no one is actually proposing this.

Good to hear. It's definitely been proposed before though -- perhaps the cult today doesn't have the same messianic fervor it did when Evan Silberman was its prophet.

I have to imagine that most anti-LaTeX posts in this thread are the result of somebody either thinking that people were arguing for writing/submitting/distributing to TDs packets in LaTeX format, or acting pre-emptively because they thought the conversation was headed in that direction.

[AKKOLADE]

Right, and I also used it in undergraduate - not nearly as much as you, I'm certain, but I did spend some time with it. I recognize that it's extremely useful for certain uses; it'd probably be pretty swell for my chemistry lab reports, as a case in point. But I know how it works, and I also know how people work with computers. That isn't rocket surgery.

All I'm saying is that its usefulness for quizbowl is extremely limited. The requirements of formatting a packet aren't that tough; it doesn't involve anything more complex than a tab every once in awhile in Word. I don't know why people would use LaTeX for a qb packet, but if you want that, and if your output matches what the tournament hosts require, I don't give an inscrutable crap.

Fake edit: and yeah, I was getting preemptive. I've seen things, man; I've seen LaTeX discussions.

[Cody]

I don't think we came to the conclusion that it would be a "dumb move." After all, LaTeX is a consistent markup and thus fairly straightforward to parse. The limitation is in how useful that would be to people.

Perhaps "dumb move" was too strong a wording, but it doesn't have as much of an advantage as other available systems that are easier for the "average person" to use use:

That said, there are practical considerations to keep in mind. Obviously, we want to make life as easy as possible for the writers and the editors. They shouldn't have to learn a complicated markup, and there's really no need to do so. Moreover, writing in markup when you don't know it all that well can cause problems because you mis-type something or you forget a closing tag or whatever.
...
Most of the objections I've heard voiced to employing such a format are not really valid, and I hope that this kind of markup renders the final objection (too complicated to learn) also null.

After which the entire conversation changes.

Not, mind you, its usefulness to the majority of the quizbowl community in its raw form, but rather to people who can come up with a good way of using it as a backend for producing something the community will find valuable, like pretty-looking PDFs of packets.

Wouldn't a far superior solution be to find something that is useful to the majority of the quizbowl community in its raw form, so that it could be widely adopted and implemented and have a better chance of becoming a de facto standard? I'm not saying AFT or Markdown or format X is definitely what we should be using, but it seems far more useful to people who write questions than LaTeX. Also, I imagine, to editors of tournaments.

Edit: clarity about "or something"

[grapesmoker]

Right, and I also used it in undergraduate - not nearly as much as you, I'm certain, but I did spend some time with it. I recognize that it's extremely useful for certain uses; it'd probably be pretty swell for my chemistry lab reports, as a case in point. But I know how it works, and I also know how people work with computers. That isn't rocket surgery.

By the by, I'm extremely pleased to learn that my alma mater is running tournaments and such. I should drop by campus when I'm in San Diego next.

Anyway, I think LaTeX, like any regular plain-text markup is quite useful. The major problem for parsing things in Word is that Word is extremely inconsistent with how it does things. Just because you see the same thing doesn't mean the same representation is being used under the hood. One very obvious example of this is that sometimes, for reasons I still do not understand, some packets end up having random newlines sprinkled all over the place. Other packets end up with weird non-printable characters; again, where these come from is unclear. Word in general doesn't give you access to the document semantics. That is, there's no way to know what's a tossup or what's a bonus, it all has to be inferred from its relative position in the document, which is crazy when you think about it. In general, it is surprising how brittle Word formatting is; I encourage anyone who doubts this to take a look at the packets from, I believe, 2007 Ghetto Warz, which must have made a special effort to break formatting.

[grapesmoker]

Perhaps "dumb move" was too strong a wording, but it doesn't have as much of an advantage as other available systems that are easier for the "average person" to use use:

I think we've always spoken of LaTeX as a preferred back-end, not as something everyday users would write in.

That said, there are practical considerations to keep in mind. Obviously, we want to make life as easy as possible for the writers and the editors. They shouldn't have to learn a complicated markup, and there's really no need to do so. Moreover, writing in markup when you don't know it all that well can cause problems because you mis-type something or you forget a closing tag or whatever.
...
Most of the objections I've heard voiced to employing such a format are not really valid, and I hope that this kind of markup renders the final objection (too complicated to learn) also null.

After which the entire conversation changes.

I guess I'm not really sure what you're saying here. I think I stand by what I said in that thread. It is a historical fact that for a long time, people really did write packets in plain text, using just the sort of formatting I described in that thread. So we know that it's neither impossible nor particularly onerous. Can we get people to adopt it? I have my doubts, I guess. What I'm thinking is much more likely is that an online submission system such as QEMS will use LaTeX as a backend and give people a web-based frontend that allows them to write in rich text.

Wouldn't a far superior solution be to find something that is useful to the majority of the quizbowl community in its raw form, so that it could be widely adopted and implemented and have a better chance of becoming a de facto standard? I'm not saying AFT or Markdown or something is definitely what we should be using, but it seems far more useful to people who write questions than LaTeX. Also, I imagine, to editors of tournaments.

Well, as I've said in the very thread you linked to, yes, of course. However, as I've also said in that thread, I want to come up with a solution that removes the need for any end-user to actually do their own markup. Which is why I think something like a web-based QEMS with rich-text editing is a viable solution to this problem. The only real thing that needs to be hashed out among the nerds is what the final representation will be. In other words, a standardized QBML/LaTeX markup for packets.

[jonah]

Yeah, I thought I made this clear by saying "putting aside the question of getting people to use it", but apparently not: I am not expecting anyone to write or submit their packets in LaTeX. If they want to and editors want to accept it, great. The purpose of this is twofold: (1) for the output of tournament production systems that are more fancy than Google Docs and email—one of which is among the next few items on my list of quizbowl projects—and (2) for the use of any tournaments that do have an interest in packets that are pretty and easily parseable.

Also forthcoming: A proposal for a more comprehensive XML application for quizbowl (I'm soliciting suggested names other than QBML) along with a script to turn that into LaTeX.

[grapesmoker]

Also forthcoming: A proposal for a more comprehensive XML application for quizbowl (I'm soliciting suggested names other than QBML) along with a script to turn that into LaTeX.

Is something wrong with the name QBML? And what deficiencies in the existing schema are you trying to correct? I'm genuinely curious about what you think is missing from the current specification.

As for a script for LaTeX conversion, I believe I have that somewhere. In any case, recursively crawling an XML tree with Python is a trivial task.

[theMoMA]

An internet tool that would allow users to submit plaintext questions (with markup for simple, necessary formatting like required answers and italicizing) would be good for this. Using something similar to HSAPQ's QEMS program, individual players from packet-submitting teams could submit questions as they finish them (packet discounts starting when the packet is complete), the editors could edit them in plaintext, and the result could be outputted as a nicely formatted PDF (or RTF for those who prefer to be able to edit in Word, or as a simple markup document for easy archive parsing). Editors would get more questions sooner in an easier forum for editing and everyone would get pretty PDFs.

[grapesmoker]

An internet tool that would allow users to submit plaintext questions (with markup for simple, necessary formatting like required answers and italicizing) would be good for this. Using something similar to HSAPQ's QEMS program, individual players from packet-submitting teams could submit questions as they finish them (packet discounts starting when the packet is complete), the editors could edit them in plaintext, and the result could be outputted as a nicely formatted PDF (or RTF for those who prefer to be able to edit in Word, or as a simple markup document for easy archive parsing). Editors would get more questions sooner in an easier forum for editing and everyone would get pretty PDFs.

The key to getting a system like this up and running and also making it useful is to have the correct permissions logic. This is something that in that past I've found somewhat difficult to do. Once I'm done with Nationals, I might take it up again; it might also be a good idea for anyone who is interested in working on something like this to join forces and divide labor.

[jonah]

Is something wrong with the name QBML?

Just that it's already doubly taken: by your specification and by NAQT. I don't see this as a very significant issue, but it would be nice to distinguish things by name. If nothing else, QBML2 works.

And what deficiencies in the existing schema are you trying to correct? I'm genuinely curious about what you think is missing from the current specification.

I'm assuming this is the existing schema, since I haven't seen anything else. Anyway—a few things, few of which are hugely significant (and most if not all of which could be omitted in many cases), but all of which I think would be nice to have:

• Formatting tags. Preferably semantic ones akin to \majtitle{} and \mintitle{} in the LaTeX specification. Also, a tag for pronunciation guides.
• Powermarking.
• Attributes in the <tossup> and <bonus> tags to indicate question-level metadata like category information, type of answer (person/work/conflict/...), location, language, time, etc. Additionally or alternatively, an id="" attribute to refer back to the question's key in a database table.
• Other semantic tagging for things like quotes, and perhaps even lead-ins and giveaways.
• Better modes of indicating alternate/promptable answers. For example, <alt type="accept" restrict="beforemention">...</alt> and <alt type="prompt">...</alt>.

I've actually got most of this done in the form of a sample document and a stylesheet; I just need to clean up a few things. I'll probably post that next week.

An internet tool that would allow users to submit plaintext questions (with markup for simple, necessary formatting like required answers and italicizing) would be good for this. Using something similar to HSAPQ's QEMS program, individual players from packet-submitting teams could submit questions as they finish them (packet discounts starting when the packet is complete), the editors could edit them in plaintext, and the result could be outputted as a nicely formatted PDF (or RTF for those who prefer to be able to edit in Word, or as a simple markup document for easy archive parsing). Editors would get more questions sooner in an easier forum for editing and everyone would get pretty PDFs.

Yeah, this is the eventual goal. I'm working in steps here, with the side benefit that some parts of it might be useful to people independently of other parts.

[ezubaric]

An internet tool that would allow users to submit plaintext questions (with markup for simple, necessary formatting like required answers and italicizing) would be good for this. Using something similar to HSAPQ's QEMS program, individual players from packet-submitting teams could submit questions as they finish them (packet discounts starting when the packet is complete), the editors could edit them in plaintext, and the result could be outputted as a nicely formatted PDF (or RTF for those who prefer to be able to edit in Word, or as a simple markup document for easy archive parsing). Editors would get more questions sooner in an easier forum for editing and everyone would get pretty PDFs.

This is what my now defunct Jerome system did. (Although it did it used LaTeX markup internally in the submission box.) A friendlier input format would have likely made it easier for people to adopt. I also got too ambitious with it, and it died after being pulled in multiple directions.

[grapesmoker]

Just that it's already doubly taken: by your specification and by NAQT. I don't see this as a very significant issue, but it would be nice to distinguish things by name. If nothing else, QBML2 works.

I certainly don't mean to steal that name; can't say whether NAQT cares about the overlap but I'm guessing not.

Anyway—a few things, few of which are hugely significant (and most if not all of which could be omitted in many cases), but all of which I think would be nice to have:

• Formatting tags. Preferably semantic ones akin to \majtitle{} and \mintitle{} in the LaTeX specification. Also, a tag for pronunciation guides.
• Powermarking.
• Attributes in the <tossup> and <bonus> tags to indicate question-level metadata like category information, type of answer (person/work/conflict/...), location, language, time, etc. Additionally or alternatively, an id="" attribute to refer back to the question's key in a database table.
• Other semantic tagging for things like quotes, and perhaps even lead-ins and giveaways.
• Better modes of indicating alternate/promptable answers. For example, <alt type="accept" restrict="beforemention">...</alt> and <alt type="prompt">...</alt>.

The first three seem sensible (as they are consistent). The next one seem like serious overkill. You don't really need to tag leadins and giveaways, I think. As for the "accept x before y" that seems like a pretty ambitious differentiation, although I think that if you want to feed this to some automated reader later on, you do need to know what the minimal set of acceptable things is. Looking forward to seeing that stylesheet.

[jonah]

The first three seem sensible (as they are consistent). The next one seem like serious overkill. You don't really need to tag leadins and giveaways, I think. As for the "accept x before y" that seems like a pretty ambitious differentiation, although I think that if you want to feed this to some automated reader later on, you do need to know what the minimal set of acceptable things is. Looking forward to seeing that stylesheet.

The lead-in and giveaway tags would be so that someone could say "show me all the lead-ins from tossups on the Payne-Aldrich tariff act". That and the answer-acceptability tags are things that I would consider extremely optional, but I'd like the specification to support lots of reasonable things that might not actually get used very often, so if people want to do something neat they can.