To the pros: Have you been following tex.stackexchange.com/questions/434273/… ? There is a lot of discussion going on under this question.

twitter.com/PHDcomics/status/1003307602648842240?s=09

@PauloCereda ^^^

@CarLaTeX You are mean

@AlanMunn Currently 1146 comments on Hacker News on the github acquisition. (@percusse @DavidCarlisle)

@HaraldHanche-Olsen Not surprising, but like @DavidCarlisle says, in reality what change do people expect (GitHub has it's own issues as a company, despite a good product)

@Skillmon oh no

@CarLaTeX ooh

@PauloCereda @DavidCarlisle Goliardia :)

@DavidCarlisle Do we need 'Suggest an edit' at the bottom of each FAQ? Should be doable ...

@JosephWright not sure, there was one on the default layout of course. the link would be useful sometimes but I was thinking more of "source on azure github" rather than "suggest an edit" as we don't want to encourage a stackexchange model of continuous edits (at least, not after this week's edits done:-)

@DavidCarlisle Could be 'Suggest an improvement', 'Log an issue' or similar: just thinking we might want some easy 'Fix this page' link

@JosephWright unless we want to encourage in-browser edits of the markdown, probably it's enough to have a fixed "source" link along with "about" and "contributors" in the sidebar to the top of the git repo.

@DavidCarlisle I suspect a lot of people will be leaving GitHub ... Norbert is already talking about setting up a TeX-facing Git system ...

@DavidCarlisle Ah, could be

@JosephWright we can certainly add one and if it encourages bad edits we can remove it...

@JosephWright shame as having everyone in one place is more useful than the details of what that place is...

@JosephWright gitlab was tweeting it was their best ever weekend for acquiring new projects from githib....

Am I the only one not worried or not caring too much about MS owning GitHub? :)

@PauloCereda No: I don't see an issue either

@PauloCereda In the end, the key is that the sources can be moved, and as long as they keep an API for accessing (and so exporting) issues, who cares ...

@JosephWright :)

@JosephWright Yes, indeed. :) And as @TeXnician said, we can always run our own Git server. :)

@PauloCereda For a lot of people, MS will never be acceptable, even though GitHub and others are backed by goodness knows who, and have all sorts of dodgy behaviours at a corporate level (that MS and indeed Apple have avoided in the main)

@PauloCereda I can't see any issue for open source projects. If you are a commercial competitor to microsoftand have code in an enterprise github private repo I can see you might be concerned about the sysadmyns having root access working for microsoft

@PauloCereda That's already what Norbert's suggested: the issue isn't the Git hosting (trivial), it's all of the support stuff, which is what makes GitHub/BitBucket/GitLab work

@PauloCereda running a git server is the easy bit, git isn't what made github useful...

@DavidCarlisle Even then, if you are using it that way I hope you have good legal cover (in the end, that's your real protection)

@JosephWright Yes. And we can run our own Gitlab server. :)

@DavidCarlisle Exactly: I see people are talking about GitLab, but I'd be more tempted to move to BitBucket (always a good website)

@PauloCereda Then one of us ends up bogged down in that ...

@JosephWright ooh

@JosephWright sure but could you afford to take MS to court (about anything..)

@PauloCereda What's pretty easy, but setting up runners for CI and GitLab pages on an own server is pretty tedious…

@DavidCarlisle If you go that way, almost no-one could use the enterprise stuff: most start-ups are much smaller than GitHub as-is ...

@DavidCarlisle Not guilty your honour.

@TeXnician Exactly

@TeXnician That's true. :)

@DavidCarlisle Talking of legal issues, did you watch the Jeremy Thorpe series on BBC1?

I don't even have BBC0 in here... :(

@DavidCarlisle Anyway, back to the FAQ: I think we look pretty good now, and I do hope to talk about it (perhaps a workshop) at TUG2018 (@PauloCereda, @StefanKottwitz)

@JosephWright oh no

I WANT TO GO TO THE COFFEE BREAK

@PauloCereda bbc.co.uk/…

@PauloCereda If the committee like the idea, of course

@JosephWright go ahead. :)

@JosephWright ah

@JosephWright no skipped it so far

@JosephWright sure but it sounds smaller and your board probably haven't heard of it, which is an easier sell than "we have put the entire source code of our project on the servers of our major competitor" ...

@DavidCarlisle True, true

@JosephWright yes maybe needs a couple of passes tagging questions

@DavidCarlisle MS must have thought about this, unless of course all they want is the large-scale Git technology, and everything else is going to get zapped

@DavidCarlisle See mail

@JosephWright and also, if they got wind that the owners were amenable to selling, they'd rather buy it than someone else buying it (which is the only other alternative)

@DavidCarlisle That's also true

@AlanMunn Well, it is hardly surprising that lyrics with more words in total have more unique words as well. But does that explain the whole difference in the lower graph? To answer that question, we would need some statistics showing the number of unique words of a text versus total word count. I'd suspect that some sort of power law would be a good fit. Surely, such data has been collected and studied? (@percusse)

@HaraldHanche-Olsen You underestimate the repetition in rap and hiphop songs.

The striking part for me is that there is a clear winning formula by the music industry that shapes all genres

@percusse except I didn't estimate it in the first place …

@HaraldHanche-Olsen Well, it is hardly surprising that lyrics with more words in total have more unique words as well this assumes that implicitly

more words doesn't take into account the repetitions. Hiphop despite its repetitiveness have more unique words by far

In that sense hiphop listeners are more literate :)

@JosephWright on the faq I think possibly the one thing I miss are next/previous links so that you could browse the site as you would the pdf, just reading sequentially (useful when reviewing if nothing else) I think it would just mean abstracting some of the code in the index into the sidebar of the default page, so you could reconstruct the page order to know which one is next....

@percusse I am sure @egreg counts himself in this group.

@DavidCarlisle Is that group Abelian? If not, no way.

@egreg hmm you mean listening is a commutative operation and people can listen to you hip hopping? Interesting thought.

@egreg ooh the Abel bloke

Isn't him who got shot because of a lady?

@PauloCereda No, he died of tubercolosis. You're referring to Évariste Galois.

@PauloCereda you mush be thinking of Galois.

@egreg, @HaraldHanche-Olsen oh that's right!

I think both died very young, that's why I got confused.

@PauloCereda “Je meurs pour une infâme coquette”.

@DavidCarlisle Hmm: how are you defining 'next' and 'previous'?

The default flare of my newly created profile on GitLab is the same as on TeX.SE. Cannot believe in a coincidence. Anybody else noticed the same?

@AlexG Do you use the same email address?

@samcarter yes, I do.

@AlexG Gravatar then

@AlexG en.gravatar.com/site/implement/images

@AlexG The image is some hash of your email address.

@JosephWright Thank you. Didn't know TeX.SE and others outsourced this.

@JosephWright order as given in index.html (which is arbitrary but means you can read the whole thing by hitting next without having to go back to table of contents, scroll to where you were, then choose next entry)

Did anyone else try to use GitLab's importing service for repos. Because all attempts to import my repos from GitHub failed. Maybe GitLab is currently over-loaded?

@AlexG @JosephWright the really strange thing is the one generated for me is in fact a reasonable facsimile of my face....

@AlexG Probably it is ... I've nto tried

@DavidCarlisle :)

@AlexG Take it you don't like the current news?

@JosephWright Exactly.

@DavidCarlisle Ha ha, lol.

@AlexG I wouldn't jump ship just yet:-)

@AlexG I'm with @DavidCarlisle: for open source projects, I'm not really concerned, but I would be if I was using it for proprietary code

@JosephWright It's because I like GiLab's logo much more :)

@AlexG :)

@AlexG I'm trying one out ...

@AlexG I see they have a layout very similar to BitBucket ...

about.gitlab.com/2018/06/03/microsoft-acquires-github

@JosephWright just in case.... eightportions.com/2016-08-21-Migrate-blog-to-GitLab-pages

@DavidCarlisle Did occur to me ...

@DavidCarlisle In the end, if the feeling is to move, I guess we all will: it's a social question

yep..

Cannot wait to see Clippy being featured in Github. :)

@JosephWright ^^ dito

Coming soon: Microsoft arara™

A friend of mine sent me an email: "I don't trust this Microsoft + GitHub merge. Sent from Windows Phone."

Meanwhile...

@JosephWright ^^ it's your duck avatar. What were you doing out there? :)

blog.github.com/2018-06-04-github-microsoft

@JosephWright it's all going to be OK, it says so... ^^

@DavidCarlisle Off to GitLab then ...

@JosephWright For real?!

@PauloCereda Partly serious: we've already seen @AlexG and @norbert start on moving their code, and the key to GitHub's success is that people have congregated there, more than any technical aspects. In particular, if @norbert does set up a 'TeX GitLab' system, we might all end up there ...

@JosephWright I can migrate arara there too. :)

Complementing @DavidCarlisle's post: blogs.microsoft.com/blog/2018/06/04/…

@PauloCereda I'm going to 'wait and see': as @DavidCarlisle says, the real issue is likely to be for closed-source people

7.5 BILLION

Holy cow

@DavidCarlisle May I ask to request to consider making my question a community wiki ? tex.stackexchange.com/questions/434916/… . Newcomers who struggle to choose between the two packages mentioned there can benefit immensely

@PauloCereda if someone offered you $7.5Bn for arara, would you sell?

@DavidCarlisle I'd sell for 100 boxes of chocolate. :)

We ducks love chocolate

@Krishna You have more access to that question than I do, if you posted it.

@DavidCarlisle thank you. I am unsure if the questions meets the standards to justify being a community wiki.

@Krishna I can't see why you want that actually, it doesn't make any difference other than you would get no rep points for any votes.

@Krishna Having read the post... it's verging on being too broad or "opinion based" for closure, there is no real question there at all...

@DavidCarlisle Its not for the rep points, it is for visibility. Maybe I did a very quick search (as I am writing a thesis now), and I coudn't find something similiar to justify making a decision between the two. You know these manuals are very big.

@DavidCarlisle most CW posts are like this? eg, the comparison between text-editors. The idea is to engage the experienced users to provide their experiences with these packages. The user can then decide for himself. Reading both those 100+ page manuals just to make a decision is a no-go, especially when there is a wealth of knowledge here in the community, why not leverage it?

@JosephWright: Microsoft right now: I won't peek your private repo for a pound. MP sketch. :)

@Krishna are cw posts more visible? I hadn't noticed that they stay on the front page any longer than any other post. Anyway I'm not a moderator so I have no more say in this than you do (or actually less say as you own the post)

@Krishna I think the question starts from the wrong premise. "But once a user has decided to use the pgf engine for their graphics," Deciding on the back end is the last decision that should be made, not the first. You could implement tikz (and so pgfplots) over a system other than pgf (for instance l3draw, or luatex a native lua back end) that's implementation detail that doesn't matter at the high level you are asking your question....

@Krishna If you specified actual requirements perhaps they are met by asymptote, or pstricks, why pre-judge and commit to pgf?

@DavidCarlisle good points. I read the section on typesetting consistent graphics in the pgf manual. Really strong points. There could be many reasons the user chooses to do pgf. Prior experience in creating illustrations with tikz may be one. asymptote is C++, and that scares away many people, including my professor (who calls it ugly).

@DavidCarlisle I thought a lot of comparison-type questions are community-wikis. I didn't dare to make one for fear of drawing the irk of experts here, who might disagree with my line of thought. Anyway, I'll leave it like that. These manuals are large tomes, and an informed choice to read which one first and in detail makes a difference :)

@Krishna I suppose your professor has looked at the source code of tikz to make an informed comparison? :-)

@DavidCarlisle :)

@DavidCarlisle can you post running code tex.stackexchange.com/questions/434532/…

@DavidCarlisle haha. No, he was just making a statement in a programming context, not in a documentation context. I just skimmed through the l3draw package. You guys rock with the l3 progress. Cleaner code it may be, but l3draw is marked as highly experimental, and it obviously doesn't ship everything needed to get on with my drawings now :)

@Krishna I can probably provide everything for the 'back end' of pgfplots ... the stuff that is not done is almost all not required

@JosephWright Thank you for chipping in. Would you recommend l3draw for a thesis being written today (for plots and artistic illustrations)? If so, great - I have 3 manuals to read.

@Krishna I just mentioned l3draw because its an example of why your question is mis-placed. The document markup you choose is what matters, if someone comes along and replaces the pgf back end from your drawing package then would you consider not using it any more?

@Krishna Probably I could hack up some basic pgfplots stuff quite quickly ... but I'm trying to sort out siunitx v3 (under pressure from one FMi)

@DavidCarlisle We (I) are going to end up writing our own TikZ system, aren't we (I)

tikz data-visualisation? pgfplots? I thought MetaPost was the only option. :)

@DavidCarlisle and @JosephWright . I am sorry for any stress caused by these questions. My intention was to simply sort out the issue of which manual to read first (among currently existing solutions), and thought that the community can help me (and others) to make this decision. l3 is a long-term research project, and obviously you are doing this wonderful thing despire holding day jobs. My goals and aims are a bit more modest, and that is to decide on a plot-drawing package

@JosephWright then you'll have to add some texfaq answers describing it...

And wait for it....to stop procrastinating and get on with my reading and more importantly, thesis writing :)

@DavidCarlisle Yup

@JosephWright but first you need to take a week to recreate the faq build in gitlab:-)

@DavidCarlisle :)

@DavidCarlisle Need to wait for their import system to recover from the current load!

@JosephWright, @DavidCarlisle ^^

@PauloCereda :)

@JosephWright Too late to remove the GitHub patch from my laptop cover. :)

@JosephWright ^^ :)

Today is a funny day. :)

@JosephWright Did you saw tex.stackexchange.com/q/434928/2388?

@UlrikeFischer No, but I know we need to sort it

@UlrikeFischer We have a (stalled) mail thread on the team list about how to name/conceptualise the underlying functions

@JosephWright ouch

@JosephWright:

From my Gitlab account. :)

Bonus points for gitlab: my username was still available, while I had to add a number to my name at github.

@PauloCereda I just tried to make an account but it says my email domain is not authorized. (gmx.com)

@AlanMunn oh no

@AlanMunn: you can sign with GitHub. :D

@PauloCereda I tried that and it failed too. :(

@AlanMunn wow

@PauloCereda So I used my gmail account. But weird.

@PauloCereda It's not clear that it's worth moving, but it can't harm to set up the account.

@samcarter my bonus point is that I could use u_fischer with underscore there like for my other accounts (I have an account already since last year). github forced me to u-fischer.

@AlanMunn Yes, you are right. There's a strong commotion, but I will wait a bit longer.

@AlanMunn good idea I just signed up to grab my user name:-)

@DavidCarlisle realDavidCarlisle?

@AlanMunn LOLOLOLO

DavidCarlisleOfficial

@DavidCarlisle on a more serious note, you had a notDavidCarlisle one, don't you?

hehe, good idea, even though, I hardly ever use git (only to clone something), I might as well just make an account, if nothing else other than getting the appropriate username ;-)

@UlrikeFischer :) very good reason for bonus points, too!

OOH ALLURA allura.apache.org

@PauloCereda that person was deleted I think

@DavidCarlisle oh no

@PauloCereda but gitlab.com/davidcarlisle

@DavidCarlisle ooh

@cereda Member since June 22, 2016

@PauloCereda strangely he looks like a user here, not like a user with the same name on github

I had an account before the chaos. :D

@DavidCarlisle Damn internet people

@AlanMunn DavidCarlisleWithAShiftKeyJustLikeBarbara

@UlrikeFischer Did github decline your underscore in username with Missing $ inserted.?

@PauloCereda What's the allura this?

@AlanMunn LOL

I've created a new room to dicuss the GitHub business: chat.stackexchange.com/rooms/78439/github-take-over

@samcarter ;-)

@JosephWright pin it

Something more enjoyable than the github business ;-)

@PauloCereda Done

@marmot ooh marmothub

Everyone trying to import their projects into GitLab? ;)

@JosephWright That is the result of Github buying Microsoft in order to use Comic Sans in their web page, I guess.

@mickep :)

macOS Mojave

@PauloCereda Looks good?

@JosephWright will report back. :)

@PauloCereda Great

@JosephWright: Dark mode confirmed.

@PauloCereda BLACK!

@JosephWright :)

@JosephWright No I'm looking at the rush from a safe distance. Also I'm waiting that microsoft steels my ducks and make a logo from them ;-).

@UlrikeFischer ooh duck stealing

@UlrikeFischer I'm trying a couple out, but just for testing (I'll zap them as/when necessary)

@PauloCereda Stay tuned!

Microsoft partners with Stack Overflow

@StefanKottwitz oh no

@PauloCereda I like this screenshot from that link:

@StefanKottwitz This is ironic. The one part of Stack Overflow that almost everyone agrees is crap is the one MS is getting involved with. (Or maybe we're the only ones who think it's crap?)

(see the credits at the end)

@AlanMunn Nope, I think that was general

@AlanMunn got not getting, that's an old post

@StefanKottwitz :)

@DavidCarlisle That's 2 years old. No agreement yet on the price I guess :-)

@DavidCarlisle Ah right. So it's done well. :)

@UlrikeFischer Beware: You will be unable to use GitLab pages with that name…

@TeXnician ?

@JosephWright meanwhile back on github, I have to decide how much I like the next/previous links and whether to push for them upstream..... they work (see next here davidcarlisle.github.io/texfaq.github.io/FAQ-startup ) but they have a big impact on build time as seems there is no global state so you have to harvest the list of files and essentially generate the full table of contents ordering on each page.

@JosephWright they make it a lot more comfortable to review the site...

@DavidCarlisle We don't build the pages ...

@JosephWright no but we wait for gh to build them before seeing if an edit looks good

@DavidCarlisle True ...

@JosephWright I should time it but it's gone from "don't really notice how long it takes" to "it takes several minutes"

@DavidCarlisle Hmm, a bit of a pain

@DavidCarlisle Personally, I don't particularly see that the marginal idea of 'next'/'previous' is worth it here

@JosephWright of course it's possible my liquid coding is sub-optimal (it took a few goes to get it working) but the basic plan is a copy of the code in index.html, and googling for how to save a collection list globally just turns up people complaining it's not possible...

@DavidCarlisle I suspect it's not possible: Jekyll is really meant mainly for a particular task (blogs) without complex programmatic structures

@JosephWright seems that way.

@JosephWright I certainly missed them while reviewing the site, but I wasn't expecting the build time to take so much of a hit, for a normal user of course they'll come in via search so traversing the site isn't so much of an issue.

@DavidCarlisle Exactly: really for users there is no ordering of questions that makes any useful sense (cf. a blog where there are dates)

@JosephWright well yes but when encyclopedias were books I'd often just read on past the item I'd started to read just because...

@DavidCarlisle Yeah, sure (much the same for me reading papers)

@JosephWright another possibility I might try would be to have a link back to the table of contents that is # id linked to the link to that page rather than the top of the index, then you can easily select the next page.

@DavidCarlisle Might be lower load

@JosephWright or most likely just leave it on my fork, so i can navigate on my fork as I want, and not touch the main build...

@JosephWright there wouldn't be any extra time at all I think just a matter of putting id=thispageid on each of the outgoing links in the toc and having a link to index#thispageid in each faq page.

@DavidCarlisle I see TUG2018 is going to work a bit like UK-TUG: 'Joseph comes up with some talks'

@JosephWright I'm expecting you to video record and take minutes too

@DavidCarlisle :)

@DavidCarlisle I did take notes at TUG2015 ...

@DavidCarlisle texdev.net/?s=tug2015

@DavidCarlisle Kaveh's not going ... better pack my mic (@PauloCereda) ;)

@JosephWright my next/previous links show interesting bug the previous link on davidcarlisle.github.io/texfaq.github.io/FAQ-oldfontnames is 404 because the permalink in FAQ-parmoderr.md is short of an r ... I wonder if we should mechanically check those permalinks...

@DavidCarlisle Odd: they were all auto-generated from the file names!

@JosephWright just checked, it's the only one.