You're reading the public-facing archive of the Category Theory Zulip server.
To join the server you need an invite. Anybody can get an invite by contacting Matteo Capucci at name dot surname at gmail dot com.
For all things related to this archive refer to the same person.
Antonin Delpeuch (Apr 12 2020 at 19:58):One rule I have about communication is to boycott closed-access publications, both as an author and as a reviewer. We are lucky to work in a field where there are high quality open access options, so it is an easy rule to follow.
Antonin Delpeuch (Apr 12 2020 at 21:12):If you are doing that too, consider adding it to your webpage to give that more visibility (I just do it like this: https://www.cs.ox.ac.uk/people/antonin.delpeuch/#pledges, some of my friends put more effort into it: https://a3nm.net/work/research/openaccess/, https://desfontain.es/privacy/personal-open-access-policy.html)
Lê Thành Dũng (Tito) Nguyễn (Apr 12 2020 at 21:37):Oh, I hadn't heard about https://tcs4f.org/ before visiting your web page, so thanks @Antonin Delpeuch !
Lê Thành Dũng (Tito) Nguyễn (Apr 12 2020 at 21:38):I'm going to add the open access pledge thing to my web page too
Lê Thành Dũng (Tito) Nguyễn (Apr 12 2020 at 21:39):(Full disclosure: there is actually a preprint with the IEEE template and my name somewhere on the Internet, but that wasn't my decision and I'm pretty relieved that the submission was rejected (deservedly so for purely technical reasons as well))
Jules Hedges (Apr 12 2020 at 21:40):I haven't signed the formal pledge, but at the bottom of my publications page I wrote the message "We need to gradually stop viewing publications that are not open access as genuine publications"
Jules Hedges (Apr 12 2020 at 21:41):I actually have several papers that would be closed access if I didn't host them somewhere else, because at the time I submitted them to conferences I had no idea who was going to publish the proceedings
Lê Thành Dũng (Tito) Nguyễn (Apr 12 2020 at 22:43):Antonin Delpeuch said:
One rule I have about communication is to boycott closed-access publications, both as an author and as a reviewer. We are lucky to work in a field where there are high quality open access options, so it is an easy rule to follow.
I think it's a bit more complicated than that. First of all, people here come from many different fields (this wide applicability is part of the beauty of categories :smile:). More importantly, people who care about their career might "need" publications in closed-access venues because of the prestige they carry (e.g. LICS in type theory or automata theory) – sadly this is how your work is judged nowadays…
(That is one of many reasons why I decided that I would never apply for any kind of permanent academic position: to escape this kind of incentive.)
Fabrizio Genovese (Apr 12 2020 at 23:19):Nguyễn Lê Thành Dũng said:
Antonin Delpeuch said:
One rule I have about communication is to boycott closed-access publications, both as an author and as a reviewer. We are lucky to work in a field where there are high quality open access options, so it is an easy rule to follow.
I think it's a bit more complicated than that. First of all, people here come from many different fields (this wide applicability is part of the beauty of categories :smile:). More importantly, people who care about their career might "need" publications in closed-access venues because of the prestige they carry (e.g. LICS in type theory or automata theory) – sadly this is how your work is judged nowadays…
(That is one of many reasons why I decided that I would never apply for any kind of permanent academic position: to escape this kind of incentive.)
I agree, but in cases such as LiCS I think it would be a good compromise to start your presentation talk by saying that you don't agree with their proceedings policy and that you suggest them to evaluate alternatives. If everyone would do this, LiCS would probably switch to open access proceedings within a year.
Fabrizio Genovese (Apr 12 2020 at 23:20):So a good policy, similar to what Jules suggests, is that if you submit at conferences that have closed-access proceedings, then you use part your presentation time to "denounce" this.
Lê Thành Dũng (Tito) Nguyễn (Apr 12 2020 at 23:58):I'm afraid that "within a year" might be too optimistic. The Cost of Knowledge petition was 8 years ago, and there are still a lot of papers appearing in Elsevier journals (at least in my field, I'm thinking of TCS, I&C, APAL…)
Lê Thành Dũng (Tito) Nguyễn (Apr 12 2020 at 23:59):(And now I feel old, I thought that was more recent before checking)
(=_=) (Apr 13 2020 at 04:00):This is related to my prior arguments about the feedback loop in academic publishing, particular in CS vs mathematics (actually, CS vs. the rest of the disciplines, but I digress).
The root of the problem, as I see it, is that an academic's career depends on frozen publications and not living bodies of work. It makes lives easier for the bean counters, but the act of fossilising papers in journals, so that you can have a record of work and priority, also has a tendency to break the feedback and dissemination system that advances knowledge within a field.
(=_=) (Apr 13 2020 at 04:04):There are ways to game this system, even if it is done innocently. I think it's been observed that some papers are published twice in CS: once in a conference proceedings, and once in a "proper" journal (sometimes with heavy revisions and/or additions), and usually with really similar titles. There is justification to do this, namely the authors want to provide a more complete and/or updated paper, but this is frowned upon in maths and other disciplines. I see this as a solution to work around the rigid nature of the publication system.
(=_=) (Apr 13 2020 at 04:19):There are reasons I can think of for wanting a publication system that works this way:
(=_=) (Apr 13 2020 at 04:28):However, digital repository systems with timestamps (e.g. arXiv) mostly settle the issue of priority, and makes obsolete the problem of scarce page space. If you add a version control system and a feedback system to that (e.g. Github), you can have a permanent record of the evolution of your work, without the need for an arbitrary cut-off date imposed by the journal system. The Github system also provides a means of ascertaining the impact of your work (stars, forks, etc.), and so does Google Scholar with its citation count. This seems to take care of the "nobody is using my work and I didn't know that" problem.
Gershom (Apr 13 2020 at 04:30):In at least some branches of CS, the situation is somewhat different. Conference proceedings are sufficiently refereed and selective that people mainly count conference papers. This is now reflected in there being special editions of journals that serve as conference proceedings. Standalone journal publication is a "luxury" usually undertaken by established researchers who want to collect and codify a research program for posterity in a longer, more synthetic fashion.
In general I think that most disciplines have moved towards a preprint culture for active research (perhaps with conference and workshop presentations serving some of that role), and then more rigorously vetted things for posterity / evaluation by committees.
An important issue is that math is very different in a certain regard. Refereeing a math paper is a much more serious and involved task, I imagine, since it carries a stronger stamp of "here is some delicately engineered thing that we believe is True and you can Rely On".
(=_=) (Apr 13 2020 at 04:33):Gershom: I'm looking beyond that. I'm really not interested in this "CS v Maths" business.
So what I'm saying is that we have all the means now to build a publication system that provides:
I think this would be a step forward from what we have right now, in CS, mathematics or any discipline.
Gershom (Apr 13 2020 at 04:43):The model you're suggesting already seems to be the "overly journals" model people are using/developing on arXiv. I think its a good idea, and I don't know what is really missing from it except I suppose time to develop and "sink in". It also doesn't seem to have room for conferences and workshops in the bits you're discussing, and I think those are nice and useful as well, of course.
(=_=) (Apr 13 2020 at 04:43):Gershom said:
An important issue is that math is very different in a certain regard. Refereeing a math paper is a much more serious and involved task, I imagine, since it carries a stronger stamp of "here is some delicately engineered thing that we believe is True and you can Rely On".
Er... no. See John's comment here:
I always feel computer scientists overestimate how much mathematicians think their work is perfect, or even care. It's not like anyone is gonna die if the proof of some esoteric conjecture later on turns out to have a flaw. If the result is important enough, people will keep looking at it and probably eventually find the mistake.
In really important subjects, like politics or economics, everything is and always has been a complete shitshow. It's sort of weird that people get freaked out by mistakes in proofs.
In fact, the open-source software movement was supposed to promote the production of "delicately engineered thing[s] [...] you can Rely On". I was proposing exactly that for maths in that topic. See what happened next.
(=_=) (Apr 13 2020 at 04:50):Gershom said:
The model you're suggesting already seems to be the "overly journals" model people are using/developing on arXiv. I think its a good idea, and I don't know what is really missing from it except I suppose time to develop and "sink in". It also doesn't seem to have room for conferences and workshops in the bits you're discussing, and I think those are nice and useful as well, of course.
No, it's not just the overlay journals, and it does provide for conferences and workshops. OpenReview.net is an excellent instance of an active open review system that enables conferences and workshops to create what are essentially overlay journals. I can imagine it being part of my model.
I don't like how journals work now, because they have a cut-off date: all feedback and peer review stops after the publication is accepted. This is like stopping patches for software once you've decided it's Version 1.0.0. As I've pointed out with the Voevodsky example, this breaks the feedback loop.
(=_=) (Apr 13 2020 at 04:56):One other thing: errata. In the current journal system, it's really hard to track down errata, unless someone has put in the effort to make things visible. With the kind of model I'm proposing, all you need to do is look at the log, created by the version control system, of issues raised and resolved.
Antonin Delpeuch (Apr 13 2020 at 07:13):Yeah I am really not claiming that my own way to boycott things is the only meaningful one. One thing that I find particularly effective is refusing to review - because when editors ask you for it, you have a chance to have a conversation with them.
Jules Hedges (Apr 13 2020 at 09:51):When you trace the problem to it's root, the actual limiting factor is hiring and promotion committees. Specifically, that they're usually made of dinosaurs. That's the one thing that's stopping everyone from just multilaterally switching completely to preprint culture
Jules Hedges (Apr 13 2020 at 09:55):If you want to be hired, a paper in Advances in Mathematics counts for more than a paper in TAC, even though Advances is Elsevier and so should be considered to not really exist by reasonable people
Jules Hedges (Apr 13 2020 at 09:58):This leads to what we in the business call a "social dilemma", ie. a real world situation with similar incentives to the prisoner's dilemma
Fabrizio Genovese (Apr 13 2020 at 10:48):Jules Hedges said:
This leads to what we in the business call a "social dilemma", ie. a real world situation with similar incentives to the prisoner's dilemma
And as in the prisoner's dilemma, actually you can obtain pareto efficiency with strong cooperation :P
Jules Hedges (Apr 13 2020 at 11:05):Yep, and the Cost of Knowledge pledge is exactly a mechanism of cooperation
Lê Thành Dũng (Tito) Nguyễn (Apr 13 2020 at 15:28):Gershom said:
Conference proceedings are sufficiently refereed and selective that people mainly count conference papers. This is now reflected in there being special editions of journals that serve as conference proceedings.
Are they really comparable to journals? The interaction that you can have with the reviewers is still severely bounded by the time pressure. I feel like this is completely backwards: instead of formally turning conference proceedings into journals, wouldn't it be better to move to publishing in actual journals?
Lê Thành Dũng (Tito) Nguyễn (Apr 13 2020 at 15:31):Rongmin Lu said:
One other thing: errata. In the current journal system, it's really hard to track down errata, unless someone has put in the effort to make things visible. With the kind of model I'm proposing, all you need to do is look at the log, created by the version control system, of issues raised and resolved.
Even for already published papers, it would be great to have some sort of centralized system for "bug reports"; currently, if you want to know whether a paper from the 1980s is correct, you have to ask knowledgeable people about issues that might be known in the "folklore" (or attempt to redo all proofs yourself).
Lê Thành Dũng (Tito) Nguyễn (Apr 13 2020 at 15:35):Jules Hedges said:
If you want to be hired, a paper in Advances in Mathematics counts for more than a paper in TAC, even though Advances is Elsevier and so should be considered to not really exist by reasonable people
BTW, does anyone know why TAC doesn't deliver DOIs for its papers? (Not that this has any relevance to scientific quality.)
(=_=) (Apr 13 2020 at 15:50):Jules Hedges said:
When you trace the problem to it's root, the actual limiting factor is hiring and promotion committees. Specifically, that they're usually made of dinosaurs. That's the one thing that's stopping everyone from just multilaterally switching completely to preprint culture
When you trace the problem to its roots, the answer was in something you retweeted earlier. The committees merely implement the dinosaur, they're not (entirely) made up of dinosaurs.
A nice article by @Alex_Danco that explains how academic research really works, how it became the way it is, and how Twitter might play an important role in disrupting it. Good read for non-academics to understand incentive structures of the institution. https://danco.substack.com/p/can-twitter-save-science
- hardmaru (@hardmaru) (=_=) (Apr 13 2020 at 15:55):And it's explained in more detail by Alex Danco in an earlier blogpost: it's "positional scarcity", which he claims is "one of the two primary kinds of emergent scarcity that we always, inevitably get in conditions of abundance".
https://alexdanco.com/2019/09/07/positional-scarcity/
(=_=) (Apr 13 2020 at 16:12):Antonin Delpeuch said:
Yeah I am really not claiming that my own way to boycott things is the only meaningful one. One thing that I find particularly effective is refusing to review - because when editors ask you for it, you have a chance to have a conversation with them.
And how did those conversations work out? Did they promise they'll "look into it", that they "understand"; or did the editorial board resign en masse?
Because if there wasn't meaningful action as a result of your conversation, your boycott wasn't a meaningful one. It is far easier to politely fob you off, and scores of others just like you, than to actually act in a way that would contribute towards achieving the aim of the Cost of Knowledge movement. I hope @Faré would forgive me for tagging him in this conversation, but he has written a pretty good essay about this.
(=_=) (Apr 13 2020 at 16:25):Nguyễn Lê Thành Dũng said:
Gershom said:
Conference proceedings are sufficiently refereed and selective that people mainly count conference papers. This is now reflected in there being special editions of journals that serve as conference proceedings.
Are they really comparable to journals? The interaction that you can have with the reviewers is still severely bounded by the time pressure.
They are traditionally superior to CS journals, because they have a lower acceptance rate and, ironically enough, because the non-cognoscenti will only accept journals in hiring and promotion decisions.
I agree that, as with journals, the interaction with reviewers is severely limited.
Fabrizio Genovese (Apr 13 2020 at 16:28):Rongmin Lu said:
Nguyễn Lê Thành Dũng said:
Gershom said:
Conference proceedings are sufficiently refereed and selective that people mainly count conference papers. This is now reflected in there being special editions of journals that serve as conference proceedings.
Are they really comparable to journals? The interaction that you can have with the reviewers is still severely bounded by the time pressure.
They are traditionally superior to CS journals, because they have a lower acceptance rate. I agree that, as with journals, the interaction with reviewers is severely limited.
Am I the only one that reasons along the lines of: If I can fit it in 12 pages then it goes to a conference, otherwise it goes to a journal?
Jules Hedges (Apr 13 2020 at 16:30):No, you're not the only one
(=_=) (Apr 13 2020 at 16:30):These days, you put out a 12-page conference paper, and then put the rest on the arXiv.
Fabrizio Genovese (Apr 13 2020 at 16:30):But seriously, we are all familiar with the butchering procedure one has to go through to submit to a conference, but sometimes you just can't!
Jules Hedges (Apr 13 2020 at 16:31):Conferences are the crack cocaine of publications.... your paper gets published fast, so it always feels like a good idea at the time, now 8 years later I have practically no journal publications
Fabrizio Genovese (Apr 13 2020 at 16:31):I mean, maybe it's just me, but even if I am ok in giving the proofs as an appendix, I am not ok in giving definitions as an appendix. So if what I am writing cannot be squeezed in 12 pages even when severely butchered and without proofs, I just don't bother
Fabrizio Genovese (Apr 13 2020 at 16:32):I guess one of the main reasons why conferences are so important is because if you submit to a conference it's easy to claim funding, esp. for PhD students. So you actually submit to see your peeps.
Lê Thành Dũng (Tito) Nguyễn (Apr 13 2020 at 16:32):Fabrizio Genovese said:
Am I the only one that reasons along the lines of: If I can fit it in 12 pages then it goes to a conference, otherwise it goes to a journal?
That's not how you do it: you're supposed to take your 50-page paper, put all the explanations and proofs in the appendices, and then you get 12 unreadable pages and people will be highly impressed by your intelligence :>
Jules Hedges (Apr 13 2020 at 16:32):In the last year I decided that I'm fed up of conferences and I'm going to focus on journal papers now. And then immediately found I have students of my own now who need conference papers
Fabrizio Genovese (Apr 13 2020 at 16:33):Nguyễn Lê Thành Dũng said:
Fabrizio Genovese said:
Am I the only one that reasons along the lines of: If I can fit it in 12 pages then it goes to a conference, otherwise it goes to a journal?
That's not how you do it: you're supposed to take your 50-page paper, put all the explanations and proofs in the appendices, and then you get 12 unreadable pages and people will be highly impressed by your intelligence :>
Yes, this is precisely what I try to avoid. Also because when I read papers like this I end up cursing and swearing so much against the authors that it's probably not worth it. My luck is already what it is without an evil eye on me. :P
(=_=) (Apr 13 2020 at 16:57):Nguyễn Lê Thành Dũng said:
currently, if you want to know whether a paper from the 1980s is correct, you have to ask knowledgeable people about issues that might be known in the "folklore" (or attempt to redo all proofs yourself).
Under the status quo, this is a feature, not a bug.
Antonin Delpeuch (Apr 13 2020 at 17:10):@Rongmin Lu I am not claiming my small boycott will change anything on its own. I have done other things: getting elected to the scientific board of my previous university and passing an open access policy there, building a platform to help researchers upload preprints to open archives, adding links to free to read versions of articles in Wikipedia articles (in a semi-automated way - we made tens of thousands of edits with that tool so far), crashing a presentation of an Elsevier representative in a university… None of these have achieved much in the grand scheme of things! The truth is that I don't actually care about the issue anymore, I am very happy not to stay in academia if it is not viable to only publish in OA. So my boycott is more of a matter of principle than a real aim to change anything.
Antonin Delpeuch (Apr 13 2020 at 17:10):@Nguyễn Lê Thành Dũng about TAC and DOIs, I have been in touch about that with the editorial board, and the previous managing editor was not keen for ideological reasons. Perhaps that will change after the recent switchover, perhaps not.
Jules Hedges (Apr 13 2020 at 17:12):There were real breakthroughs last year. University of California and Max Planck Society cancelled all their Elsevier subscriptions, for example. I suspect most of that is being driven from the bottom, librarians are listening (maybe indirectly) to their academics
Jules Hedges (Apr 13 2020 at 17:12):I think 2019 is the year that the fight against publishers really went mainstream
Fabrizio Genovese (Apr 13 2020 at 17:20):A big help is also given by libgen, btw. It makes subscriptions essentially useless, and is way more ergonomic than many uni library services (at least this was the case compared with the one Oxford, for instance)
Joachim Kock (Apr 13 2020 at 17:36):The problem with proceedings papers as a publishing form is not only the page limits but also the deadlines. You are forced to publish in yearly 15-page installments, instead of releasing your work when it is ready and has been ironed out, and instead of letting the material dictate the form and length of the exposition.
Antonin Delpeuch (Apr 13 2020 at 17:43):( @Nguyễn Lê Thành Dũng still about DOIs for TAC, if you also care about the issue, make sure you inquire about it to the new managing editor - if this is something that is frequently requested by prospective authors, they might do something about it…)
Lê Thành Dũng (Tito) Nguyễn (Apr 13 2020 at 17:45):I'll be sure to keep that in mind if I ever have something that I could submit to TAC :-)
John Baez (Apr 13 2020 at 18:30):Jules Hedges said:
When you trace the problem to it's root, the actual limiting factor is hiring and promotion committees. Specifically, that they're usually made of dinosaurs. That's the one thing that's stopping everyone from just multilaterally switching completely to preprint culture.
Maybe in some universities a hiring and promotion committee in one department would have the unilateral right to change how they did things... but it sure ain't true at the University of California, and probably not at most other big universities: there are rules governing these things, and all decisions get reviewed by other committees which must follow rules of their own!
If the math department just decided one day to stop caring about publication and promote based on preprints, all our decisions would be knocked down by CAP - the Committee on Academic Personnel. And if they tried to revolt and stop following the usual rules, their decisions would be knocked down by the dean of the relevant college. And if they tried to revolt... well, you see what I mean.
John Baez (Apr 13 2020 at 18:33):Rongmin Lu said:
When you trace the problem to its roots, the answer was in something you retweeted earlier. The committees merely implement the dinosaur, they're not (entirely) made up of dinosaurs.
They're definitely not entirely made up of dinosaurs. (Am I a dinosaur?) The real dinosaurs are the rules governing various institutions. These rules can be changed. but only by collective action taken by lots of persistent people, including people in positions of authority.
These rules do change - it's not impossible! It takes lots of committee meetings.
Faré (Apr 13 2020 at 21:10):@Rongmin Lu Sorry, I wrote a pretty good essay about what? Do you mean my essay about "Passivism" vs Activism? That's quite far from Category Theory :-)
PS: http://fare.tunes.org/liberty/passivism.html
Faré (Apr 13 2020 at 21:24):At my last accepted peer-reviewed conference paper, the ACM demanded that I either assign them copyright or pay them $900 so their lawyers would "protect" my rights, or else I wouldn't be published. I told them a resounding "fuck you". But I could afford it.
Faré (Apr 13 2020 at 22:03):@John Baez In hiring decisions, do you count positively articles that were accepted at a conference but not published by the ACM (or other publisher) because the author refused to submit to the publisher racket?
John Baez (Apr 13 2020 at 22:19):I'm in a math department. If you publish with the ACM you don't get a job. :upside_down:
John Baez (Apr 13 2020 at 22:20):In our department, like most math departments, conference proceedings are treated sort of like trash.
John Baez (Apr 13 2020 at 22:20):Like "well, he has some conference proceedings, but what did he really publish?"
John Baez (Apr 13 2020 at 22:20):It's very different from computer science.
John Baez (Apr 13 2020 at 22:22):In math people just don't publish serious stuff in conference proceedings.
Lê Thành Dũng (Tito) Nguyễn (Apr 14 2020 at 00:12):I suppose that's a big reason for the existence of the Proceedings of the ACM in Programming Languages journal (http://sigplan.org/PACMPL/)
(=_=) (Apr 14 2020 at 02:09):Nguyễn Lê Thành Dũng said:
I suppose that's a big reason for the existence of the Proceedings of the ACM in Programming Languages journal (http://sigplan.org/PACMPL/)
That's a big reason for the existence of CS journals (yes, plural) in general.
The AMS has the Proceedings of the AMS as well. PAMS is definitely not trash: see here, here, or here.
But it's not "really" a conference proceedings, it's a proper journal. :triumph: :sweat_smile:
(=_=) (Apr 14 2020 at 02:11):John Baez said:
In math people just don't publish serious stuff in conference proceedings.
They used to, just not now. There are so many old papers I've tracked down only to lose them, because they were published in some obscure conference proceedings that's not available anywhere anymore.
John Baez (Apr 14 2020 at 02:12):Proceedings of the AMS is indeed nothing like a conference proceedings. It's just a journal: the AMS has two main ones, the Transactions and the Proceedings. I've only published in the Transactions. I forget the difference - something to do with the length of the papers.
There's also the Proceedings of the National Academy of Sciences and the grand-daddy of them all, the Proceedings of the Royal Society, neither of which are conference proceedings.
(=_=) (Apr 14 2020 at 02:12):Faré said:
Do you mean my essay about "Passivism" vs Activism? That's quite far from Category Theory :-)
See the topic title. Not exactly CT, is it? :sweat_smile:
(=_=) (Apr 14 2020 at 02:14):John Baez said:
Proceedings of the AMS is indeed nothing like a conference proceedings. It's just a journal: the AMS has two main ones, the Transactions and the Proceedings. I've only published in the Transactions. I forget the difference - something to do with the length of the papers.
And CS conference proceedings are just annual versions of journals, aka "annuals".
(=_=) (Apr 14 2020 at 02:20):Joachim Kock said:
The problem with proceedings papers as a publishing form is not only the page limits but also the deadlines. You are forced to publish in yearly 15-page installments, instead of releasing your work when it is ready and has been ironed out, and instead of letting the material dictate the form and length of the exposition.
That's a feature, not a bug. :upside_down:
(=_=) (Apr 14 2020 at 02:26):Antonin Delpeuch said:
None of these have achieved much in the grand scheme of things!
Because the time isn't ripe for such changes, and it will only be right once the environment changes in such a way that the new way of doing things is demonstrably and overwhelmingly better than the old way of doing things. See Faré's essay that he mentioned here.
The truth is that I don't actually care about the issue anymore, I am very happy not to stay in academia if it is not viable to only publish in OA. So my boycott is more of a matter of principle than a real aim to change anything.
I guess I have mixed feelings about that. On the one hand, it is sad that this has happened. On the other hand, it's probably the right decision for you as an individual.
(=_=) (Apr 14 2020 at 02:31):Jules Hedges said:
There were real breakthroughs last year. University of California and Max Planck Society cancelled all their Elsevier subscriptions, for example. I suspect most of that is being driven from the bottom, librarians are listening (maybe indirectly) to their academics
Most of the decisions are driven by budget concerns, and quite a bit of that is due to the growing administrative apparatus that is taking up more and more of an institution's budget. I reckon librarians are confronted both with budget constraints and the demands of academics, and as the budget becomes more and more curtailed, the cheaper alternatives would become more appealing.
(=_=) (Apr 14 2020 at 02:35):John Baez said:
These rules can be changed. but only by collective action taken by lots of persistent people, including people in positions of authority.
These rules do change - it's not impossible! It takes lots of committee meetings.
Yay! More committee meetings! :tada: :rolling_on_the_floor_laughing:
Lê Thành Dũng (Tito) Nguyễn (Apr 14 2020 at 10:36):Rongmin Lu said:
That's a big reason for the existence of CS journals (yes, plural) in general.
I don't know many other journals that are explicitly designed to contain only conference proceedings. The only point seems to be to make its status/prestige legible outside the community (so as John said this is nothing like Proc. AMS).
Lê Thành Dũng (Tito) Nguyễn (Apr 14 2020 at 10:39):Perhaps I should quote directly from the source:
https://www.acm.org/publications/pacm/introducing-pacm
Proceedings of the ACM on X (PACM) is a new journal series with first issues planned for 2017. This series is suitable for those ACM SIG-sponsored conferences that adapt their review processes to be comparable to those of journals. It has been launched in recognition of the fact that conference-centric publishing disadvantages the CS community with respect to other scientific disciplines when competing with researchers from other disciplines for top science awards and career progression, and the fact that top ACM conferences have demonstrated high quality and high impact on the field.
Lê Thành Dũng (Tito) Nguyễn (Apr 14 2020 at 10:46):(Though I suspect that the name is deliberately chosen to remind one of both conference proceedings (for those who publish there) and the other Proceedings of <whatever learned society> (for non-CS colleagues and bureaucrats))
(=_=) (Apr 14 2020 at 15:26):Nguyễn Lê Thành Dũng said:
Rongmin Lu said:
That's a big reason for the existence of CS journals (yes, plural) in general.
I don't know many other journals that are explicitly designed to contain only conference proceedings. The only point seems to be to make its status/prestige legible outside the community (so as John said this is nothing like Proc. AMS).
Yeah, that was my point. There are CS journals, and their existence seems to be about providing status/prestige that's legible outside the CS community, as is explicitly stated in the passage you've quoted.
Lê Thành Dũng (Tito) Nguyễn (Apr 14 2020 at 20:18):Are you making this claim about old journals such as Theoretical Computer Science, Information & Computation, etc? I'd find it hard to believe that this was the initial impetus for their creation as for the Proceedings of the ACM series.
Lê Thành Dũng (Tito) Nguyễn (Apr 14 2020 at 20:21):Related, but getting back to the open access topic: it seems that there are a lot more OA conferences (thanks to LIPIcs) than journals. So if you're sick of conferences but don't want to sign away your copyright to Elsevier / Springer / ACM / CUP, you're left with few options (for computational logic you've got LMCS, in some cases TAC, and… ?)
(=_=) (Apr 15 2020 at 05:09):The point of having journals and other selective periodicals is to confer visible legitimacy/status/prestige through curation, and if that legitimacy isn't visible outside of a certain community, then they have failed. PACM is providing a curation service that explicitly addresses a problem peculiar to the CS community, but the problem is only peculiar in form, not in substance.
Henry Story (May 17 2020 at 09:54):I was pointed here from a thread I started on programming in applied CT. I'll repeat the point made there here, as it is really about Open Source and peer review.
Henry Story said:
There is a major story in the UK breaking now about the code behind the Imperial College Model for Covid-19 which I also wrote up in Code, Models and Covid-19, where I put it in a philosophical context with references to mathematical work too (eg. Agda, Coq, Idris, Scala).
Fabrizio Genovese (May 17 2020 at 12:45):This is a surely important topic, but I think it's worth saying that it is not as important as we believe it is today
Fabrizio Genovese (May 17 2020 at 12:46):What I mean is that today the biggest part of the tech business is done with data, and we do not have many "open data avocates" around. Like, for many people the data issue is not really a thing. This has paradoxical applications. For instance, you can have companies like Google that open part of their sourcecode, which is tho absolutely useless for everyone else
(=_=) (May 17 2020 at 12:47):Fabrizio Genovese said:
This is a surely important topic, but I think it's worth saying that it is not as important as we believe it is today
Well, I think it might be important for the future of ACT, or at least Jules' grant application:
At least one """good""" thing comes out of this -I'm definitely going to refer to this whole thing in some future grant application
Fabrizio Genovese (May 17 2020 at 12:47):So there's this famous effect in machine learning where a simpler algorithm can outperform a very complicated one when the dataset is huge. This is because the simpler algo is more flexible, and as such more difficult to train, but better suited at fitting the data when the dataset is big.
Fabrizio Genovese (May 17 2020 at 12:48):The result is that I can claim to do open source stuff, and open up algos to the community, which are worth nothing tho unless you have my datasets, which I will guard in any way I can
(=_=) (May 17 2020 at 12:48):Part of the problem is that the original code was written by a non-professional software developer, i.e. the scientist himself. https://twitter.com/neil_ferguson/status/1241835454707699713
I’m conscious that lots of people would like to see and run the pandemic simulation code we are using to model control measures against COVID-19. To explain the background - I wrote the code (thousands of lines of undocumented C) 13+ years ago to model flu pandemics...
- neil_ferguson (@neil_ferguson) Fabrizio Genovese (May 17 2020 at 12:48):So what I'm trying to say is that open source is for sure an important issue, but in 2020 it should really be seen through the lenses of "open source + open data".
(=_=) (May 17 2020 at 12:49):The problem in this case is that the code hasn't been open source for 13+ years, let alone the data, so it has undergone no peer review.
Fabrizio Genovese (May 17 2020 at 12:50):Yes, I am not referring to this case in particular, but at the topic, which is "open source"
Fabrizio Genovese (May 17 2020 at 12:50):I guess that one of the reasons why I'm saying this is that there is this common misconception now that "open source" is basically the ethic choice in compsci. What I'm saying is that you can do open source and still be not ethic at all, in a variety of different ways
(=_=) (May 17 2020 at 12:53):Fabrizio Genovese said:
I guess that one of the reasons why I'm saying this is that there is this common misconception now that "open source" is basically the ethic choice in compsci. What I'm saying is that you can do open source and still be not ethic at all, in a variety of different ways
Open source allows for peer review, which removes one particular ethical problem, i.e. whether the code is properly scrutinised. This is especially important in this particular case, when the code in question is produced in academia and then used in public policy.
(=_=) (May 17 2020 at 12:54):Yes, there are a variety of different ways in which open source code can be used for non-ethical means, but this is irrelevant to the matter at hand.
(=_=) (May 17 2020 at 12:59):Fabrizio Genovese said:
So what I'm trying to say is that open source is for sure an important issue, but in 2020 it should really be seen through the lenses of "open source + open data".
I think, in the context of this Zulip topic, that "source" includes "data" in its scope. After all, this topic was initially about academic peer review, and in academia, there is a trend towards asking for datasets to be made open as well.
Antonin Delpeuch (May 18 2020 at 18:43):By the way, we just launched a public pledge: "No free view? No review!"
Antonin Delpeuch (May 18 2020 at 18:43):https://nofreeviewnoreview.org/
Fabrizio Genovese (May 18 2020 at 18:52):Antonin Delpeuch said:
This is great. I'm signign instantly.
Jules Hedges (May 18 2020 at 19:08):So, I intentionally never signed Cost of Knowledge, because I wanted to keep myself free to apply common sense. Now I have to go through all that thought process again
Jules Hedges (May 18 2020 at 19:12):One obvious problem springs to mind. (@Antonin Delpeuch probably should have thought of this when you asked me for feedback before, but I never got round to it sorry) --- The most common case in cs is to be asked to review for a conference, and there's no way to tell who's going to publish the proceedings
Jules Hedges (May 18 2020 at 19:14):Case in point: I had a paper accepted in MSFP (mathematically structured functional programming, a small workshop) in 2014 and it was published in EPTCS, which is open access. Then I had another paper accepted there in 2018 and it was published in ENTCS, which is Elsevier. There was no way to know this when I submitted!
Jules Hedges (May 18 2020 at 19:15):Since I hadn't signed Cost of Knowledge it didn't matter, just make sure it's freely avaiable and job's a good'un. But if I'd signed it, officially I'd have to withdraw it in order to not break the pledge
Antonin Delpeuch (May 18 2020 at 19:16):In general conferences announce how the proceedings will be published - if not, just ask :)
Jules Hedges (May 18 2020 at 19:17):Yeah, I know that now
Antonin Delpeuch (May 18 2020 at 19:18):None of the boycott pledges I know are retroactive, so that should not prevent you from signing up to them
Jules Hedges (May 18 2020 at 19:20):Another weird case I'm aware of: LiCS 2018 proceedings are published by the ACM and closed access, and LiCS 2019 proceedings are not published (!)
Kenji Maillard (May 18 2020 at 19:21):so LiCS 2019 are officially not accessible at all ?
Jules Hedges (May 18 2020 at 19:22):As far as I can tell! The LiCS 2018 site has a "proceedings" link and the 2019 site doesn't, and google returns nothing
Jules Hedges (May 18 2020 at 19:22):This is taking the semantics of "published" a bit over-literally. In reality a LiCS paper is a LiCS paper, I don't think anyone would actually care about the difference
Lê Thành Dũng (Tito) Nguyễn (May 18 2020 at 19:33):(BTW, if I recall correctly ENTCS is actually gold open access with a rather small article processing charge despite being owned by Elsevier)
Lê Thành Dũng (Tito) Nguyễn (May 18 2020 at 19:34):For LICS 2019: https://ieeexplore.ieee.org/xpl/conhome/8765678/proceeding
Lê Thành Dũng (Tito) Nguyễn (May 18 2020 at 19:35):Since it's an ACM/IEEE co-sponsored conference, they alternate between ACM proceedings on even years and IEEE proceedings on odd years
Jules Hedges (May 18 2020 at 20:41):Nguyễn Lê Thành Dũng said:
For LICS 2019: https://ieeexplore.ieee.org/xpl/conhome/8765678/proceeding
Thanks! Somehow I never found that despite looking several times
John Baez (May 18 2020 at 21:07):I don't think anyone actually comes to your door and kneecaps you if you take one of these pledges and for some (hopefully good) reason breaks it at some point.