TeX, LaTeX and Friends - 2015-04-06 (page 3 of 3)

« first day (1618 days earlier) ← previous day next day → last day (3309 days later) »

9:00 PM

@FaheemMitha yes

@SeanAllred From a TeX mailing list:

> Would you consider converting the repo to git and hosting on GitHub?

No, I don't want to do that. If anyone feels compelled to use git,
they can use git-over-svn. There is no need for any contributor to use
any VC at all, as far as that goes. Just sending patches works
perfectly well here.

Playing around with infrastructure (or the organization of the document)
is easy, but a waste of time in this instance. What's needed is serious
time and attention spent on working on new content. Git and github are

@JosephWright well, they don't pay the two people who do the real work. authors and reviewers. It's a total ripoff.

@JosephWright I'd have to disagree with that.

@FaheemMitha I'm not sure you are entirely right: there is a cost here and it's more than a few pounds

@JosephWright Which makes no damn sense. I could retypeset a paper while paying myself a decent wage, email it to the peer-reviwers, and manually upload it to a server via FTP for less then that.

@Canageek But could you run a server farm to make it and the past 150 years of papers available?

9:01 PM

@JosephWright anyone we know?

@DavidCarlisle One guess

@JosephWright That isn't the cost to publish a new paper. That is legacy costs.

@JosephWright Try publishing with some of the greedier open access people. Do an approximate computation of the value of the time they spend to put your stuff up. Add some amortized cost for associated infrastructure. Then look at what they charge. It doesn't add up.

@Canageek Yes, but it has to come from somewhere

@JosephWright Also, they are tiny PDFs, and most DON'T have the old stuff up there. I've been sending out ILL requests all week.

9:03 PM

@Canageek I meant the work for digitising stuff where there are subscriptions: almost certainly there is cost-sharing

@FaheemMitha Like I said, without seeing the books its very hard to know the real situation

@JosephWright granted, sunk costs for a journal are not insignificant, but they have no resemablance to the prices they charge per paper.

@JosephWright you'd be horrified by the real situation.

@DavidCarlisle It's on TeXhax

@FaheemMitha I suspect the RSC are not far off what it really costs (once you allow for investment in new ideas): the profits can only go back into promoting chemistry and I do know what they get up to in that regard

@JosephWright I'm writing to those galapagos dudes. Shall I ask them if they could declare their font to be public domain?

@JosephWright RSC?

@JosephWright Considering how many websites I know of that offer not-dissimilar hosting services based solely on ad revenue? I'd have to see the books, but I suspect that either a) They are doing things VERY inefficiently or b) They are taking home some very large profits.

@FaheemMitha Royal Society of Chemistry, my professional body

9:05 PM

@FaheemMitha Royal Society of Chemistry (UK)

@JosephWright promoting chemistry?

@FaheemMitha Yup

@JosephWright Oh, I was speaking generally.

Not about a particular publisher.

@JosephWright Technically the ACS is under similar rules, but they manage to pay a bunch of people half a million+/year and have big executive meetings in Hawaii at luxury hotels.

@FaheemMitha Yes, but as they do publish open access costs and as I think I know where the profits go, I have a sense for what might be fair

9:06 PM

@Canageek Probably a mix of (a) and (b), but mostly (b).

So, anyone? Galapagos?

@Canageek ACS?

@Canageek ACS money comes from ChemAbs more than anything, and that does cause some tensions

@JosephWright I'm sorry to have to say this, but you are too trusting.

@FaheemMitha American Chemical Society

You need to develop a suitably dark view of human nature. Come to India. It'll help.

@FaheemMitha I'm on a committee for another area where some of these profits are being invested

9:08 PM

@FaheemMitha American Chemical Society, one of the RSC's main competitor (Along with two for-profit companies: Elsevier and Wiley)

@JosephWright Regardless, this doesn't scale across all the OA people.

@Canageek are you a chemist too?

@FaheemMitha Or at least, on a committee where we get some insight into this

@FaheemMitha Yes

@Canageek ah, ok.

@Canageek Well Wiley is a bit odd as Angewandte, etc. are actually not owned by them

9:08 PM

@JosephWright Yeah, I was never clear on what is going on with that.

So, Galapagos?

@Canageek Situation has varied over the years, so it's hard to keep a track. I'm not in the GDCh but I assume they get some money out

@JosephWright Now if they are paying for the Angewandte translation costs, I could TOTALLY see $4000/article being justified.

@Canageek :-)

@JosephWright But most chem journals: You pay to publish it, you pay to access it, and you pay AGAIN to be on one of three 'front covers' per issue.

9:11 PM

@Canageek There are tons of OA journals that are both expensive and for-profit. The ones that exist to support science are in the minority.

@FaheemMitha I can only think of two OA journals I have used as a chemist (PNAS, and Chem. Sci.)

@Canageek Oh the front covers are a pain

@FaheemMitha Plus possibly a couple of national journals (I think Can. J. Chem. might be open access, but it isn't nearly as good as it once was)

@Canageek Chem. Sci. is RSC and presumably costing me (as a member) something!

@Canageek You got in PNAS?

@JosephWright Yep. They are also my bosses most hated journal.

9:13 PM

@Canageek Yes?

@JosephWright No, but I've used papers from it.

@Canageek Try Chem. Eur. J. for a right PITA set of instructions to authors

@JosephWright Lots of reject without reviews, when other journals with similar impact factors.

@JosephWright Is that Open Access?

@Canageek No, but they have the most picky set of rules (in the template for submission, then everything out, separate graphics, etc. when accepted)

@Canageek I've not tried there, but did get rejected out-of-hand by Nature Chem. :-)

@JosephWright I've got a paper in it. They also totally messed up some of the formatting in a way that changed the meaning of an equation, and refused to fix it despite being told about it in the proof stage.

9:16 PM

@JosephWright just saw that on the bus home and thought of the recent discussion here too. :-)

@Canageek The pay to publish/pay to access business applies when you have a 'mixed' model even if no-one was making any money

@JosephWright Really, it was the fact that the speed the rejection came back meant he was rejected by an AUTOMATED SYSTEM, which he really hated. At least ACS journals look at things before rejecting them.

@PaulGessler Indeed: he has a point

PLOS has terrible service, in spite of being really expensive.

@Canageek Got rejected by JACS once in about 8 hours

9:17 PM

@JosephWright We haven't put out a JACS paper in a while. I got in at once at 7th author or so? My current one is aimed at either Cryst. Eng. Des. or Dalton, I think. Unless I get more results by the time I'm done writing it.

(Which is highly possible)

@JosephWright I want to know

@SeanAllred Take a look at TeXhax archive, then :-)

With all due respect, that's really callous of the author.

@SeanAllred Not really: he works very hard on a lot of stuff

@JosephWright This was well under and hour as I understand it. 8 hours, someone could have seen it. Less then an hour?

9:20 PM

@JosephWright I can't seem to find the thread -- what's the title?

@SeanAllred tug.org/pipermail/texhax/2015-April/021621.html

@Canageek I've got friends at RSC Publishing but no-one on Chem. Sci., so I wouldn't like to say

@JosephWright Ugh.

Can any TeX Live related person comment on this suggested text?

Hi,

I'm writing to ask that you consider making the Web-O-Mints font
explicitly available under some kind of open license. Publicly declaring
it public domain would work, for example.

These fonts are currently on CRAN, and are listed as Public Domain, even
though no specific license exists ->
https://www.ctan.org/tex-archive/fonts/webomints

I see you still have these fonts listed on

http://www.galapagosdesign.com/download/index.html

Because you don't have an explicit license for these, they cannot go into

@FaheemMitha Look up Creative Commons 0

@JosephWright The logic there doesn't work >.<

9:22 PM

@SeanAllred Huh?

@JosephWright It's like someone with impaired vision refusing glasses because 'he gets around just fine now'

@Canageek is that better than Public Domain?

@JosephWright :20947531 Yeah, we don't quite get his dislike, but his personal theory is they are doing keyword filtering and don't like coordination polymers (his main area of research).

@FaheemMitha You're missing the most important phrase: "Thank you for considering this." (or similar) and "Yours faithfully" rather than "Regards" :)

@yo' True. I'll change that. Anything else?

9:24 PM

@FaheemMitha not really I think

@FaheemMitha It is a legal text to place it as close to public domain as possible (As I recall, there is actually no legally binding way for a US author to make something public domain?). Then it has an explicit license you can reference as well.

@SeanAllred Wait until you've tried and failed to get people involved in a project :-)

@Canageek France is an example place with no public domain, not sure about the US

@JosephWright I have this unrealistic ideal that smart people working on smart things make smart decisions

@Canageek Ok, so should I recommend CC0 then?

Anyone else with comments?

@Canageek P/D exists in the US

9:25 PM

@FaheemMitha You should examine it yourself, it has been a long time since I looked at it.

@DavidCarlisle -- when tex was created, there was no such thing as a scalable font.

@SeanAllred I think he's probably wrong about Git, although quite possibly right about GitHub :-)

@SeanAllred I know it does, but I thought that putting something into it was more complicated then just declaring it public domain?

@JosephWright Oh definitely -- I can fully support the idea that github could be a passing fad

@Canageek Nope.

@SeanAllred Nevermind, I was wrong then

9:26 PM

@Canageek at least, not to my knowledge.

@JosephWright The Czech legislation has a mixed feelings about this, one of the central courts recently said that "things declared as 'public domain' are to be used without any fees to anybody".

@Canageek I've seen, used, and released projects into the public domain just by declaration

@JosephWright He is. DVCS is very powerful. Though you might not give it a workout.

@SeanAllred CC0 is the legal text that places something into the public domain in a binding fashion: creativecommons.org/publicdomain/zero/1.0

@Canageek will do. but I'd like the opinion of people more experienced at this.

9:27 PM

@Canageek I'm familiar :)

@DavidCarlisle care to weigh in?

@Canageek That counts as a declaration

@SeanAllred Yes, but it is also about a page long, not just one sentance.

@yo' I was thinking the 'give up copyright' idea: that's simply not possible in some places

I know, I'll ask #debian-mentors.

9:28 PM

@Canageek Of course -- as we've seen with the Czech legislation, public domain can mean very different things in different countries :)

choosealicense.com

@JosephWright How do you say micromole without any packages? If I use math mode the micro will be italic

@FaheemMitha I meant the 'passing fad' think: my feeling is Git will be the 'winner', so we'll see RCS/CVS/SVN/Git as the most stable 'chain'

@Canageek Yup

@JosephWright there's the "lease all rights that can be" idea :)

@Canageek Which publisher are you targetting?

@JosephWright oh

9:29 PM

@JosephWright We haven't specifically yet, just a range of journals.

@JosephWright It's unfortunate that git is so tied to github in the public mind -- it's hard to distinguish git's specific advantages!

@Canageek I would probably not worry then and put up with italic: ACS certainly do this in print

@JosephWright My boss was leaning towards Dalton as I recall, but he said to start on the materials and methods section before that.

@Canageek Ah

@Canageek You almost certainly can use textcomp, which has an upright mu

@barbarabeeton that's why I was confident in the accuracy of my assertion:-)

9:31 PM

@JosephWright I thought several journals literally said 'no packages'...which makes me wonder how to put input in without graphicx or citations without achemso rsc or biblatex..... As I recall our groups solution was submit it using all the packages we want and have no one call us on it.

@FaheemMitha I think you've got to bear in mind that there are a lot of people working on projects where all of this 'new' stuff is unlikely to actually impact, and some of them know it!

@FaheemMitha on?

@DavidCarlisle suitable license for ctan.org/tex-archive/fonts/webomints.

@JosephWright acknowledged and agreed

@Canageek suggested CC0.

9:33 PM

@Canageek They do tend to say 'no packages' but the understanding behind that is very limited. Anything from the core set from the team (so graphichx) will be fine, for example.

Well, a license that Debian would like. I'm asking on #debian-mentors just now.

@JosephWright Ah, OK.

I was thinking asking them to put it in Public Domain, but apparently that can be problematic.

@FaheemMitha Creative Commons is not recommended for software: SIL License?

@JosephWright But I'd still press that making it more accessible is good for everyone. It might not be impactful for any one person, but the project as a whole would benefit.

sigh

9:33 PM

@SeanAllred I see that

@JosephWright Don't know that one. Let me look.

@Canageek biblatex is in general discouraged by most journals

And I know, I know, you'd be supportive of the move.

I'm just a little frustrated :(

@SeanAllred The mail we are talking about is on a different area: TeX Live is not about to get a massive community take up in terms of development

@yo' I know, but I assumed they'd all have some alternative. ACS has achemso for example, and I think the RSC made their own versions of it that borrowed @JosephWright's code

9:34 PM

@Canageek Indeed

@JosephWright Oh, that's TeXLive-specific?

@Canageek RSC recommend my BibTeX style :-)

@SeanAllred it's not always the case that ease of access is good, and I think Karl's main point in that thread is that means of access is a non-issue, getting (anyone) to do the work is the issue.

@Canageek use whatever bibtex style you wish and in the final version, replace \bibliographystyle{whateverstyle}\bibligraphy{mybib} by the contents of myfile.bbl ?

@DavidCarlisle I see that, too. But I think it might be a chicken/egg problem.

9:35 PM

@yo' Oh, I'll probably do that. But it is more work then that since they all use incompatible citation comands.

@Canageek use only \cite, problem solved. The class I developed actually does \let\citep\cite and \let\citet\cite.

@DavidCarlisle I for one would probably go nuts :)

@SeanAllred go nuts if?

@SeanAllred The original point comes from some stuff that is linked to TL

@DavidCarlisle I'd go nuts over it, i.e. contribute/review/do happy dances

9:36 PM

@yo' As in, several use citations commands that aren't cite, as I recall. One uses \autocite I think, and I forget which is which.

@DavidCarlisle Yup

@Canageek that's because people use ridiculous unnecessary things.

I stand by my chicken/egg argument :)

@yo' I think @JosephWright has written at least one of the one that uses something that isn't \cite ;)

@SeanAllred if you'd contribute if it was on git(hub) contribute anyway, the costs on changing version control are all on the current maintainers, they shouldn't have to change their ways just because you want to contribute.

9:38 PM

@Canageek achemso uses natbib, as that is common, stable and has been around for years

@DavidCarlisle Of course -- the needs of the many outweigh the needs of the few -- but I'm certainly not the only one who would contribute.

@JosephWright That is \cite, right? I'll probably use that for the citations for the paper, even if it does mean I have to manually go through my .bib and put in all the accents manually.

@Canageek sorry, I changed one word to make it more clear. @Joseph only did what should be done to make things properly if people want them.

@Canageek what? why?

@yo' I thought you had to if you used bibtex?

@yo' I thought that was the advantage of BibLaTeX over BibTeX

@DavidCarlisle It's a definite fact that we could still be using the email-it-back-and-forth VCS, and that would likely have been more convenient for people at the time (or at least with RCS). Does that make the move to SVN any less valid?

9:40 PM

@SeanAllred but is there any evidence that people would contribute if it was on git who would not contribute if it is on svn or has no public version control.

@Canageek ah you mean in the bib file? well, I export all my citation from web databases and they do it properly - in 7bit :)

@DavidCarlisle Is there evidence to the contrary?

@Canageek That's more a Biber thing, but as long as you use 'traditional' input you are fine

@SeanAllred no need for such evidence, you need an argument to back up a change from the status quo

@yo' Hahahahahaha. Mendely TRIES, but it often has errors in its export. Exporting directly from the journal webpage almost never uses TeX. They either have the unicode, or just DROP all the accents completely.

9:41 PM

@Canageek they should ask @Barbara how to do things properly. Or even @barbara, for that sake.

@JosephWright I think Mendely can do that for me automatically, but journal webpages don't.

@DavidCarlisle Fair enough -- but I think examples do count, here

This is probably a dumb question, but does the stuff on ctan.org/tex-archive/fonts/webomints include sources?

I've no idea what a font source looks like.

Personally I see git as vastly more complicated than svn and while not impossible to use I'd say a git based repository has a higher barrier for people to use and contribute than an svn one.

@DavidCarlisle I suppose you didn't get this part of the conversation, but I've had a few ideas over the past few months that I wasn't comfortable putting out there formally.

9:42 PM

Talking to one of the Debian ftpmasters on IRC.

It's a culture thing, we determined.

@FaheemMitha fonts are probably considered "resources" and need not have the sources

@yo' hah. try convincing debian of that. But is that typical then? @JosephWright?

@DavidCarlisle definitely a culture thing -- I haven't the foggiest how to effectively use SVN

@FaheemMitha Pass!

9:43 PM

Git, however, seems natural to me.

@JosephWright LOL

@DavidCarlisle It's basically the same thing as svn to a first approximation.

@SeanAllred svn co <name>, svn ci, ...

you can use it as a centralized thing if you want.

@FaheemMitha that's simply false:-)

9:44 PM

@DavidCarlisle It is not.

@DavidCarlisle I don't feel it so really.

I've used both.

@DavidCarlisle The upside of git and mercurial as I understood it, was you can have a total offline repository, so if one person takes the main one offline, you aren't screwed. Is that not the case?

@Canageek There are lots of upsides to DVCS.

@JosephWright There's a difference between knowing how to use it and knowing how to use it effectively. LaTeX folks should know that even more dearly than others :)

9:44 PM

@Canageek In our case, we would simply complain to Rainer :-)

An individual may or may not care about specific ones.

@FaheemMitha a system with 1000 commands that you can use like a system of 10 commands is more complicated, even if you don't use 990 commands

2

@SeanAllred My summary is only missing svn up and svn log from my set of commands ;-)

@JosephWright what's svn up?

@DavidCarlisle Granted, git's UI is a piece of crap. But the underlying command set for centralized work is pretty simple.

9:45 PM

@JosephWright I've heard stories about disgruntled employees/contributors wiping out the source tree on their way out, which is one of the reasons you cannot alter history in git, no matter what.

@FaheemMitha That is why I'm still on Hg. GREAT GUI.

Most of the other complexity can simply be ignored to a first approximation.

@yo' Update (pull in Git parlance)

@JosephWright Many of the ideas I've come to associate with proper version control simply don't transfer well to SVN, though. How is someone supposed to denote that these changes are related to FEATURE? How is one supposed to related several changes together?

@Canageek glad to meet another hg user. great GUI as in thg?

@Canageek LaTeX SVN run by Rainer, unlikely to do this (he runs the CTAN core server too)

9:46 PM

@JosephWright ah ok. And you don't have that with git? git pull base master, where base is the repolink to the repo from which you pull?

@JosephWright How do you propose changes rather than make them then and there?

@FaheemMitha Yes! Right Click, commit, right click, synchronize.

@JosephWright Email, yes, but that's not within SVN -- and you know my thoughts on that :)

@SeanAllred Write it in the log

@Canageek Yup, but going to loose long-term (sorry)

@Canageek GUIs are gross :)

9:47 PM

@JosephWright that's a thing I miss in git as well: change the last commit message easily, and log a message with no change easily.

@Canageek I mostly use the command line myself.

@yo' Yes, that's what I meant: svn up ~ git pull

@JosephWright I'm aware. At least you can use BitBucket with Git now, which is like Github, but better.

bitbucket is silly

@FaheemMitha I'm on windows, so I only use command line for latexmk

9:48 PM

@SeanAllred You know what I'm going to say to that!

@Canageek windows has a cmd line too.

I should really switch to autolatex, I hear it is better, but haven't had the motivation

@FaheemMitha Not a good one.

bitbucket is like the google+ of VCS hosts

@JosephWright Yes, email, but you also know my thoughts on that :)

@JosephWright Also: That makes it perfect. Everyone knows the more popular software is, the worse it is.

@Canageek LaTeX stuff is all collecting up on Gitub, so I've moved my own stuff again based on likely outcome more than anything else

9:49 PM

@SeanAllred Free. Unlimited. Private. Repositories. For Academics.

@Canageek There is that

@JosephWright I usually make one for each class and keep all my notes and assigments in it to sync them accross computers.

So, again, what is normally considered source for a font?

@Canageek It's only redeeming feature -- and I don't really care for private repositories, anyway

@JosephWright Sigh. Define lose.

9:51 PM

@FaheemMitha complete and utter abandonment

I mean, I assume a font has a source.

muahahahhahahahaha

@SeanAllred Great for when I'm writing papers and want to sync them between my work and home computer.

@SeanAllred Unlikely.

@FaheemMitha not necesssarily

9:52 PM

@FaheemMitha can be editted as binary in a font editor

@SeanAllred ?

@FaheemMitha *evil laugh*

@DavidCarlisle Ok. does ctan.org/tex-archive/fonts/webomints contain sources?

@SeanAllred Fall in number of users and particularly of new users, see for example RCS, CVS

@JosephWright Well if you're going to try and be reasonable about it, sure.

9:55 PM

@FaheemMitha no, it does not. it contains PFB directly.

@FaheemMitha it contains a type1 font, that may be generated from some "source" format or it may have been drawn with a mouse in a visual font editor or any combination of the two, it's like pointing at a pdf diagram and asking if it has a source, there may be some tikz code for it somewhere or it may have been drawn

@DavidCarlisle So, basically, it is not clear.

Ok.

@FaheemMitha well it's clear it doesn't have any "source" but it's not clear if there ever was a source in that sense

@DavidCarlisle Ok.

I always assumed all the fancy fonts were generated programatically.

Isn't that how metafont works?

@FaheemMitha wrong assumption

9:58 PM

@yo' :-(

@FaheemMitha yes, but only a small portion of fonts is generated in metafont

many fonts are made as a vectorization of some rasters with post-processing

@yo' oh.

How was Computer Modern made?

@FaheemMitha Metafont, I believe

@FaheemMitha in metafont :)

@yo' do, by drawing them, basically.

@SeanAllred ok.

@yo' ok

I like CM, but it's too light.

Can barely see it on paper.

Well, I sent the email.

9:59 PM

@FaheemMitha Find a better printer :)

@FaheemMitha blame

blame somebody. It used not to be that light. It's intended for wet printing.

@FaheemMitha that's why no real font designers used metafont. It just isn't the way they think about font design.

@SeanAllred My printer is fine.

@FaheemMitha My printer prints CM just fine

@FaheemMitha laser one?

10:00 PM

@DavidCarlisle But I imagine you could do all sorts of fancy things with it.

(My keyboard, however, doesn't like me today)

@yo' yes.

@FaheemMitha I thought so :-)

@SeanAllred inkjet or laser?

Doesn't everyone use laser these days? Inkjets are so expensive.

@FaheemMitha There are things that can go wrong with both

10:01 PM

@FaheemMitha yes, like having a ink, laser, professional, draw etc version of the font ;)

Toner is expensive too, but not so much.

@yo' huh?

@FaheemMitha ink sucks if you work with scouts :-)

@yo' scouts?

@FaheemMitha there used to be a CM variant for every printer available in the 80s

@FaheemMitha yep. Ink doesn't survive. It's also sensitive to coffee stains and water spills

@yo' oh. as opposed to?

10:02 PM

@FaheemMitha laser

@yo' not any longer apparently.

@FaheemMitha no, unfortunately.

@yo' toner fades. But it is generally preferable to inkjet.

@FaheemMitha not if you get a good one. And only a small number of things I print has to last longer than ~5 years, so fading is a little of an issue.

@yo' ok

@yo' what printer do you use?

10:05 PM

@FaheemMitha I got a laser printer because the toner doesn't dry out if you don't print.

@FaheemMitha you ask me too much, some bw canon duplex laser A4 one

@yo' ok. canon makes good printers. linux drivers can be a problem.

@FaheemMitha I got one that is for a different printer but works wells

@yo' linux driver?

@FaheemMitha yep

@yo' ok

10:06 PM

came for free with cups

@yo' do you use Debian/Ubuntu/Mint, or RH something?

or *BSD? :-)

@FaheemMitha Fedora

@yo' ok

10:22 PM

@yo' \o/

ok, Amadeus is over, I gotta go home :)

10:38 PM

@yo' Was WAM able to finish the Requiem this time?

11:29 PM

@egreg He even played. :)

« first day (1618 days earlier) ← previous day next day → last day (3309 days later) »

Transcript for

Apr '156

$TeX - LaTeX$ TeX, LaTeX and Friends

\begin{chat}

join this room
about this room

00:00

06:00

12:00

18:00

all times are UTC

$TeX - LaTeX$

site design / logo © 2024 Stack Exchange Inc; legal

mobile