@AlanMunn Of course! This one is an adaptive automaton (well, it's going to be one) and it's mapped to recognize a group of people based on some caracteristics. Do you remember a game called "Guess who?" ? It's very similar. I have to draw another automaton with a "real" example, so I think it's better to explain with the other one. (cont)
The accepting states (double circle ones) represent people, let's say lockstep, Martin, you and me. the other "c" states represent some caracteristic, let's say reputation. So I can query people with rep > 3000, and the automaton will change its topology according to the input (hence the adaptive term). Only lockstep, Martin and you will be "reachable" by the automaton (the state representing me will be unavailable).
@AlanMunn I have to finish the paper until Monday night, I'll send you a copy. It's very intuitive - and quite funny! The "boring" part is the theory. :-)
@Paulo: In my regex code, I am using automata. Do you think that providing automata as a stand-alone feature would be useful? Are automata a useful programming tool for users or package writers? (I've only ever used them as a theoretical object, or for that regex code.)
@AlanMunn ... I am, I know. Actually, I'd probably schedule automata for after the completion of the floating module. But since I'm looking at them right now, I might try to code things in a manner that makes coding for automata later easier.
@BrunoLeFloch Bruno, I share your thoughts on using them as theoretical objects. Since the language recognized by a automaton/finite state machine is a regular one, they are pretty good for regex - they grow in space, not in time. In the worst case, if I'm not wrong, it's O(n) for string acceptance, considering n as the lenght of the string. Unfortunately, apart from our theoretical views and the extraordinary regex use, I don't see a better application for casual users or even programmers.
@PauloCereda Ok. Thanks for listening to Alan ;-). The complexity is O(<length of string>*<size of automaton>) in a programming language that provides O(1)-access arrays. In TeX, I'm not too sure what my algorithm gives, but I doubt that I reach that when tracking submatches.
I'm already abusing of TeX's toks registers and other registers, accessing them directly by number, and manipulating pointers myself.
E.g., if a programmer uses a boolean in his code, that might be represented by a two-state automaton. One problem I face when dealing with automata is that my "world" is quite volatile. Sometimes, I have to microprogram things and the thing grows a lot. :-)
@BrunoLeFloch I'd love to see a nice automata implementation!
@BrunoLeFloch Joseph was not kidding when he said you are the Optimization Jedi Master. :-)
@BrunoLeFloch yes, they are, according to the input symbol. There might be a function associated with a transition and triggered to remove/create states/transitions.
@PauloCereda Most of my LaTeX3 work has been to go through the modules one by one and go from O(n^2) to O(n). Or whatever was possible. Joseph told me once that before my changes the siunitx doc took 30s to compile, and now only 9s.
4
@PauloCereda That's cool. But probably difficult to reason about, isn't it?
@BrunoLeFloch wow! I wonder what you can do on PC overclocking or car tuning. :-)
@BrunoLeFloch Not at all. :-) I'll try to translate to English the section of my master's thesis about adaptive automaton. There's a nice example in which you can see how easy it is. :-)
@PauloCereda Trodding along. I don't really have a precise project. At the moment, my supervisor teamed me with an old friend of mine, and we're working on localization of a supersymmetric gauge theory on S^2 (sphere). Too close to string theory in my view, but I shouldn't say that too loudly :).
@AlanMunn I get no error if I say, before \usepackage{pgfkeys,pgfmath}, also \input{pgfutil-common}. Definitely not user friendly. Maybe @Martin can tell more.
@egreg The PGF package have poor package handling code in it. You can't load pgfmath on its own any longer, for example. This is caused by the mixed support for plainTeX, LaTeX and ConTeXt. All generic files should check if they were loaded already by themselves.
I mark the following question as duplicated in relation to Martins comment. What do you think: http://tex.stackexchange.com/questions/9740/how-can-i-add-vertical-space-to-a-beamercolorbox-to-make-it-align-with-another-on
That's true. But in the comment Martin provides a link to the correct answer: http://tex.stackexchange.com/questions/14220/how-to-modify-columns-column-environments-so-they-resize-automatically-to-the-lar/21591#21591
@Marco Indeed. But until now I don't think we closed older questions after a canonical answer had been provided to a similar newer question. You could ask a question at meta what our policy should be.
Here's an example where I tried to provide a canonical question/answer to float numbering. So far, we haven't closed the (about half a dozen) earlier questions that provided parts of my answer.
(This is intended as a consolidated reference for one of the "frequently asked questions" at tex.sx.)
Some document elements (e.g., figures in the book class) are numbered per chapter (figure 1.1, 1.2, 2.1, ...). How can I achieve continuous numbering (figure 1, 2, 3, ...)?
And vice versa: Some...
@lockstep I don't believe that the order of appearance is important for judging what duplicate to keep open. In this case, the more recent question is quite detailed and the accepted answer is very good; moreover no answer to the older question was accepted, which can leave in a casual reader the impression that the problem has not yet been really solved.
I'm not a {Open|Libre|Neo}Office user, but I may have need to convert a Word document to LaTeX, and Writer2LaTeX looks like it will do the trick. Does anyone have any experience with these three incarnations of OpenOffice on the Mac, and recommendations? Mac users especially: Is NeoOffice up to date relative to the others?
Also, any advice on Writer2LaTeX use. The document doesn't have any math; what I would like is minimal markup if possible, since much of the distinct formatting will be done with a package.
@AlanMunn I had NeoOffice installed, but I think it's better to give LibreOffice for Mac a try. Nothing against NeoO, but since it looks like a LibreO fork, it might not cover everything.
I've never tried it before, but it seems pandoc is able to convert odt or rtf to tex.
@PauloCereda Thanks. And I couldn't quite figure out the Libre vs. Open Office difference. There are definitely some political issues related to Oracle, but it wasn't clear from googling whether there were substantive differences between the two. I'd forgotten about pandoc; that also might be a way to go.
@PauloCereda Actually, it looks like pandoc can go from TeX -> odt/rtf but not the other way around.
@lockstep Now, now, no fighting boys. "the last thing I hit before you" is ambiguous between your intended meaning and "the last thing I hit before I hit you" :-) (One of the hazards of having a linguist around.)
@lockstep Not really, all you need is a robust distinction between nominative and accusative in this case. But maybe German is like English and the accusative in this structure can still yield both interpretations.
e.g. John met Mary before me. = John met Mary before I met Mary or John met Mary before he met me. Can German say that?
@JosephWright The second sentence is the third paragraph is missing an βisβ: My impression [...] is that this is quite a laid-back task compared ... β
@JosephWright I don't know enough about the different drivers to know if the solution to the beamer shading problem is simply to have pgfsys-xetex input pgfsys-dvipdfmx.def. Do you know?
@egreg As I understand it, you prefer to find out the solution yourself whenever a (La)TeX question arises. But it would still be interesting for "the rest of us" to follow the process through a (self-answered) question by egreg posted at tex.sx. (And such a question would be a worthy no. 10,000.)
I've implemented some of the things I mentioned in my earlier answer, but I felt this was different enough to get a separate answer.
Once again, tex.sx provided me with a couple of things I needed for this. Bruno Le Floch wrote some code that rotates arguments at a random angle. Martin Scharrer ...
Why do I have the impression this letter is written by The Joker?
If I add a comment "Why so serious?", it will probably be flagged.
@PauloCereda You don't have enough reputation to read the deleted answers, but one of them was this: stackoverflow.com/questions/1732348/… which is a masterpiece.
could be closed, since the OP decided not to pursuit his/her intent. However, I didn't found an appropriate reason to close it in the list of possible reasons to close a question.
I flagged, but should I (we) also vote to close it?
@PauloCereda But it's really funny (and basically true.) HTML is a context free language (I think) and good luck processing it with a regular expressions. I think the answerer had answered one too many questions of "how do I process HTML with a regex" and decided to pretend that he snapped.
@AlanMunn It's funny indeed. And HTML is really a context free language, so regex won't be able to process it, only the "regular" parts, forget about syntactic nesting. :-)
(This is intended as a consolidated reference for one of the "frequently asked questions" at tex.sx.)
Some document elements (e.g., figures in the book class) are numbered per chapter (figure 1.1, 1.2, 2.1, ...). How can I achieve continuous numbering (figure 1, 2, 3, ...)?
And vice versa: Some...
I wonder how to use \textsc globally in WARSAW theme. Thanks in advance.
