@PiFisher a .... long time ago, when Alan was writing it:-)

@PiFisher have you seen tug.org/TUGboat/tb11-2/tb28jeffrey.pdf

@Dav

@DavidCarlisle Yes, I started by reading that.

I suspect my issue is not really understanding the differences between \def and \newcommand, but it might go deeper. Alan wrote that so he could order numbers he already had, whereas I'm trying to compute numbers. To do this, I need to be able to represent arbitrary non-negative integers, so I decided to represent an integer by a list with that many elements.

@PiFisher and?

(I started by writing everything I need for this project in Scheme, hoping it'd be fairly straightforward to go from one Lambda Calculus implementation to another.)

looks like you are missing some braces

The first line shouldn't need to be changed, right?

Or should Increment and Decrement also get more braces?

in \Increment#1 you are only passing the first item of the list #1 to \Increment not the whole list, you want \Increment{#1} and similar elsewhere.

@PiFisher it's like going \fbox abc instead of \fbox{abc} it just boxes a

@PiFisher IsEmpty and Tail the same... and Cons (for teh second argument)

@PiFisher run some examples with \tracingonline=1 \tracingmacros=1 and you will see what gets passed as argument to each macro call.

How do \tracingonline=1 and \tracingmacros=1 work? I don't have a local installation of TeX, because I've been working from my Chromebook with v2.overleaf.com.

@PiFisher it puts stuff in the log (and because of tracingonline also the terminal) not sure overleaf really has terminal output but you can presumably see the log file

but it's 2am and time for bed:-)

Sleep well, and thanks for the pointers.

@UlrikeFischer Got your mail. This should be easy to add, I'll take a closer look tonight.

@samcarter Thanks ;-)

I HATE EMPTY POCKETS

Bring David Tennant back

@JosephWright ^^

Greetings. Is there way to label "[]" on the arrows in tikz-cd? I tried \left[ \right], but it works only on nodes but not for arrows.

@PauloCereda I've not seen it yet ...

I found a way. Never mind.

@PauloCereda I actually saw it last night, I think last time I watched it it was the bloke with the scarf. But you can practice your Yorkshire accent now, useful for MP sketches:-)

@samcarter the tape for the duck-gift-paper

@UlrikeFischer :) Nice combination - you don't actually need a gift in the box, nobody will dare to tear through the paper :)

Is this a good place to get feedback and discuss implementation for a package?

@DavidPurton I've made very good experience with posting the first draft of a package to codereview.stackexchange.com I got very helpful feedback.

@samcarter @DavidPurton It also depends on the size of the package; if it's big, perhaps it's better to show only the parts to be discussed.

@samcarter, I have had a go at implementing CMYK shadings for PGF which comes up from time to time on this site. Is that too specialised for a general site? I wondered if my ideas would benefit from having someone who actually understand PGF look at it (as opposed to me who is really just guessing). I can successfully deal with all relevant questions I have tried from this site.

The code is not that big. It's just basically adjusting a variety of existing PGF macros. But I'm not really sure if my overall strategy is good or not.

github.com/dcpurton/pgf-cmykshadings

@DavidPurton Give Code Review a try; post here the link.

@DavidPurton Ohhhhh, that is very good idea for a package! This problems occurs quite regularly with beamer. I just wonder if "someone who actually understand PGF" actually exist?

@samcarter, that recent beamer question (tex.stackexchange.com/q/453336/87678) that you commented on prompted me to have a go. It works nicely in that case.

@DavidPurton That is very good to hear! I have some examples of related problems saved. I can test them with your cool package tonight and report back.

@samcarter you never have problems like that with picture mode

@samcarter, you'll have to load it before beamer so that internal shadings are defined correctly, but you should get the right output if you do \RequirePackage{pgf-cmykshadings} before you load beamer. I guess you might want to pass cmyk to xcolor as well to ensure solid colours are always CMYK.

@DavidCarlisle picture mode is bad for the pharmaceutical industry:

@DavidPurton Ok, I'll keep that in mind.

The code review link is at codereview.stackexchange.com/questions/205145/…

@samcarter @DavidPurton while it is certainly good if shadings works also with cmyk. Why it is needed for beamer? who cares about cmyk for a document meant for the screen?

@UlrikeFischer, yes this is true. That beamer question just happened to bring up the topic again for me. I have previously used the pgfplots work around in print contexts and that method is a right pain in the neck to use. It just prompted me to look at the code again and see if it could be done without too much pain.

@UlrikeFischer The problem is that users use colours defined in cmyk and then complain that the shadings are messed up :)

Perhaps @ChristianFeuersänger might be interested in taking a look at this too, since he is the source of most of the pgfplots workarounds to this problem.

May be a bit of a tanget, but speaking of "load order" . . .

@SapharKoshet you were going to add something here?

Waiting on me? Ok. Meta Question on "load sequence." New "Answer" is off topic. Can't propose an edit. Could "flag" it, but don't know what that entails. Any suggestions?

@samcarter hmmm, I don't think this is right. There is no problem loading \documentclass[xcolor=dvipsnames]{beamer} after pgf-cmykshadings. But it definitely doesn't work if you load pgf-cmykshadings after beamer. Some (at least) shadings are already defined as RGB and will only change to CMYK if you subsequently redefine them.

@SapharKoshet I remembered the meta question after making the comment. still the pgf shading question isn't about loading order so a load-order tag wouldn't be suitable there

Right! I see the difference in "tagging" Qs vs As. There's a "new" "Answer" to that Q&A. It addresses proposed edits on a different topic (incompatibility). It's not a "load order" "Answer." I don't think the "Answer" belongs in that thread. Getting some perspective on how to approach correcting that.

@SapharKoshet sorry I don't understand your comment. by There's a "new" "Answer" to that Q&A. which post are you referring to?

@DavidPurton Please try the following MWE

\RequirePackage{pgf-cmykshadings}
\documentclass[xcolor={dvipsnames,cmyk}]{beamer}

\usetheme{Frankfurt}

\setbeamercolor{section in head/foot}{fg=OliveGreen}

\begin{document}
\begin{frame}
\frametitle{Frame title}
Some text
\end{frame}
\end{document}

@samcarter. Oh I take that back. The colours don't get defined. This is one disadvantage of my method. You'd have to load xcolor before pgf-cmykshadings too

@DavidCarlisle The Q&A at (La)TeX Meta. As "asked," "load sequence." As suggested by Wright, "load order."

@DavidPurton Oh you are right. Loading xcolor before should work - I did not think of that. Sorry!

@SapharKoshet As Joseph said, anyone can tag a question with load-order if the question is about that, but I don't see many questions about that anyway so I wouldn't see the tag being used much.

Maybe I should pick this up another time. In this particular instance, I'm not talking about "tagging." I'm talking about that particular Q&A and the new "Answer" to it, which new "Answer" is off-topic.

@SapharKoshet it doesn't look off topic to me. It exactly answers your question why there isn't such a tag

Well, it's about the "incompatibility" tag, not the "load order" tag, but, if that's sufficient, then it should remain. Thanks for having a minute for this.

@DavidPurton Upvoted and starred, so other chat residents can have a look at it.

@SapharKoshet no, it's not. there is no incompatibility tag mentioned, it is explaining that load order (or load sequence as it's called in the question) is usually a feature of answers not questions so not much use as a tag.

@samcarter, this is a weakness of my implementation. You can produce RGB or CMYK gradients (even in same document). But one has to be the default. It isn't smart enough to choose on the basis of what colours you pass to the shading definition. Perhaps the choice could be made on the basis of the current xcolor model. But if no model is specified and the the shading is defined with a mixture of RGB and CMYK colours it would still need to pick one.

Even if the shading is all specified in one colour space, the way the current macros work the colours are not parsed until after the colour space is chosen.

The proposed (and rejected) edits are 100% regarding the "incompatibility" tag. I was proposing edits there to learn about how this "tagging" system works, and I've learned that it has to do with the Q, not the A. In the "load order" issue, the A is what raises the "load order" matter. Q may not know to ask about it.

@DavidPurton I just thought this too ;-). Imho if a color model is specified it should use it, and if not use the one from the starting color ...

@SapharKoshet well obviously the links don't go to existing load-order tagged issues as there is currently no such tag, but the text of the answer makes it clear that the issue that tagging applies to questions, the links are just examples of rejected edits where a tag suggestion based on an answer has been refused, it doesn't really matter for the example which tag that was.

@UlrikeFischer, I think choosing the model based on what is specified would be easy to do. But doing so on the basis of the first colour specified would involve changing a lot more of the internal PGF shading parsing mechanism.

Very good, then. Thanks for having a minute.

@UlrikeFischer, and colours can be specified as grayscale too, which complicates things further.

@SapharKoshet in any case I think the idea that there is any generic load order issue and that there is any relation between answers mentioning load order between different sets of packages is misguided.

Easiest solution to the whole shading problem: just allow two colours, black and white and no shadings :)

wrong one:-)

@DavidCarlisle Might I suggest a review of the parallel thread into which our comments are interspersed. Maybe you might want to ask them about "load order" issues.

@DavidCarlisle You want to quote the picture mode comment instead :)

@samcarter yes but hand to eye and trackpad coordination failed :(

@DavidPurton yes. Well in any case imho a shading between mycolor(RGB) to myothercolor(cmyk) should do something more or less sensible without the user having to worry to much.

@DavidCarlisle Maybe a text based browser would help and eliminate the trackpad problem

@SapharKoshet that is a good example of what I mean, any issues in that thread are completely unrelated to (say) issues about loading hyperref before cite or whatever

@samcarter I do miss emacs w3-mode

@DavidCarlisle My preliminary study indicates a number of "load order" issues that have nothing to do with hyperref. I get the impression that this parallel thread is discussing one such non-hyperref matter "as we speak."

@DavidCarlisle If this is of interest to you, maybe you can suggest where I could upload for your perusal what I've found so far.

@DavidCarlisle Would also solve the problem with the "responsive design" of the main site. After fixing some problems at the beginning, I they now just wait until all users of the old design are death and nobody remembers what a working site looked like

@UlrikeFischer, that option isn't supported by standard PGF at all. My code will at least output a CMYK shading.

@SapharKoshet The "parallel thread" is discussing more a problem within a package than loading order problems.

@SapharKoshet what list of package load order recommendations? I don't know you could ask a "what order should I load packages" question on the main site then self answer with a list I suppose. But as I say grouping these things as "load order issues" rather than issues with the individual packages, rather misses the point I think.

@samcarter Thanks for having a minute. Is there an issue worth marking as such, do you think, obviously not via a "tag," as a "load order" issue?

@samcarter which is always the case.

@DavidCarlisle there is a question "Add a plane to a tikzpicture ..." on the main site ...

@UlrikeFischer you should close it as a duplicate of my plane.

@DavidCarlisle Ok. I'm looking solely for a way to identify the issue in its most appropriate form. I think a lot of time can be saved where the typesetter using this or that package is aware of a "load order' issue. "How" that is addressed is pretty much what I'm looking for guidance regarding. If more specifics are in order, great. What do you suggest? Something more package specific?

@SapharKoshet No, there is nothing to be marked as "load order" issue.

@samcarter I'm more than open for the rest of the story, here.

@SapharKoshet there is no single "issue". Some packages may make conflicting definitions and in some cases perhaps adjusting the load order may help, but lumping that with discussion of two unrelated packages that happen to have their own conflict doesn't really help anyone.

@DavidCarlisle I see exactly what you're saying, and I agree with you. So, maybe specific package-related "comments" would work or suffice?

@SapharKoshet Which rest of the story are you referring to? There is no rest other than discussion about pgf shadings.

@SapharKoshet that's what happens now, someone asks about hyperref and possibly package adjustment is part of an answer, but there is nothing special about that part the answer that needs to be treated in a special way. there is no use in anyone searching for all answers mentioning package order.

@samcarter I'm sorry to come into the middle of that with a different context. I think what I'm looking for is what Carlisle is helping me with.

@DavidCarlisle I see what you're saying, and I agree, that would be far too broad and, therefore, not at all helplful. So, something that's package specific is what we're talking about, and, perhaps even more to the point, is that something that should be posted at or with the CTAN information or in (La)TeX SX threads?

@SapharKoshet no, there are 1000s of packages on ctan and so 1000000s of possible package orders there is no general issue to solve, nor any general solution. If two packages define \foo then the end user has to decide (a) if it is wise to load both packages and (b) which order to load them, depending on which definition of \foo is required. So any conflict resolution is document specific.

@SapharKoshet there are thousands of answers, blog, tutorials handling various aspects of load orders already. Why do you believe whatever you want to add to the lot will solve all problems?

@UlrikeFischer Thanks for the question. I think what I'm looking for is what Carlisle is helping me get some perspective on.

@SapharKoshet which is not to do anything.

@DavidCarlisle Ok. So, it's a SX matter that simply has to be addressed as it comes up in this or that particular instance. What I was finding was (considerably) more general in nature. I'll go reread those discussions and perhaps raise this with you again as the matter is reviewed further.

@SapharKoshet you can look, but it's hard to think there is anything to be said that hasn't already been said in the 30+years that latex has been around.

@DavidCarlisle :) :) I can assure you, as one of the newest members, that it's not so much a matter of "it's having being said" as it is "finding what's been said" in time for it to matter. :)

@DavidCarlisle And, there are the persistent issues not only with "new" packages but also with changes to "existing" packages that don't necessarily allow for "backwards compatibility."

@SapharKoshet in practice I would say the most common reason people get package conflicts is that they copy and copy and accumulate junk and end up with 500 line preambles loading 400 packages they don't use so "don't load packages you don't use" is by far the best advice to give and will solve the mast majority of problems in actual documents.

@DavidCarlisle :) I've seen a number of those MWEs, and I have an understanding of exactly what you're saying. The "load order" matter, then, beyond hyperref, is something that's really too document-specific to know how to address in a "general setting," and needs, then, to be left to that specific case, is what I'm learning.

@DavidCarlisle Ok. One more thought down this path. Is it worth generating a collection of those types of matters into one particular location?

@SapharKoshet as I said just now, no I don't think there is anything in common

@DavidCarlisle You've provided me the perspective on this matter that I've needed. I'll keep my notes on these matters in the event they are relevant to a document I'd generate, and focus on a new project. I really do appreciate you're having a minute. It matters.

@JosephWright looking at the l3build code I guess there is no sane way to run tests with luatex53 --fmt=lualatex53 instead of lualatex?

@UlrikeFischer you can add luatex53 to checkengines, but then it wouldn't know it was luatex and do the extra luatex normalisation (which might not matter, depending) , or you can temporarily install luatex53 as luatex. I think I've done both in the past

@DavidCarlisle but it I add luatex53 and it works I would still not have the right format. I would have first to create a lualatex format with the engine.

@UlrikeFischer ah, I forgot: i was doing this in latex base so it was making a format anyway as part of the test build

@UlrikeFischer but if you have made such a format, can't you just add it to checkengines? or is the issue you need a lualatex53 command not luatex53 --fmt=lualatex53 ?

@UlrikeFischer I'd have to code it in by hand I guess: something similar to the JIT version

@DavidCarlisle That I have no lualatex53.exe is an additional problem. My first is that even if I had the exe I don't see how to get an engine -eluatex53 option to map to a "lualatex53" call.

@UlrikeFischer code is

    not match(engine, checkformat) then
    format = " -fmt=" .. gsub(engine, "(.*)tex$", "%1") .. checkformat

@UlrikeFischer ooh

@DavidCarlisle 4th doctor?!

so I gues choice is set lualatex53 as the checkengine (and arrange there is a command of that name) or modify that gsub so it doesn't assume it can do the right thing just by stripping tex from the end of the engine name

@PauloCereda waasn't counting. Possibly. Did I miss anything?

@DavidCarlisle Nah they rewrite story all the time. :)

@DavidCarlisle hm. If I call the format luatex53duck, shouldn't then engine=luatex53 and checkformat="duck" work?

@DavidCarlisle, I got enough functions working that I can no longer keep using Overleaf, as the computations are too slow on their servers and I keep timing out. Hopefully this doesn't bode too poorly for when I finally try to solve the Discrete Logarithm Problem on my wife's desktop in LaTeX.

Maybe people who just post images and ask for code should get images of the code as answers. :)

seems sensible to me

@AlanMunn ^^ :)

@PauloCereda :)

@JosephWright@UlrikeFischer Very likely next texlive will have "luatex" with lua 5.3 , and perhaps "luatex54" with lua 5.4 perhaps l3build shoudl have some permanent flag for testing with (jit)luatex(nn) if variants are going to be a permanent feature.

@PauloCereda disaster! what does that do to these statistics?

@DavidCarlisle ooh statistics

@PauloCereda "statistics" sounds posher than "adding"

@DavidCarlisle Yes looks like a good idea ;-). (I'm not sure if jit testing is so needed, Hans seems not like it ...).

@UlrikeFischer I was trying to test the ffi interface earlier in the year and it still seemed better with the jit version although it was supposed to have been added to the non jit as well.

@DavidCarlisle ooh

@UlrikeFischer is it clear why adjustbox loads pgf in luatex?

Oh, that should be \babylist{#1}{\n}{#3}

@PiFisher there is no scope associated with macro expansion so that is only local if contained in some group. (but doing non-expandable assignments seems at odds with using lambda.sty)

@PiFisher also \let\n{ defines \n to be { I think you meant \def not \let here.

@UlrikeFischer that's what I was thinking, most of the relevant things that pdftex can do luatex can do as well, so it seemed a surprise to use a sledgehammer like loading pgf.

@DavidCarlisle but as any document with a duck needs tikz anyway ... ;-)

@UlrikeFischer only because @samcarter still hasn't updated that package to use my picture mode ducks.

@DavidCarlisle Hmm, I guess

@PiFisher then \edef probably (if the things in the arguments are safe in an expansion context) certainly you don't want \let as I say.

I haven't heard of \edef before. If it stands for 'expanded define', or something like that, I guess it's exactly what I want.

@PiFisher that is what it stands for and all lambda.sty expressions work by expansion so would be safe in an edef, but your \DLP uses assignments so would not itself be safe to use in an \edef.

@PiFisher lots of answers about that on site eg:

nice diff message ;-) @DavidCarlisle ^^^^

@DavidPurton I did some test and the cmykshadings work really well! To safe you from other users doing the same mistake like me and complaining about undefined colours, how about adding a sentence to your readme file that your package should be loaded after xcolor if special xcolor options are needed?

Good evening to all user.

Excuse me for this question.

@samcarter, good to hear. I'll update README.md to note that. I think I'll also change it so that if the xcolor model is RGB then an RGB shading will be used, otherwise it will use a CMYK one.

For this question I have received 2 downvotes and I have not understand this. In this period I am very busy at school and I don't connect very often. I am inexplicably getting many negative votes on old questions and answers.

@DavidPurton Welcome and good evening.

@marmot Hola'. How are you? Again I have the same problem with downvotes.

@DavidCarlisle Hi, how are you?

@Sebastiano seems a perfectly reasonable answer to me.

@DavidCarlisle I thought about you at school this morning. An English teacher told me that when she wishes everyone a good evening, she doesn't need a "to". Is it true?

@Sebastiano where would you add "to"?

@Sebastiano that looks pretty unreasonable you can ask the staff to investigate.

@DavidCarlisle Is correct in English language: good evening to everybody?

@Sebastiano probably not (but you should ask @AlanMunn about grammar, I'm English so have no idea:-), I might use "to" in a phrase such as "he wished good evening to everybody" but then "to" goes with the verb "wished", but if i was actually giving the greeting I'd just say good evening everybody, or more liklely just good evening (but Alan will probably laugh at my analysis:-)

@DavidCarlisle I have understand :) mille grazie.

@Sebastiano Please do not raise chat flags for issues with votes on a site. Raise a custom moderator flag on one of your posts on the site itself.

@DavidCarlisle Hahahahaha. (Just meeting your expectations). @Sebastiano I agree with David (and the analysis). In the context of greeting people here, the 'everyone' is vocative, so you don't need the preposition.

@ACuriousMind Hi, where i can put the flags? Into my answers and questions? Thanks

@Sebastiano if you flag anything in chat you will invite anyone with >10K rep on the entire network to this room to sort out the issue, that's rarely a good idea....

@Sebastiano don't flag at all there is a contact email to raise with teh staff

@AlanMunn "vocative", see I just thought it didn't sound right. You sound like my sister.

@Sebastiano Below any of your questions/answers there's a flag link. Use that and choose "custom" in the menu that appears, then explain the issue there. Alternatively, you can use the "Contact us" link in the footer of the site

@DavidCarlisle Where is the link for the contact email?

@Sebastiano If you want to ask a mod something, just come into chat and ping @JosephWright or @StefanKottwitz. They are in chat regularly and will respond to you.

@Sebastiano but as we have told you before site mods can do nothing here they don't have a lot more information than you, you need to either let it go or raise with the staff.

@AlanMunn Hi, and thank you very much. I trying it.

@DavidCarlisle Don't confuse naming things with explanation. :)

I leave the chat to perform the operation and greet you all with esteem and affection. Have a nice rest.

@Sebastiano We worked on the issue last time weeks ago and it should have stopped at least, for a while.

@Sebastiano We will have to see if there is something different now. We have limited view and limited options, but could moderate somehow.

@Sebastiano It may be that we are on our own. If you think, that the site owners are interested in TeX or are interested in user experience and user opinions, let me know. ;-)

Hey all!

Is there a math operator for the k index? i.e. we have $\imath$ and $\jmath$ but no $\kmath$. How could we define something like that?

I know k is not dotted

Oh, I see this fantastic manual‌​, page 10 point 7., and there says "There is no such thing as \kmath since k is not dotted"

@Sebastiano Sorry to hear that. But it seems that @StefanKottwitz is looking at this. (You probably know that these sites are a big sociology experiment gone bad, so not much one can do ;-)

@manooooh $k$ :-)

@DavidCarlisle $\check k$ :)

@StefanKottwitz Please do not take away my illusion. It would be very hard for me to believe that some users, though these are very few, are really as nasty as they appear to be. So I think of those as being paid for doing this in the experiment.;-)

@StefanKottwitz But thanks for the link!

@marmot It's life. It's the Internet.

@marmot Competition brings good and bad.

@StefanKottwitz I can logically understand why some people show scientific dishonesty in real life: they think that stealing ideas from others will give them a better job, higher salary etc. And yes, there are (even famous) cases in which this works. But here it is just the meaningless number: reputation score. Even more, those people hurt themselves because others realize it and do not upvote them that much. Really don't understand.

Does anyone know how to change the dashed line between the upper and lower parts of a tcolorbox to something not dashed?

@AlanMunn yes

@AlanMunn An answer for a linguist: "Does anyone know..." -- yes, for example the author would know ;-)

@marmot I had enough faith that he wouldn't require explanatory text on the answer:-)

@marmot haha

@DavidCarlisle Part of linguistics is pragmatics. If I were a logician, then maybe this would be an appropriate response.

@DavidCarlisle Yes, @AlanMunn is now blessed with two very helpful answers. ;-)

@AlanMunn I guess change the numbers here:

@marmot Hard to say if you cannot read all but just see a specific flag for mod attention. I dealt with 970 flags in the last month. Most are easy and maintainance though.

% draw segmentation line
\def\tcb@drawlower@standard{%
  \begin{pgfscope}%
  \pgfsetdash{{1mm}{1mm}}{0.5mm}%

@DavidCarlisle Yeah, I found that too. But that seems to imply it's basically hard coded.

@AlanMunn nothing is that hard coded once it's in your emacs buffer

@AlanMunn see /tcb/segmentation style= in the docu.

@UlrikeFischer Yes, I've tried that but it seems to have no effect on the standard line, just on the \tcbline.

@UlrikeFischer which is presumably this pointing at same macro

  segmentation engine/standard/.style={segmentation code=\tcb@drawlower@standard},

so it comes to the same thing, you need to provide an alternative definition of that whole thing

@AlanMunn are you using an enhanced skin?

@UlrikeFischer Not really. I've loaded skins to gain access to this key, but otherwise no.

@UlrikeFischer I guess I'll use a workaround with just a \tcbline instead of \tcblower.

@AlanMunn then try with enhanced. I can't look now.

@UlrikeFischer Ok. Thanks. No worries, it's not that important at the moment. It just seemed like an obvious thing to want to change.

@UlrikeFischer Ah, that worked. Thanks a lot.

@DavidCarlisle Does anything look obviously wrong with this?

\def\IsEven#1%
  {\Isempty{#1}%
    {\True}%
    {\edef\oneLess{\Decrement{#1}}%
    \Isempty{\oneLess}%
      {\False}%
      {\IsEven{\Decrement{\oneLess}}}}}

It works if I get rid of the \edef line and replace all instances of \oneless with \Decrement{#1}.

@PiFisher as I said before lambda.sty is all about using expandable definitions so having an \edef there looks rather suspect,

@PiFisher as expected, although I didn't run it, just looked odd.

@DavidCarlisle Okay. I guess I'll just stick to keeping everything expanded, at the cost of computation time.

@PiFisher what's the main use case, do you actually need expansion-only definitions, Alan's code was mostly a thought experiment on what is possible. In some cases it is necessary to work by pure expansion but usually that isn't the case.

@DavidCarlisle my use case is trying to solve the Discrete Logarithm Problem. For homework, if we write code, we're supposed to turn in a copy of the code in addition to the result, and I thought it would be cute to turn in LaTeX source along with the PDF it created. Unfortunately, I'm having trouble getting \Multiply{\Three}{\Six} to compile in a reasonable time, but I'd still like to get the macros to be correct.

@PiFisher yes but why do you have to start from an abstract model of the integers as lists in purely expandable macros rather than use tex numbers when you can use \numexpr 3*6 to multiply and \ifodd to test if something is even or odd?

@DavidCarlisle Because I started this before knowing that \numexpr 3*6 existed, and now that I've seen you write that, I think I'm in too deep to restart. :P

Before I found Alan's paper, I had already written a few operations that took counter names as inputs and changed the values of those counters. The algorithm I have for solving DLP involves creating two lists, finding a shared element, and computing the index in each list, which is where I got stuck previously. I found lambda.sty while hoping to find an implementation of arrays.

@PiFisher it's an interesting mathematical exercise, but most people assume the integers exist, they don't start from nothing and work through 400 pages of Whitehead and Russell to get to 1+1 =2.

Most people don't have to spend 400 pages to convince themselves that Add({{}},{{}}) is the same as {{},{{}}}?

What do most people do with their lives?

@PiFisher They use Word, so I'm told

That made me laugh.

I'll probably try to work with built-in numbers next time I do something silly like this, but I think I've already dug far enough with this project to turn back now.

I've successfully determined that $2^3\pmod{3}$ is 2.

And $2^4\pmod{3}=1$! Here I was expecting a timeout error.

Does anyone know if $\im$ mean "Image"?

@manooooh what's the context?

@PiFisher linear transformations/map :)

I know the $\ker$ command, and for the image... $\im$?

Seems like a reasonable interpretation. If there's a symbol you've seen used for this and you're trying to mimic it, I'd recommend looking at detexify.kirelabs.org

I've also seen simply $Im(T)$ for the image of $T$.

@PiFisher yes yes I already knew it! Thank you

@PiFisher yep... but I read a topic in TeX.SE that says "why the kern is written in lowercase and Image in uppercase?"

That makes sense for me... don't you?