Not signed in (Sign In)

Not signed in

Want to take part in these discussions? Sign in if you have an account, or apply for one below

  • Sign in using OpenID

Discussion Tag Cloud

Vanilla 1.1.10 is a product of Lussumo. More Information: Documentation, Community Support.

Welcome to Math2.0
If you want to take part in these discussions either sign in now (if you have an account), apply for one now (if you don't).
    • CommentRowNumber1.
    • CommentAuthorJohn Baez
    • CommentTimeFeb 16th 2012
    • (edited Feb 16th 2012)

    Chris Lee and Marc Harper want to implement some of the ideas described here:

    I’m going to help them, and we’d like to know if you want to help, too.

    Lee and Harper are good at programming, but their plan is to start something quickly and improve it later as people start to use it and give feedback. So, your feedback is welcome now… but the focus should be on keeping it simple and flexible, not on detailed fine-tuning or fancy features.

    Also, if you dislike the general idea of this system, I urge you to spend your energy developing a different one yourself. Rather than seeking universal consenus, we need a healthy dissensus in which different groups of people develop different systems, so we can see which ones work.

    • CommentRowNumber2.
    • CommentAuthorJohn Baez
    • CommentTimeFeb 16th 2012

    Chris writes:

    I propose that the review site would present each paper with a basic tabbed interface, initially with just two tabs, Recommendations and Discussion. The main difference is that whereas recommendations are automatically forwarded to the recommender’s subscribers, discussion comments are not; they are simply displayed on the Discussion tab for the paper. The Recommendations tab would show who recommended the paper, for what area(s), and what they said about it. The Discussions tab would show the usual threaded discussion interface that everyone is used to.

    I. The Recommendations tab would show the following simple interface that lets you view and make recommendations. It should be designed to look like the arXiv main page for a paper (e.g. sidebox for accessing PDF, PostScript etc. for the paper), but with the main content of the page shown in the following order (see explanatory notes below on the interactive elements):

    • paper title, authors
    • [+] abstract Flagging papers that interest you will help this system predict other papers that may interest you. Subscribing to people who recommended this paper will alert you to other papers they recommend. Writing recommendations that other people find useful will attract them to subscribe to receive your future recommendations and papers.

    • [STAR] Rated must-read by Leonard Euler, Terence Tao and 37 others [+]

    • If this paper interests you, flag the group(s) you would recommend read it:

    [ ] alg-geom: rec’d by Maxim Kontsevich, Pierre Deligne and 21 others [+]

    [ ] abelian varieties: rec’d by Herbert Spencer… [+]

    Add a topic-group: _____________

    • News & Views

    [+] Pierre Deligne: The most exciting paper since…

    [+] Qiaochu Yuan…

    [+] 5 others

    Tell the world! Write a News & Views:

    Interface Notes:

    • clicking +/- sign shows or hides the text of the abstract.

    • clicking the STAR icon toggles whether you rate the paper must-read or not; clicking the +/- sign toggles viewing of the list of other recommenders.

    • the list of candidate topic-groups would be the superset of what other people (including the authors) suggested as groups that would be interested in this paper, and the groups the user belongs to.

    • Add a topic group would let the user enter text to search the database of existing topic-groups and pick one or alternatively create a completely new group.

    • “News & Views” is just a provisional title, suggestive of alerting others to something new and interesting and of communicating your opinion about what matters. “News & Views” is what Nature calls these kind of highlight pieces.

    • A News & Views item is implemented essentially a discussion item that gets forwarded to its authors subscribers, who in turn can recommend it to their subscribers, and so on. (in this sense it is like a publication in its own right, but of course we won’t allow people to write News & Views on a News & Views item!). As a discussion item, it can be commented on by others just like any discussion item).

    • expanding a News & Views will show its text, as well as a link for viewing comments / commenting on it in the Discussion tab.

    • you can add a News & Views by giving a DOI or URL for something published elsewhere (e.g. on your blog), or by simply entering text into a textbox here.

    [cont.]

    • CommentRowNumber3.
    • CommentAuthorJohn Baez
    • CommentTimeFeb 16th 2012
    • (edited Feb 16th 2012)

    Chris writes:

    II. The Discussion tab would be a generic threaded discussion for this paper. As long as people can add new topics for discussion and comment on existing topics, it will be adequate for an initial release. Over time it would add peer review features (e.g. you could flag a discussion item as raising doubt about the validity of a major claim in the paper). However that is not needed at first.

    III. User Registration:

    • users would be asked to register with email address that matches their arXiv account, and would then receive an email for activating their account. This way we link each user to their papers etc. in arXiv.

    • to seed the system for recommending papers to them, new users would be asked to list recent papers they considered must-read for their own work, and / or other researchers whose interests they consider to be most similar to their own.

    • users would be asked to select topic-groups that fit their research interests. Initial topics can be suggested by the system based on the papers and people they listed, e.g. if any of those people already listed some topic-groups, offer those as options. Of if any other people listed those same papers or people, offer the topic-groups that those people listed. Of course the user can add his own topic-group terms.

    • the site would follow the Amazon model of remembering the user (i.e. based on their last login to this site on this computer, show their personal recommendations), but authenticate them (i.e. ask for password) if they want to publish a recommendation.

    IV. I suggest the look and feel of the site draw from two main sources:

    • mainly, make it look like arXiv ;

    • where appropriate, copy the clean, simple model of Google Code, e.g. http://code.google.com/p/pygr/.

    This is just a first attempt; let me know your thoughts and questions…

    • CommentRowNumber4.
    • CommentAuthorJohn Baez
    • CommentTimeFeb 16th 2012
    • (edited Feb 16th 2012)

    Chris writes:

    Hi John,

    sure, feel free to communicate this on Math 2.0. The sooner other people start giving feedback, the sooner we start learning what works for people… In addition to asking people for feedback, you might also ask whether anyone wants to join us in aspects of the actual work, e.g. (in increasing order of effort):

    • trying out the system and reporting problems or requests;

    • working on the graphical esthetics of the site (i.e. images and CSS);

    • working on front-end code (HTML templates) or back-end code (e.g. database).

    Marc and I want to throw together a prototype ASAP and would love to have people to talk with and work with in this effort.

    • CommentRowNumber5.
    • CommentAuthorScott Morrison
    • CommentTimeFeb 16th 2012

    Yay! I highly approve. I think I’m too busy to consider actually helping write code, but I love being opinionated about user interfaces, so I’ll try to help out with alpha testing.

    Anything that gives me a better thing to look at each morning that the daily arXiv email counts as “useful at the point of use”, and I think this could quickly fit the bill.

    • CommentRowNumber6.
    • CommentAuthorJohn Baez
    • CommentTimeFeb 16th 2012

    Thanks, Scott. We’ll count you as an alpha tester.

    Anyone who can write code?

    • CommentRowNumber7.
    • CommentAuthorAndrew Stacey
    • CommentTimeFeb 16th 2012

    I do not highly approve - in fact, I highly disapprove! But I completely agree with the last paragraph in John’s first post that we need to try things out and I’d certainly approach it with an open mind - I’ve been known to change my mind before. So I’d be happy to test out the system. I can also attest to Scott’s being opinionated about user interfaces!

    I feel I should say that I already have what I consider a fairly good system for working with the arXiv. My reference database grabs a list of the new stuff on the arXiv each day and then I go through it (through an interface that looks pretty much like the standard arXiv list, but with a few little bits added) and decide which ones to add to my database - I can add keywords too if I like. Then I also add the reference to one of my nLab webs. All of these are cross-linked.

    My reference database is behind a password so you can’t try it out, but here’s a screenshot:

    RefBase arXiv integration

    The nLab web is http://ncatlab.org/lspace (see, I do read more than just Douglas Adams), a typical entry is: http://ncatlab.org/lspace/show/Algebra%2BHomotopy%3DOperad (this was from today’s list). The idea is that if ever I want to make notes on a particular article, then I’ll do so there. The fact that I haven’t says more about me than about the system!

    (Hmm, that last remark’s gotten me thinking - but that might be something for a different thread.)

    Any chance of Chris and Marc joining us here to discuss it?

  1. I’ll gladly help, too. After Scott is in favor, Andrew sceptical, I think I’ll be undedcided for now ;)

    • CommentRowNumber9.
    • CommentAuthorKevin Walker
    • CommentTimeFeb 16th 2012
    I would be happy to help out. I might be able to help with coding, if that sort of help is desired.
    • CommentRowNumber10.
    • CommentAuthorMarc Harper
    • CommentTimeFeb 17th 2012
    Marc here. Chris and I have started repositories on github (https://github.com/cjlee112/spnet and https://github.com/marcharper/spnet) to start building a prototype. If anyone wants to help, please join in and contribute! Our initial goals are commenting and recommendation systems for arxiv papers.
  2. Marc, Chris, John great initiative ! Hope to be able to help.

    @Andrew, I am sorry I did not quite understand why do you think it is not good idea ?

    May I ask several questions:

    1) the biggest concern - if site would be created, can we built the community who will use it ?
    (Personally I greatly want to have site where I can read comments and leave them).

    2) is it open source project ?

    3) I think that there should NOT be automatic transfer of the paper from arXiv to the site - I mean,
    some people do NOT want their papers to be publicly discussed. (There are such comments at Gowers blog).
    I think that only if the author wants he can send his paper to the SITE.

    4) I am sure that comments should be allowed to be written anonymously without registration.
    Otherwise people will be afraid to criticize papers and other reasons...
    I think small moderation and voting comments would be enough to keep the order - look at Mathoverflow. (Personally I use real name everywhere, but you can see many people on MO
    making good contributions under nicks).
    • CommentRowNumber12.
    • CommentAuthorMarc Harper
    • CommentTimeFeb 18th 2012
    > 1) the biggest concern - if site would be created, can we built the community who will use it ?
    (Personally I greatly want to have site where I can read comments and leave them).

    There is no plan to limit who can use the site, and we hope that communities will grow around common areas of interests!

    > 2) is it open source project ?

    Yes it is open source.

    > 3) I think that there should NOT be automatic transfer of the paper from arXiv to the site - I mean, some people do NOT want their papers to be publicly discussed. (There are such comments at Gowers blog). I think that only if the author wants he can send his paper to the SITE.

    While we have not been planning to have papers automatically appear from the arxiv, I do not see the point of preventing discussion on any particular paper (from the arxiv or otherwise). My personal opinion is that the idea that one would publish a paper but not want there to exist public discussions about it seems perverse -- why publish something if not to have others read and discuss it? Trying to limit discussion seems rather disingenuous regarding the validity and legitimacy of the work.

    It also does not seem possible. If people cannot discuss a paper on the proposed site they will just go elsewhere -- there are dozens of places now that a paper can be publically discussed, such as on Google+, sites like reddit, community blogs like the n-cafe, and on personal blogs. I do not see any realistic way to prevent public internet discussion of *anything*. Moreover, if we only allow authors to submit their own papers, we would have to put significant effort into verifying identity of users/submitters, which is somewhat at odds with having "anonymous" users. This would also require (at least one? or all?) author(s) to be users of the site before discussion could be allowed.

    I know we are not going to be able to please everyone, of course, but hopefully we can make users feel comfortable using the site while also maintaining high quality discussions. I do not want to limit the utility of the site just to please a small minority (if that is the case here) of users that have another option (not uploading to the arxiv, or not publishing at all).
    • CommentRowNumber13.
    • CommentAuthorMarc Harper
    • CommentTimeFeb 18th 2012
    > 4) I am sure that comments should be allowed to be written anonymously without registration.
    Otherwise people will be afraid to criticize papers and other reasons...
    I think small moderation and voting comments would be enough to keep the order - look at Mathoverflow. (Personally I use real name everywhere, but you can see many people on MO making good contributions under nicks).

    While I think the mathoverflow model works in many ways, the type of discussions at MO are of a different type than what is proposed for this site. It is one thing to leave a negative comment in an open discussion of a possible theorem, and quite another to leave a negative comment on a published work in plain view of many people looking for information on that work. Unregistered anonymous comments would make the SPNet extremely vulnerable to tampering and gaming and require a huge amount of moderation.

    There are other issues as well -- what if a user with a nick rather than their real name wants to upload a paper? Wouldn't they have to 'out' themselves in the process or else have multiple accounts? I really don't want to allow people to have accounts where they promote their own work and another where they try to discredit their detractors anonymously.

    Regarding moderation of comments, I think that the community should handle it in general (perhaps even moreso than the mathoverflow model), but I don't know if comment voting is really the way to go. Community consensus != correctness, especially if "anonymous" downvoting is allowed -- this could also really open a system to gaming (or require huge amounts of monitoring). What is to prevent a coordinated group of users acting anonymously to suppress a legitimate work, or to suppress legitimate comments (or to upvote illegitimate comments)? I also have no desire to moderate comments for legitimacy or validity and can easily imagine myself being asked to arbitrate a discussion far outside of my expertise. Regardless of my opinions on the matter, I am open to any comments that others have regarding how to set up a useful and robust comment system.

    Note that we plan to eventually have an issue tracking system for papers going through "peer review" so that the author can interact with reviewers to resolve issues in a productive way. This provides a better format for serious and civil discussions of legitimacy of aspects of the paper.

    Anyway, user identity is being actively discussed. I think that once the site has enough users and data that it could probably automatically determine users doing some of the things I expressed concern for above, and in fact I have argued in favor of allowing so-called anonymous (yet registered) users if the activities of users can be "weighted" somehow based on the quality of their participation. The reality is that reputation matters, and if someone important recommends your paper that may matter to a lot of other people moreso than a recommendation by an anonymous user. I am very much opposed to unregistered users and I see the value in requiring verified identities for users. If others have opinions on this topic I would like to hear them.
  3. @Marc Thank You for detailed answers.

    Concerning (1) - I asked the question " Will YOU contribute comments to papers ? " on this site
    let us look...

    Concerning (2). (Should author allow his paper to be discussed or not).
    I absolutely agree that "one would publish a paper but not want there to exist public discussions about it seems perverse". But "people are strange" and we should take this into account.

    "If people cannot discuss a paper on the proposed site they will just go elsewhere -- "
    absolutely agree, but this is problem of "elsewhere" sites,
    if one comments "this paper is junk" on the site - the author of paper
    will have definitely bad opinion about the site as a whole and it will stop him to contribute...

    Concerning (4) - "anonymous" actions

    At least is my understanding correct he "anonymous" can view comments ?

    a) "anonymous" voting - not allowed like MO - for sure
    b) Okay I agree that may be for the first years "anonymous" forbidden to comment - is more safe and easy, but for future this does not seem to be optimal ...
    c) Soft version - may be "anonymous" comments will not appear immediately, but only after moderators/authors approval or may be even correction

    ------
    You write " what if a user with a nick rather than their real name wants to upload a paper? "
    what means "upload paper" ? I thought that papers will be in arXiv , is it correct ? what means "upload" then ?
    • CommentRowNumber15.
    • CommentAuthorHenry Cohn
    • CommentTimeFeb 18th 2012
    >My personal opinion is that the idea that one would publish a paper but not want there to exist public discussions about it seems perverse

    If the site becomes really successful, that may be true, but for example I'd prefer not to have automatically generated pages for my papers on it, because I think the hassle will greatly outweigh the likely benefits.

    As you mention, there are already any number of places one could discuss papers online, but people seem to have practically no desire to do so, and the fraction of papers that get discussed publicly online is very nearly 0. Many people agree that discussions would be great in the abstract, but they don't seem to have anything to say that might motivate them to start such a discussion. Having a dedicated web site would presumably lead to more discussions, but even so I expect there would be interesting, substantive discussions for just a tiny fraction of all papers.

    However, that doesn't mean the site wouldn't get used. There would be crackpots, amateur kibitzers, lazy students, academic spammers ("if you liked this paper, you're going to love my latest paper!"), people starting tangential discussions that aren't really about a given paper, etc.

    So what I'm afraid of is that this could end up creating a low-quality, somewhat embarrassing discussion site associated with each of my papers, which I'd then have to monitor.

    Incidentally, I think there actually is genuine demand for substantive mathematical discussions online, but they can rarely be characterized as being about one particular paper.
    • CommentRowNumber16.
    • CommentAuthorDavid Speyer
    • CommentTimeFeb 18th 2012
    I agree with everything Henry says.

    In particular I want to emphasize that Internet communities don't moderate themselves.

    I think it is a terrible idea to set up any sort of discussion forum without people prepared to serve as moderators, and the amount of work to moderate a discussion thread for every paper on the arXiv would be huge. Several times, people on meta.MO have requested that MO set up an associated pure discussion forum, and the organizers have always responded that they would be glad to do the technical work if some experienced users would volunteer to moderate. There haven't been any takers. I think we're going to see the same thing here.
  4. @David Speyer 16
    "and the amount of work to moderate a discussion thread for every paper on the arXiv would be huge "

    It is not obvious for me... who cares about what mathematicians write in their papers ?
    "amateur kibitzers, lazy students, academic spammers" have much better targets than spam
    on the papers which almost no one reads (remember Gowers post - average number of readers for math paper is less than one).
    • CommentRowNumber18.
    • CommentAuthorScott Morrison
    • CommentTimeFeb 18th 2012
    • (edited Feb 18th 2012)

    @marcharper,

    I think it’s actually a good idea to invest more effort in identifying users, and respecting their wishes. Although thoroughly contrary to current trends on the internet, I think a site like this would benefit very significantly from having a very “professional” feel. (Not the interface, rather, the expectations on the community.) I think MathOverflow has benefitted significantly from our relative intolerance of incivility (on the main site, at least), and the effort we’ve put into encouraging people to use their real names, to establish credible identities.

    A big problem you’re going to have with the site you propose is “throwaway” comments, poorly thought out, or simply downright rudeness. If everyone knows who everyone else is, this is hugely reduced. People put far more effort into thinking carefully about what they’re saying. (As a little example, whenever now I need to write something in a forum that is potentially ’dangerous’, I steal Matt Emerton’s habit and actually sign my name explicitly (best regards, SM, or yours sincerely, SM, depending on the occasion. This one extra step ensures that I also double check my text for tone and potential misunderstandings, simply because I really am signing my name to that text!)

    I would strongly encourage a system in which logging in requires you to match your identity with an arXiv author. (Yes, you can’t comment if you don’t have a paper.) I’m not actually sure that only the author should be able to open up comments on a paper, but I think it would be an excellent idea to allow authors to withdraw their paper from your site. It’s a matter of basic respect, and silly to say “but they shouldn’t use the arXiv if they don’t want their paper discussed here”. (It is of course very reasonable to say that, if you leave our the “here”.)

    • CommentRowNumber19.
    • CommentAuthorScott Morrison
    • CommentTimeFeb 18th 2012

    @Alexander Chervov, having seen MathOverflow, I think that the amount of moderation required (or perhaps more relevantly, the optimal amount) is vastly underestimated by most.

    I’m not saying that the “official” moderators have a huge workload. Indeed, I’m lying on a beach in the Seychelles most days. At MathOverflow the software enables a rather large class of people to manage this moderation work, and the “official” moderators only need to step in in corner cases where the default system breaks down for some reason.

  5. Scott Morrison wrote "...but I think it would be an excellent idea to allow authors to withdraw their paper from your site. It's a matter of basic respect ... "

    I think it might be good solution. I would add the two details:

    1) once author submits paper to the site he canNOT withdraw it - submission is once and forever,
    otherwise commenters will be unhappy - if they spent efforts to write comments - but then someday all their efforts become deleted.

    2) If paper is not submitted to the site by the author - than there should be very clear warning for commenters -
    you are writing comments on your own - they might be deleted from the site if author withdraws paper
    (may be before deleting their should archiving feature - and sending such archives to the commenters e-mails, in order their work will not be lost completely - if they want they can put it in their blogs).
    • CommentRowNumber21.
    • CommentAuthorzskoda
    • CommentTimeFeb 18th 2012
    • (edited Feb 18th 2012)

    I do not understand this withdrawing business. If I discuss some publicly available material, why would the author of that material be able to withdraw my discussion ? It is not about somebody’s personal life, it is about mathematics. If there is a warning that my comment may be erased for such a reason, without my own withdrawal, I would consider the site unstable and not contribute. The papers are anyway not supposed to LIE on the site, but one includes just a link to the arXiv. Scott says it is a matter of basic respect to withdraw. But it includes the disrespect toward the commenter. Erasing is I think worse offense to a contribution than criticism.

    It is a different thing if an author announces and “pays” commenting by some sort of “refereeing credit”. Unless that is the case, why would an author open the discussion, or having anything to do with it, except being able to comment himself, and argumentatively defend his points of view to unlimited extent.

    On the other hand, if it is not an arXiv-available paper, but a paper specially author-sent to the site for commenting, than probably a different logic applies, I do not know in that case.

    • CommentRowNumber22.
    • CommentAuthorMarc Harper
    • CommentTimeFeb 18th 2012
    Regarding discussion takedowns: The ArXiv does not allow one to withdraw papers -- why would someone be allowed to remove a discussion that just refers to a paper? I am assuming there would be a reporting feature for derogatory/throwaway comments (that anyone could report, not just the author). Take a site like reddit (which is not the model, but is a 'discuss anything site') -- any article from the New York Times or any other publication can be discussed there and there is no takedown mechanism. From what I have seen, people are very happy to get attention for their works.

    In the case of a takedown request for a multiple-author paper, what if one or more authors does not want the discussion removed? What would the policy be in that case -- take down if any of the authors want it taken down? Once you publish something on a site like the ArXiv you cannot, practically or in reasonable expectation, to exert absolute control over the use, discussion, and dissemination of your work. I have no intention of crippling the features of the SPNet just because of a few "strange" users (as Henry says).

    @Henry We were not planning to have automatic pulling from the ArXiv, but once someone links to a specific paper and wants to start a discussion, we'd pull the necessary details from the ArXiv. I also agree that a lot of papers would probably not have much or any discussion. This is another reason not to automatically populate such a site from new or existing ArXiv submissions.

    We could email the authors of a submitted paper automatically once a discussion began. The intent is not to have yet another thing for an author to manage, rather to also be a valuable resource for the author. We hope to eventually incorporate a peer review system (separate from discussions) so that authors will have additional incentive to interact with the site.

    @Scott -- We are thinking of matching usernames to ArXiv usernames if at all possible. There are, however, many authors that do not use the ArXiv, and if such a site becomes successful and wants to open up to areas other than math and physics, this could be an issue long term (especially if the SPNet hosted papers for fields outside the scope of the ArXiv or if users have submitted multiple papers at "ArXiv-like" sites). I would not want to limit the users of the SPNet to a subset of ArXiv users, however in the beginning would probably be the case.

    Also, I absolutely agree with the desire for such a site to be "professional" in look and feel, and that strong identity leads to that. I also see how some people would be afraid to "speak their minds" if their identity is attached, and while I feel this is a good thing, I also feel there is too much emphasis on 'never being wrong' in academic communities, so I understand this fear. Identity will reduce moderation demands (how much only time will tell).

    @Alexander " what if a user with a nick rather than their real name wants to upload a paper? "
    In this case I mean what if someone posts under a nick but then wants to start discussion on one of their own papers -- would they have to make another nick? I would think that the author of a paper would want to identified as such during a discussion, but as I said before I really do not want users to be able to split their activities into multiple accounts.

    "if one comments "this paper is junk" on the site - the author of paper
    will have definitely bad opinion about the site as a whole and it will stop him to contribute..."

    Comments like this would be suppressed by moderation and reporting features. Negative comments are fine in a discussion if they are civil and present evidence to back up their claims, but simply saying "This paper is junk" would not be allowed.

    @zskoda I agree with everything that you said.

    Thanks to everyone for the useful comments and discussion so far!
  6. To Marc
    The site is very very necessary for students but they do not have papers in atxiv.
    How to be if you allow only arxiv based guys leave comments and ask questions? ???
    • CommentRowNumber24.
    • CommentAuthorMarc Harper
    • CommentTimeFeb 18th 2012
    @Alexander Students can get ArXiv accounts already. Endorsements are just needed to upload papers, and students can get those as well.
  7. this comment is about websites for commenting scientific papers in general, rather than completely focussed on the present proposition. The following workflow sounds reasonnable to me:

    (1) you write a paper and make it available (on the arXiv, say),

    (2) you refer to it at some discussion website, in order to get some feedback, then revise accordingly (or withdraw, or rework out, etc.); people who where interested by the talks you gave about it can also comment here,

    (3) you submit it to a journal, old-fashion or new model, sending the arXiv reference as well as the commenting site discussion,

    (4) the referee can take advantage of the commenting page to ease its work, can possibly use it to interact directly with you. Ideally, the site should enable journal referees to have a
    temporary account, which they could use only to discuss the given paper, and which would
    identify them as referees without giving their names, so that the author understands what is at stake when answering the comments.

    Greg Kuperberg indeed pointed out to me that anonymous refereing has a cost, because the editor has to act as a third party. Such a system could be used to preserve the anonymity while freeing the editor from this burden. In this respect, it has the advantage to be immediately useful, even if only few editors are willing to use it. And since it is designed to save their time, it would harness their selfishness. Moreover, in most cases I think that neither the author nor the referee would disagree to use the site if the editor asked them to.

    Last, it would probably be good to have a possibility of discussing the paper privately on the site, so that the referee can ask questions and be answered without fearing to embarrass the author. At the end of the process, a report explaining why the referee thinks the paper is interesting, what it achieves, etc. could be published in the public part of the commenting section. If the answers are not satisfying and the paper is not accepted, the question is wether the referee should be able to write anonymously a negative report that would be publicly available. I think it sounds like a bad idea, but that's an option.
    • CommentRowNumber26.
    • CommentAuthorMarc Harper
    • CommentTimeFeb 18th 2012
    @Benoit Chris and I have been discussing a "validation" system for the SPNet which is intended to be a type of peer review where validators raise issues (like tickets in an open source project) to which the author and the other validators would respond. Presummably these system would be private to the validators and the author (at least during the validation process). I think that this serves most of the purposes that you describe if say traditional editors and reviewers are allowed to participate, though I think even they are unnecessary given a critical mass of users. It also separates recommendations (meaning "I think this paper is interesting") from validations ("I think this paper is technically/rigourously correct"). To each objection raised by a validator, the author could respond by explaining -- in an active forum with the other validators -- either that is not an issue or what steps the author has taken to resolve the issues.

    The workflow would be as follows. An author uploads a paper to say the ArXiv and submits it for comments/recommendations. Based on the topic area, keywords, the authors relationship to other users, etc. the paper is suggested to certain other users of the site. The paper perhaps receives some discussion and perhaps some recommendations and the author could decide to ask for validators. Then users can volunteer to become validators, ultimately deciding to validate the paper or not. Users would get "credit" on the site by having their validation activies reported (e.g. this user has validated X papers in the past six months) much like people put such activities on a CV. Given a critical threshold of affirmative validations, the author tags the updated version of the paper as "published".

    From here, a journal editor just has to go look for topic-area relevant papers that have been recommended and/or validated through the SPNet. This forces journals to compete for papers through a transparent and fair system that gives ALL the credit to the academics involved (authors, recommenders, and validators) rather than the current system that lacks transparancy and accountability and forces authors to compete for journals and uses academics for free labor.
    • CommentRowNumber27.
    • CommentAuthorEric
    • CommentTimeFeb 19th 2012

    Hi Andrew,

    That is awesome, but why behind a password? Linking to instiki is a nice touch.

    • CommentRowNumber28.
    • CommentAuthorJohn Baez
    • CommentTimeFeb 19th 2012
    • (edited Feb 19th 2012)

    Needless to say, I volunteer to serve as an alpha-tester, beta-tester and moderator on this site, if people want me to do those things.

    Henry wrote:

    As you mention, there are already any number of places one could discuss papers online, but people seem to have practically no desire to do so…

    That seems counter to my experience. Besides having written 300 issues of This Week’s Finds myself, most of which discuss 1-10 arXiv papers, I read a lot of discussion of arXiv papers on math and physics blogs and (these days) Google+. The main problem I encounter is that it’s hard to find these discussions because they’re spread out all over the landscape.

    Also, people in the quantum information community told me they used to wake up and eagerly use Dave Bacon’s software to see which new papers on quant-ph other people had read and liked.

    So, I think this site is quite likely to take off, starting with some communities and spreading to others.

    One reason this site should take off is that it will do most of what the arXiv does, but also more. If I understand it correctly, it’ll still let you read and search for papers as on the arXiv. It won’t let you submit papers - but for that, it’s easy to make a link to the arXiv submissions webpage. So, there shouldn’t be a big mental price in using this website if you already use the arXiv. The only big difference is that you can, if you want, comment on papers, recommend them, and read other people’s comments and recommendations.

    In short: the arXiv plus some extra functionality.

    • CommentRowNumber29.
    • CommentAuthorAlexander Chervov
    • CommentTimeFeb 19th 2012
    • (edited Feb 19th 2012)
    @John Baez
    "The main problem I encounter is that it's hard to find these discussions because they're spread out all over the landscape."
    I guess it was before arXiv introduced "trackbacks" or at least trackbacking becoming easy and popular.

    ---

    I am very positive about the idea. But at the moment I do NOT see "killing competitive advantages"
    with respect to trackback system of arXiv + bloging. I am wrong ?
    More details: 1) author submit paper to arXiv and 2) by himself create blog-page
    3) create trackback to this blog-page - so everybody who wants to discuss can discuss on this page.
    This is easy solution which we can start right now.

    What prevents everybody doing this, right yesterday ?

    Imho the problem is that community is not ready to take "commenting/discussing" seriously, but not the lack of technology.

    So the important problem which I see is creating community of active people who will start
    doing this.

    ---

    Henry Cohn wrote similar things (thread "Will YOU contribute comments to papers ?" post 12)
    "The critical question I see for paper commenting is this: how will it be better than writing a blog post about a paper and getting a trackback linking to it from the arXiv? Zoran has made the important point that blogs aren't a permanent part of the scholarly record, but I don't see that as a bad thing. If people are submitting "official" comments that will be permanently attached to my article and archived forever, I'll be much more concerned about moderation than if it is just some person with a blog."
    • CommentRowNumber30.
    • CommentAuthorAlexander Chervov
    • CommentTimeFeb 19th 2012
    • (edited Feb 19th 2012)
    @Marc Harper
    May I ask you clarify what will be the policy - what comments are acceptable or not ?

    My feeling that it is unsolvable problem - I think that to find the "level of politeness" which will
    satisfy significant part of academia people is practically impossible.
    So I am really afraid that if want to satisfy many people - we should only allow very boring comments. As for me that would be unsatisfactory.

    " Negative comments are fine in a discussion if they are civil and present evidence to back up their claims, but simply saying "This paper is junk" would not be allowed. "

    Sorry I did not mean exactly saying this "paper is junk" - this is of course should be forbidden and can be controlled by automatic software search for rude words.

    But my feeling is that average academia people, are much more sensitive than average people on this site. John Baez wrote the same (at thread "< Peer Review, Refereeing > = 0" post 4) "Many of here are somewhat more bold or adventurous or tough-skinned or foolhardy than the average mathematician, so we should always keep that in mind when imagining what systems others will like to use."

    Another examples:
    From Thread "Quick, easy and fun solutions" post 67,
    Colin Gopaul wrote "people have openly stated they avoid MO due to the "tone" of some posts"

    Another example, me and Zoran Skoda see the value of the site in a opposite ways,
    at "Will YOU contribute comments to papers ?" post 18.
    " "Closer the format and mission of a comment is to a mini-version of a paper, I will be more enthusiastic about it. Closer the format to the popularity contest or chat group, I will be less enthusiastic about it."
    If you write a comment which is "mini-version of paper" - I think it is of value to put it in arXiv as a separate paper ! "
    • CommentRowNumber31.
    • CommentAuthorzskoda
    • CommentTimeFeb 19th 2012
    • (edited Feb 19th 2012)

    Sasha, I wrote in another thread, and I will explain here with an example. One can not disrespect the standards of stratified community and write a short comment as a paper and mix it with papers. One needs a category in which his action is not confusing the community and attaching stigma. Typical example is when I read a proof and find it unsatisfactory. Maybe the author had no time or space to write the full proof. I spend two days to reconstruct the full proof and would like this to make into useful thing for my students, for future. If I write a paper on the improved proof, the author of the original paper will probably say: look a fool, he writes trivial details, clear to every expert. Well, I am not an expert, I usually read papers in a variety of areas and often lack some well known background, Does that mean that I should put myself on shame, or to keep quiet ? If there is a good category where I feel INVITED to write a comment and full proof, probably the format and the surrounding will make the author and the community feel different about “my” completed proof. The author will say, look my paper seem to be well received, people write about it in arxiv:comment section, they fill more details and advance my idea and are interested in every detail. I do not want to be a jerk who says that say Connes or Kontsevich wrote an incomplete proof but want to be a guy who appreciates and works toward making their great contributions more accessible and extended in the way everybody would like. Sending me to the obscure discussion groups or writing sites of my own *who is this guy to open his own web page to preach that those guys wrote an incomplete proof) or to write papers to look a ridiculous jerk is not helping community ever hear about my improvements. Nor will I feel motivated to really improve to the level of a minor, but constructive, comment.

  8. @Zoran, that is Okay, but still can you and me coexist on one site ? :) I mean if I write "non-serious" comments - you say you will leave site, but if community will welcome only "serious" (as for me = "boring") comments, I will be unhappy.
    Let me ask again:

    Who will decide what is acceptable comment or not ?

    And clearly seeing that probably impossible to give criteria of acceptable comments which will make everybody happy - how to solve it ?
    • CommentRowNumber33.
    • CommentAuthorzskoda
    • CommentTimeFeb 19th 2012
    • (edited Feb 19th 2012)

    Sasha, am I writing Chinese ? To iterate, I advocate community with many niches clearly stratified. I want to distinguish a chat comment written in 25 seconds from a comment mini-article redoing a proof which requires a couple of days of serious work and wants to be stable and citeable. They can not exist on the same footing though they can be linked to the same site. But their stability, format, expected content etc. should be organized differently. I also asked what is the difference from forum functionality that you advocate ? Also what do you mean coexist if the category for the things I suggested is not created ? If you just create interlinked blog with arXiv it will still be just a chat/blog. Chats already exist and I feel talking on a new chat is waste of time.And clearly seeing that probably impossible to give criteria of acceptable comments which will make everybody happy - how to solve it ?

    And clearly seeing that probably impossible to give criteria of acceptable comments which will make everybody happy - how to solve it ?

    I said it multiple times: as long as you have stratified formats and categories people are wise enough not to mistaken a research article from a comment mini-paper, comment from an elaborated response, response from a huh-chat replica etc. Like we do in math: we label things as theorems, lemmas, facts, proofs, remarks, historical notes and so on.

  9. @Zoran stratification - within one site ? if yes how will you do it ? Moderators should do it ?
    • CommentRowNumber35.
    • CommentAuthorzskoda
    • CommentTimeFeb 19th 2012
    • (edited Feb 19th 2012)

    34 Sasha, I have suggested many ways already. One was to create the categories and formats. Create category arxiv;comments with its own numbering and allow LaTeX-enabled wiki format and pdf as possibilities there and already we have something. It has nothing to do with moderators. I extensively talked above about SELF-preclassifying, about a possibility to moderate myself into elaborated system. Most people are smart enough to classify wisely and save the need of most of the moderation. One size fits all would not do it.

    We are NOW mainly arguing about design of future purposes and formats. It better be done now, in advance, and not at the repair stage (mechanics and moderators). I would like this aspect of the discussion to advance. I do not understand questions for impossibility of any distinctions, formats, categories etc. It is so obvious, there are hundreds of possibilities raised around (not only at Math2.0).

    • CommentRowNumber36.
    • CommentAuthorJohn Baez
    • CommentTimeFeb 19th 2012
    • (edited Feb 19th 2012)

    Kevin Walker wrote:

    I would be happy to help out. I might be able to help with coding, if that sort of help is desired.

    Thanks, Kevin! Yes, Marc and Chris are good at coding but they’ll be wanting help… I gave them your email address, and urged them to say hi. I know all 3 of you somewhat, Chris the most, and can vouch for all of you as being cool dudes - at least if you trust my judgement.

    • CommentRowNumber37.
    • CommentAuthorJohn Baez
    • CommentTimeFeb 19th 2012

    Alexander Cherov wrote:

    I am very positive about the idea. But at the moment I do see “killing competitive advantages” with respect to trackback system of arXiv + bloging. I am wrong ?

    Maybe you mean you do not see them?

    More details: 1) author submit paper to arXiv and 2) by himself create blog-page 3) create trackback to this blog-page - so everybody who wants to discuss can discuss on this page. This is easy solution which we can start right now.

    What prevents everybody doing this, right yesterday ?

    The arXiv does not accept trackbacks from arbitrary blogs. I don’t which blogs they accept trackbacks from. I know they refuse to accept trackbacks from Peter Woit’s blog, despite repeated requests on his part.

    I like the idea of trackbacks to the arXiv, but it’s not the same as being able to see, for a given paper, who has read that paper, who has recommended it, together with discussion of that paper - together with being able to see, for a given person on the system, what they have read, recommended, and/or commented on.

  10. @John Baez corrected, thank you.

    "but it's not the same as being able to see, for a given paper ... "
    I agree, however, it is not "killing application" - my concerns, not because I want to critisize, but just to analyze (may be for myself) that what will be done, will be really useful, otherwise I am afraid to spent time... I hope you can understand...
    • CommentRowNumber39.
    • CommentAuthorHenry Cohn
    • CommentTimeFeb 19th 2012
    > That seems counter to my experience. Besides having written 300 issues of This Week's Finds myself, most of which discuss 1-10 arXiv papers, I read a lot of discussion of arXiv papers on math and physics blogs and (these days) Google+. The main problem I encounter is that it's hard to find these discussions because they're spread out all over the landscape.

    I agree that there's discussion out there, but just for a small fraction of arXiv papers. (This Week's Finds is absolutely fantastic and inspirational, but it is not unfortunately representative of the rest of the internet. Incidentally, does it issue arXiv trackbacks? That would be really useful.)

    I very much like the idea of making it easy to set up, locate, and participate in mathematical discussions online. However, I don't think comments associated to particular papers is the right model. Most discussions are broader than that: they involve multiple papers, they go off on tangents, they focus on different issues than the main topics of the papers, etc. The fraction of discussions that really involve just comments on one particular paper is very small. I'm concerned that by providing a mechanism for focused comments but not for broader discussions, you're going to end up facing the issue of how to handle it when people try to shoehorn general discussions into this framework. I also moderation is going to be very contentious, but that's another story.
    • CommentRowNumber40.
    • CommentAuthorzskoda
    • CommentTimeFeb 19th 2012

    39: for broad discussions there are many math forums, mailing lists and MathOverflow already. Article specific discussions are rather not standardized yet and scattered around the internet, usually even the author of the paper does not know where they are and it is difficult to figure out at glance which papers got generally much attention, and to resolve the issues if any. Paper oriented comments at a centralized site related to arxIv is a new thing, it would cover a new niche rather than repeat existing.

    • CommentRowNumber41.
    • CommentAuthorHenry Cohn
    • CommentTimeFeb 19th 2012
    Broad discussions are also scattered around the internet and could really benefit from having a central location too. (For example, there is always tension about this on mathoverflow, about how far in that direction the site should allow itself to go. Right now, I don't think there is any forum or mailing list that has nearly the breadth or visibility of mathoverflow among professional mathematicians.) If it turns out that there really is substantial demand for focused public discussion of individual papers, then I'll be glad such a site was created. If it turns out that most of the discussion does not really fit that framework, then it may be counterproductive.
    • CommentRowNumber42.
    • CommentAuthorKevin Walker
    • CommentTimeFeb 19th 2012
    (I started this in response to #s 15 and 18 above, but didn't get around to finishing it until now.)

    I agree with much of what Scott (#18) and Henry (#15) say, but I disagree with putting any restrictions, either opt-in or opt-out, on which papers can be discussed. I would encourage (civil, polite, professional, gatekept) discussion of all papers, regardless of the author's wishes.

    For me, one of the main values of such a site would be when I'm carefully reading a (probably not very recent) paper that is a little bit outside my area of expertise. If I come across an argument that is not clear to me, it would be nice to know if it was unclear to other readers of the paper as well (perhaps because of a non-obvious typo, or perhaps simply because the writing is not very clear). It would also be nice to know of related results that appeared after the publication of the paper in question. I would not want to miss out on these benefits just because the author is cranky or shy. Where would we be if we respected someone like Grothendieck's wishes in this regard?

    I like Scott's idea of having strict gatekeeping for who is allowed to participate. Perhaps the gatekeeping could be relaxed later, once the tone of the site has been established.

    I think Henry is correct that a site like this will provide new opportunities for irritating people to say irritating things in public. But I think the benefits will outweigh these drawbacks. Wind back the clock to just before the introduction of the telephone or the internet. Someone might point out that these new inventions will provide myriad new opportunities for criminals and other malefactors, or that they will be huge time sinks. And they would be correct. But nevertheless on balance the benefits of the telephone and the internet outweigh the drawbacks.

    One model that has not been mentioned yet is Amazon product comments. Most products have no comments at all, and for the (more popular) products that do have comments, most of the comment are low-quality and not very useful. Despite all this, I find Amazon comments to be very useful. It's easy to ignore the dumb comments and zoom in on the helpful ones. With proper gatekeeping and moderation, I think that the arXiv comment site can have a much better ratio of good comments to lame ones.

    I agree with Henry that many potential online discussions are not about a single paper, but that's a reason to develop another site in parallel, not a reason to not develop this proposed site.
    • CommentRowNumber43.
    • CommentAuthorHenry Cohn
    • CommentTimeFeb 19th 2012
    Regarding another site in parallel, why not a combined site? Instead of having official comments on paper X, it could allow all sorts of discussions, with trackbacks to all the discussed papers. This seems strictly better: it encourages more discussion, it better accommodates the types of discussion people seem to want to have, it makes it easy to find all discussions involving a given paper, etc. I don't see any way in which forcing people to classify everything as a comment on one paper would be preferable. Am I missing something?
  11. @Henry Cohn, Sorry I cannot imagine interface you keep in mind ?
    If discussion is paper based - the idea is clear, decribed by Chris and Marc.
    • CommentRowNumber45.
    • CommentAuthorHenry Cohn
    • CommentTimeFeb 19th 2012
    I don't see the interface as a big obstacle here. It could be implicit (automatically noticing every arXiv mention and issuing a trackback) or explicit (choosing some topics), although I'd recommend implicit. The key thing is that it should be easy to find all discussions for a given paper, but that's not hard to achieve. Having a clear interface for paper-based discussion doesn't seem like an advantage, if many people are going to try to use it for discussions that don't really fit that framework.

    There's some real work involved in programming, but neither the architecture nor the interface seem like fundamental difficulties. I see moderation as the real issue: what should the policies actually be, and how should they be monitored and enforced?
  12. @Marc
    Henry Cohn wrote above
    " I see moderation as the real issue: what should the policies actually be, and how should they be monitored and enforced?"

    More or less the same question I asked above.
    Very waiting for yours answer.
    • CommentRowNumber47.
    • CommentAuthorKevin Walker
    • CommentTimeFeb 19th 2012
    @ #43: That seems like a fine idea to me.
    • CommentRowNumber48.
    • CommentAuthorChris Lee
    • CommentTimeFeb 20th 2012
    Hi, Chris Lee here... (sorry, when I registered for this forum I didn't realize it would only show the username and not my real name!)

    @Andrew: could you tell us what you dislike or want changed?

    @Scott et al.: I completely agree that users should be authenticated, i.e. a known arXiv user. I view real, authenticated identity as a bedrock foundation for participation in this collective enterprise, because it is the only basis for accountability. At a minimum, the system should know who is making each comment (even in cases where there is a good reason not to display that name on the public interface), so that it can track metric, relay feedback etc. A user should have a single, permanent identity. We should minimize the attack surface for creating notional logins, multiple logins, or abandoning an old identity and creating a new one.

    Regarding the challenges of keeping discussion substantive and constructive, this can only be addressed long-term by seeing what works, anchored by authenticated identity, i.e. the community can hold individuals accountable for following its rules. The following are merely suggestions:

    Just as journals push referees to take peer review seriously (e.g. by asking about potential conflicts of interest), an open evaluation site should articulate clear common-sense guidelines, e.g. comments should be relevant, specific substantive discussion of the claims and evidence, not flame wars.

    An identity will gain value over time, by building up a history of recommendations, subscribers, and relationships; what we're aiming for here is a digital embodiment of reputation. Users who are contributing sincerely to the goals of the enterprise should accrue these benefits automatically; this will happen by simply letting other users easily choose who they're going to pay attention to within their own research specialization(s). Those who contribute value (useful recommendations, validation and discussion comments) will receive value (subscribers, influence, reputation).

    I suspect that mechanisms like this will give a sort of 80-20 rule in which most of the possible moderation problems will be addressed by self-moderation (i.e. when a site has a clear culture, most people will adapt their behavior to that culture) or moderation by other users. For example, the site could encourage users to flag comments that violate the spirit of these guidelines, and the person who receives this feedback would be given a chance to either accept the criticism (and modify their comment to follow the guidelines), or to explain why they think they're within the guidelines. If the original flagger does not accept this argument, it goes to some kind of moderation mechanism (either by official moderators or by a small random sample of users). Only time can tell how much of this will actually be necessary.
    • CommentRowNumber49.
    • CommentAuthorMarc Harper
    • CommentTimeFeb 20th 2012
    @Alexander et al re: comment policy:

    There is not a point-by-point policy written out. Like Chris says, we'll have to see how it evolves, and there will be some "common-sense" guidelines that allow as much substantitive discussion as possible without permitting obviously inappropriate comments such as ad hominem attacks.
    • CommentRowNumber50.
    • CommentAuthorJohn Baez
    • CommentTimeFeb 20th 2012
    • (edited Feb 20th 2012)

    I agree that a good moderation policy and good mechanisms for enforcing it will be crucial. So here’s a question to the MathOverflow crowd: how does moderation work there?

    Is there a place I can read about this? Is there also unwritten wisdom shared and debated among a crew of moderators?

    Back when Urs Schreiber and I were among the moderators of sci.physics.research, we agonized endlessly over these issues. We eventually realized we’d been too soft. A harder line works better to strongly discourage people who enjoy testing the borderline of what they can get away with. Once a culture is established where jerks and crackpots are instantly banned, the amount of work needed to keep people nice goes down. However, the price for productive conversations is eternal vigilance. There’s always some new trick someone will try.

    We took advantage of this hard-won wisdom when we started the n-Category Café. On Azimuth, I’m even able to have people discuss climate change without acting like jerks.