I asked a question and it got a few solid answers. They all are good. Am I supposed to accept the one I like the most?

@Kurzd -- the one that is most helpful, that gives you the best answer is the one you should accept. if they are all equally appropriate, then the one that gives the best explanation of how it works is probably the one you should choose.

@barbarabeeton It has been accepted. Thanks.

biblatex.def is gone, you cannot look into biblatex1.sty ... everything changed.

@Johannes_B It's biblatex_legacy.sty, as far as I can see

@egreg Jep, but why? It was working so nicely.

@Johannes_B Blame @JosephWright

@egreg Really? I thought PLK made that decision. :-/

@Johannes_B I always agree with @egreg in such matters.

@Johannes_B Blaming @JosephWright is easier

@Johannes_B , @egreg I'm surprised as well, though I understand the decision

Probably I should look to work out how to do everything in one file with conditionals, rather than the 'quick fix' we went for when PLK wanted to split off Biber features

@Johannes_B That's all internal, of course (though there are issues, as @UlrikeFischer has noted)

@JosephWright Looking up the definition of subbibintoc is not relly an internal thing, in my opinion. The contents of biblatex.def are also needed for style authors as a basis. Reading the manual alone won't help much. I am with @Ulrike on that one, we need a better access to some features.

@Johannes_B I meant the fact it's called biblatex.def, not what's in it

@JosephWright Sorry, i don't understand.

@Johannes_B As I said, with hindsight I probably should have worked out up-front how to avoid having two sets fo files

@Johannes_B You said 'biblatex.def is gone, you cannot look into biblatex1.sty ... everything changed.', which is about file names

@JosephWright Ah.

@JosephWright Yes, my personal workflow broken.

@Johannes_B What I meant there is that only files which are documented as being loadable should necessarily have stable names. So standard.bbx as a style is important, biblatex.sty itself, yes they have to exist, but subsidiary files can be renamed

I'm not sure I understand things, but what is the motivation for still supporting/considering bibtex as a backend?

@yo' I think bibtex support will be dropped sooner or later. Rather sooner i think.

@yo' It works pretty well

@Johannes_B I think that's where PLK is going, but I'm not sure PL would have done that

@yo' BibTeX is fine for a lot of use cases and is much faster than Biber

@JosephWright Obviously it does not if it causes such headaches... :)

@JosephWright Has anybody heard from him in the meantime?

@yo' The fact new features are Biber-only is tricky, like I say with hindsight I should have suggested a different approach to that area

@JosephWright The speed is key. biber --version takes a lot of time on my machine.

@JosephWright ok, speed can be a point, OTOH, will all this TikZ, expl3 luatex-microtype etc...

$ time biber --version
biber version: 2.4

real	0m5.891s
user	0m1.800s
sys	0m0.140s

For someone like me doing physical sciences work, BibTeX is fast and does basically everything I need (it can't do dynamic sets, but I use those rarely and still think that's a solvable problem). I use biblatex mainly as 'tweaks in the document' are doable and I can access parts of the data model, not because I need any Biber features.

@Johannes_B sounds ridiculous. Why's that? Is it poorly coded or really that complicated?

biblatex had a good range of ideas going on before Biber was even written

@yo' Don't know. In addition, my system is buggy. Need to do a fresh install.

@yo' My understanding is it's the usual issue that you can do 90% of things quickly but a proper general solution is much harder. BibTeX is fast precisely because it can't do much outside of ASCII

@Johannes_B do you get that the second time? On my machine not having run it for ages I get:

$ time biber --version
biber version: 2.3

real    0m13.613s
user    0m4.343s
sys     0m4.014s


$ time biber --version
biber version: 2.3

real    0m1.166s
user    0m0.765s
sys     0m0.280s

@JosephWright yeah, and some conceptual burdens...

Note it's much quicker after the first use.

Anyone heard of this problem before:

$ time biber --version
biber version: 2.4

real	0m1.859s
user	0m1.724s
sys	0m0.092s

@DavidCarlisle Still pretty slow ^^^

apparently the wget from TL on windows will not accept the certificate for tug.org.

@yo' I'm no Perl programmer, but PLK has been asked about performance many times and says that it's all to do with properly handling all of Unicode (and the recoding/sorting/... this entails)

$ time bc --version
bc 1.06.95
Copyright 1991-1994, 1997, 1998, 2000, 2004, 2006 Free Software Foundation, Inc.

real	0m0.094s
user	0m0.000s
sys	0m0.004s
[33594][boettchj.Pearl: testing_tex]$ time man --version
man 2.6.7.1

real	0m0.500s
user	0m0.004s
sys	0m0.008s

@JosephWright not clear if perl is really the best choice especially for cross platform windows use where it means you have to ship perl:-)

@DavidCarlisle I heard Lua is available. How hard can it be for an experienced Lua programmer to port biber?

@DavidCarlisle PLK has a Perl background and wrote in what he knows. Given the fact there is no obvious alternative I'm not sure you can really criticise.

@JosephWright ok

@Johannes_B although generally speaking lua would be slower than perl I'd guess

@Johannes_B No good (or at least a lot of work): Lua doesn't ship with any (real) Unicode libraries

@JosephWright @DavidCarlisle Argh. :-(

@JosephWright true but...

@DavidCarlisle agreed. IMHO the leader in UC handling is python, but who knows...

@Johannes_B lua-users.org/wiki/LuaUnicode

@JosephWright yes, but becomes more of an issue if biblatex moves towards making bibtex harder to access

@yo' That would not be easier: TL has shipped for many years with an internal Perl set up, but Python always seems to need a stand-alone installation when I see it (for example various science-related GUIs)

@DavidCarlisle Hopefully that won't happen: I think I need to talk to PLK and see if I can suggest a better approach

@JosephWright If you wanted to beat perl for speed you'd almost certainly need C (perl is blindingly fast, really)

Probably loads of work for me

@JosephWright makes a change form luatex, and you need something to avoid chemistry:-)

@DavidCarlisle Like I said, I'm no Perl programmer so have no idea where the hold-ups are in Biber cf. BibTeX beyond 'it can do a lot more stuff'

@DavidCarlisle :-)

@DavidCarlisle Speaking of LuaTeX, do you understand Hans' reply on the list?

@JosephWright You can create executables for any system/platform, it's jist close to impossible on a machine with a different one (one of my uses of virtual box is creating win executables of python programs)

@JosephWright other than thinking "Frank will need to get used to typical Hans replies", no:-)

@yo' Looking at github.com/plk/biber/blob/master/lib/Biber.pm, there are about 3k lines to Biber itself, but a lot of libraries loaded which for a Lua version you'd need to recode. (I have been wondering about this as for an L3 implementation of similar stuff we presumably will want everything self-contained)

@JosephWright That's a bold plan...

@JosephWright How about xindy?

@yo' Yes but at some stage we will need something in this area, doing it using BibTeX will be pretty hard to justify, doing in all in macros is very hard, and using Biber itself is probably a non-starter as we are bound to want some variation in the data model, etc.

@Johannes_B Same thing (I do hope Unicode support in Lua(TeX) improves before this happens)

@JosephWright I keep thinking that for some stuff, an asynchronous approach would work, but that's a completely different thing.

@all I want to add a proper biblatex page to the wikibook. What should be mentioned and what should be left untouched?

@Johannes_B just say it's all broken and it's all Joseph's fault:-)

@DavidCarlisle :-)

@Johannes_B mention that to date, most journals do not accept biblatex

@DavidCarlisle Frank is getting the hang of it

@yo' Already done in the What to choose part on top ;-)

@Johannes_B :)

@JosephWright Fragmentation is surely a complicated challenge. :(

@JosephWright tug.org/svn/texlive/trunk/Build/source/texk/web2c/luatexdir/…

@DavidCarlisle Will watch out for a new LuaTeX build

@JosephWright I'm building from there now....

@JosephWright although Frank may yet get a 0.89.4 :-)

How many For more info consult the package manual entries are allowed on a wikibook page? :-)

@JosephWright Is there any plan for printf? Maybe I should ask someone else, right? I mean something like \str_printf:Nnw \l_new_str { %05d ~ %10s } { \l_my_int } { \l_my_str }

Or at least, padding integer with zeros to a given length

@yo' Bruno's done some experiments

@JosephWright ok, good to hear :)

@yo' We end up with questions about the aim: formatting numbers tends to need siunitx or similar

@JosephWright well, in my case, it's doi:10.1111/AP.53.01.0871, so no siunitx involved :)

gotta go, see you later

@DavidCarlisle: still harvesting points from the grandma answer. :)

@PauloCereda me too

@yo' Erm, not sure I see why there's any foramtting here

@JosephWright hmm my texlive build doesn't build anymore:(

@JosephWright well, we want the last three numbers to be padded by zeros...

@yo' I guess Joseph's point is that the DOI ought to have been correctly padded on input they aren't really numbers you have calculated are they?

@DavidCarlisle they are; I generate the doi

Am I the only one who does not have any clue on what Hans and Frank are talking about in the LuaTeX list? :)

@PauloCereda No: I suspect the people who understand are Frank, Hans, @DavidCarlisle and me ;-)

@JosephWright ooh :)

@Joseph: @David is probably the culprit. :)

@yo' back to \ifnum#1<10 0\fi\ifnum#1<100 0\fi\ifnum#1<1000 0\fi#1 then:-)

@DavidCarlisle yeah

@JosephWright how many people do you think have ever used luatex to set a paragraph with non zero looseness, and how many of those were using a non standard linebreaking callback?

@DavidCarlisle A very small number

@JosephWright that was my guess too.

Is there an equivalent to \pdfstrcmp in LuaTeX? (See tex.stackexchange.com/questions/298236/…)

@TorbjørnT. heiko has a package with an emulation in lua as does expl3 I guess (@JosephWright knows these things)

@TorbjørnT. \str_if_eq:nn or something like that

@TorbjørnT. Yes, done in Lua: load e.g. pdftexcmds or expl3, or add a version yourself

Heiko is a TeX demigod.

@PauloCereda although like many demi and full gods he's somewhat absent at present.

@DavidCarlisle Yes, but watching everything very carefully (and probably eating popcorn).

@PauloCereda minor issues like ifpdf and hyperref, not working in texlive 2016 luatex might need more than popcorn

@DavidCarlisle ooh some Black Forest cookies then! Where's @ChristianHupfer? :)

@JosephWright still trying to compile texlive (did a clean rsync of the svn head and full clean build, but that takes an hour or two....)

\def\strcmp#1#2{%
  \directlua{
    local A = "\luaescapestring{\detokenize\expandafter{\expanded{#1}}}"
    local B = "\luaescapestring{\detokenize\expandafter{\expanded{#2}}}"
    if A == B then
      tex.write("0")
    elseif A < B then
      tex.write("-1")
    else
      tex.write("1")
    end
  }%
}

@TorbjørnT. ^^^

@JosephWright @yo' @DavidCarlisle Thanks!

@JosephWright i stopped the printing I wonder what kind of phone call that was :-)

@JosephWright \usepackage{pdftexcmds} and \pdf@strcmp

@egreg Yes, I did say that too (I was just posting a package-free version)

@JosephWright @UlrikeFischer Do you think we need a question that deals with the change of the nameformat? Like we had when El Capitan was released and hell break loose.

@Johannes_B Might be a plan

@UlrikeFischer texwelt.de/wissen/fragen/16629/namensformat-mit-biblatex-33

@Johannes_B @JosephWright hell didn't get loosed yet. But it would be a good idea to think alreay about a suitable \ifbiblatexversionlesserthen{3.3}{}{}.

@UlrikeFischer I am in favour of that. :-)

@JosephWright phew texlive built:-)

@DavidCarlisle :-)

@JosephWright seems like "make clean" wasn't enough, but refetch everything to an empty directory and rebuild worked (but takes some hours:-)

C:\Users\Public\Documents\LaTeX\biblatex-chem>luatex --version
This is LuaTeX, Version beta-0.89.3 (TeX Live 2016/W32TeX)

@JosephWright you've got one too:-) should probably check in the remaining tlg and declare it passed now that has reached tl

@JosephWright I hope you get this:

$ luatex \\show\\alpha\\end
This is LuaTeX, Version beta-0.89.3 (TeX Live 2016)
 restricted system commands enabled.
> \alpha=\mathchar"010B.

C:\Users\Public\Documents\LaTeX\forks\biblatex\tex\latex\biblatex>luatex \show\a
lpha\end
This is LuaTeX, Version beta-0.89.3 (TeX Live 2016/W32TeX)
 restricted system commands enabled.
> \alpha=\mathchar"010B.
<*> \show\alpha
            \end
?

looks good

@DavidCarlisle Running tests here now

@yo' and @Johannes_B -- ams journals (and books) do not accept biblatex and won't be expected to in the foreseeable future; the declared alternative is amsrefs, which, sadly, is very much a minority option.

@barbarabeeton Speaking of journals, I'm about to find out how well chemistry journals handle LaTeX submissions: my first paper written that way (just me and my student) going in after referee comments

@JosephWright -- great gobs of good luck! let us know how it goes, please.

@barbarabeeton Once it gets accepted, I will (I got 'major revisions' though the changes requested were small)

@barbarabeeton @yo' My plan was to differentiate between manual and automatic and then BibTeX vs biblatex. From my vim-buffer: "If you are submitting to a journal, ask them for advice. Many journals provide you the style definition files for the older BibTeX workflow. If they don't use the modern approach, they won't be pleased to see biblatex."

@JosephWright -- well, that's encouraging (as far as the likelihood of acceptance, anyway).

@barbarabeeton Sure: I think we'll get in, just need to do all of the tedious stuff now (currently uploading all of the figures)

@barbarabeeton I finally added a warning sign on en.wikibooks.org/w/index.php?title=LaTeX/…. But i think i cannot most of the stuff that is on that page (i will relocate it) is so mangled up ... i don't know how to treat it.

@Johannes_B -- well, math references are helped out a lot by the fact that mathscinet will "provide" references formatted as either bibtex or amsrefs. (not likely to get biblatex there either.)

@JosephWright crossing fingers!

@Johannes_B -- i've put the page on my list of things to look at. not sure when i'll get to it, but i see the very prominent warnings at the top. and when i do get to it, i'll send my comments to you, okay?

@barbarabeeton Alright. I don't know how long i will need to rewrite the bibliography stuff. I hope @moewe can take a look at it.

I used `yellow!20!white! as the pagecolor when taking screenshots to show output. What do you think? Real wikibook example: en.wikibooks.org/wiki/LaTeX/Manually_Managing_References

looking for some advice here. @egreg has posted an excellent answer to a question regarding the relative positioning of sub/superscripts: tex.stackexchange.com/a/298295 ; i'd like to add it to the "math" section of "often referenced questions". there are two problems. one is that the question has been asked/answered before (e.g. tex.stackexchange.com/q/213674), and the other is that that "orq" section is large and unwieldy. please post suggestions.

Now that's a childish list... :) tex.stackexchange.com/q/298323/2693

@AlanMunn -- quite charming, really. but the images really need to be a lot larger to have full impact. probably one per page. (and produced as a fabric book, good for chewing, etc.)

@AlanMunn LOL

@AlanMunn what, no ducks?

@barbarabeeton Yes, it's true they don't look so good in this image.

@DavidCarlisle Trust me, I looked. But they're a niche compared to teddy bears...

@PauloCereda Thesis chapter headings ^^^ ?

@DavidCarlisle YES!

@AlanMunn A comment mentions circles in the question ... which circles?

@Johannes_B bullets for the itemize

@DavidCarlisle -- not as exciting as they might be. no variation except for the letter itself.

@barbarabeeton blame google that's what came up for "duck font"

@DavidCarlisle Oh.

Much better.

@AlanMunn YES YES YES

You guys are ruining my thesis, and I am loving it. <3

@PauloCereda by all accounts there isn't much yet to ruin, only one chapter:-)

@AlanMunn surely you could get this into an answer on the question

@DavidCarlisle Well thought. :)

If anybody wants to share a thought? en.wikibooks.org/wiki/Talk:LaTeX/…

@DavidCarlisle Done. :)

@AlanMunn luckily only one other person has already voted to close as opinion based:-)

@DavidCarlisle I didn't actually cast a vote on that. I thought my comment might get the OP to edit the question a bit.

@AlanMunn I would use snow flakes if a regular circle is too childish.

@Johannes_B Like these: tex.stackexchange.com/a/86334/2693

@AlanMunn Oh wow.

On Monday, September 30, 1996 at 12:00:00 AM UTC-7, Thomas Heim wrote: Wow, RE today. c.t.t. keeps surprising me.

@AlanMunn I did:-)

I afraid of Mico!

@EnthusiasticStudent Why?

@AlanMunn his profile photo :|

By watching his photo, I always think that he wants to eat the OP completely when he answers questions!

@EnthusiasticStudent Yes he runs a ring of bibliography style hackers among other things.

@barbarabeeton Not completely the same, but very similar.

@egreg -- okay then. should i add both to the list? i'm thinking that a reorganization of that long list might include some extra sub-categories -- coding matters, spacing and positioning, haven't come up with good "wrappers" for the rest.

By the way, can somebody tell me what the current biblatex version is?

@DavidCarlisle: picture mode to rescue: tex.stackexchange.com/questions/298350/how-to-draw-this-in-tikz

@Johannes_B 3.4

@Johannes_B Well in the github message I wrote "with older biblatex versions (v3.2) but fails with error in the newest version (v3.3)." And as I'm normally precise when I write such things ...

@UlrikeFischer The pdf states 3.4 that's why i am asking :-) See @JosephWright's post

biblatex.sty in texlive says \def\abx@version{3.3}

@Johannes_B @Johannes_B ^^^^

@UlrikeFischer I know. It was a trick question ;-)

by the way, wanna answer texwelt.de/wissen/fragen/16629/namensformat-mit-biblatex-33?

@barbarabeeton Wow, i am using the wikibook as the basis of an answer: latex-community.org/forum/viewtopic.php?p=93897#p93897 :-D

@Johannes_B -- well, you did check first to make sure it works properly. (never said it was all bad, just that it really needs a thorough housecleaning.)

@barbarabeeton Look at the version history ;-)

I try to use real words. I don't like Blah Blub examples.

@Johannes_B -- impressive. nice examples. that deserves an "attaboy".

@barbarabeeton Awww :-)

@JosephWright Frank's playing in our playground:-)

@PauloCereda This user tex.stackexchange.com/users/12065/waldir-leoncio reminds me of another Waldir that you might remember of. ;-)

@PauloCereda Strangely it makes me remember of a guy called Paolo Rossi. :)

@egreg Ouch. :)

@PauloCereda Beating a team with Sócrates, Falcão and Zico is not something you forget. :)

@PauloCereda On the other hand, Rossi played for Juve at the time.

@egreg :)

Is it legal to send someone the PDF of my TB article?

@yo' -- to quote the copyright statement (inside front cover of every issue): "Copyright to individual articles within this publication remains with their authors, so the articles may not be reproduced, distributed or translated without the authors'~permission." since you're the author, you have that control.

@barbarabeeton Thanks (I was just asked a copy by someone who: "I would like to be a member of TUG, but my budget won’t let me.")

@yo' -- understood. (it's a financial hardship for a lot of potential members, unfortunately.) make that person knows that all of tugboat becomes "open" after a year. (or, more precisely, when the third issue following the issue in question is ready for publication, the issue in question is "opened up".) and if a person who is not a member writes something for tugboat, that person automatically gets a paper copy of the issue, as well as the final files.

@barbarabeeton "Mail sent." :-( If I communicate with him further, I'll let him know!

btw, I should change my delivery address at TUG! (I'm not anymore at my alma mater)

(ah good, I gave my personal address this year)

Working on my biblatex style updates

@JosephWright blaming the biblatex team for the inconvenience? :-)

@DavidCarlisle No

All done now :-)

Didn't have to change biblatex-ieee but did do all of the others

Any Windows users here?

@AlanMunn possibly, depends how you define things..

@AlanMunn Go away!

@AlanMunn Yes

@DavidCarlisle I need to get someone to install Python on a Windows machine. There are two versions of the installer available: Windows x86-64 MSI installer (for AMD64/EM64T/x64, not Itanium processors) and Windows x86 MSI installer. Are most machined 64bit these days?

@egreg :P

@AlanMunn Yes, but you can quickly check by looking for C:\Program Files (x86): 64-bit installs have this, 32-bit ones don't

@AlanMunn most certainly most home machines, only really people with "big systems" that avoid updating still have win32

@AlanMunn 32-bit installer will work on 64-bit Windows so unless you need >1.3Gb address space should be fine

@DavidCarlisle @JosephWright Ok, that's kind of what I thought (on both counts).

@JosephWright I'm not sure I understood this though. What exactly would they look for to see that they have a 64bit system?

@DavidCarlisle Also people who want stability :-) (At work most of our instrument PCs are WinXP, 32-bit, as upgrading might be difficult-to-impossible)

@JosephWright These are people with their own laptops running Windows 8 or 9.

@AlanMunn If they have a 64-bit system they'll have both C:\Program Files and C:\Program Files (x86) as many Windows programs are 32 bit

@AlanMunn Doesn't matter, you can always tell by looking for the two locations

@JosephWright yes you probably count as big systems:-) (or small systems attached to expensive equipment, which comes to same thing)

@JosephWright Ok, so a 32 bit machine will lack a Program Files (x86) directory. Got it. Thanks.

Anything else I might need to know about running Python on Windows? (If any of you have any experience with it).

@AlanMunn i use the cygwin one (64bit) although I also have python for x64 installed (for use with the pycharm IDE)

@DavidCarlisle I'd prefer not to have them install cygwin as well if possible.

@AlanMunn the windows one seems fine (and pycharm seems OK too, although I mostly got it as a suggestion for the boy) at work we use cygwin for everything anyway (x windows, make, bash, python, perl, ..... so it's not really an overhead:-) (I only have cygwin tex on here as well, not a native windows tex)

@DavidCarlisle Ok. Yes, I'm sure that cygwin would help for other things, and we may need to go that route, but I'll try the windows one first. Part of the issue is that I'm here and they are in Buenos Aires... so I don't have access to their machines readily.

@AlanMunn You can also simply look in the system info. On windows 10 right click on the start menu and then system ...

@UlrikeFischer Good to know. Thanks. (As you can tell, I'm not a Windows user at all...)

@AlanMunn You need only a bit faith. Such an info must be somewhere (you can also simply search for "system". This will lead to the systeminfo too.)

@UlrikeFischer True, but I'm doing this long distance with people who are likely running a Spanish locale system. So ample faith will be needed. :)

@AlanMunn you could send @PauloCereda, it's only down the road and I'm sure the language differences are minor:-)

@DavidCarlisle Yes, a mere 2250km down the road... :)

@AlanMunn how long is a piece of string?

@DavidCarlisle Very long.

@AlanMunn but @PauloCereda loves to ride on a shaky bus.

@DavidCarlisle Experimental picture. :)