I’m writing this mostly for students and postdocs contemplating an academic career, just so they know what they’re getting into. It’s not all about working quietly in your office, sheltered from the business realities of “real life”. If that’s what you expect, be warned that you might not get it. If on the other hand you enjoy working with people and find the gloomy ivory tower cliches unappealing, I can reassure you that they bear no resemblance to my own work.

Before someone gets on the “LOL professors complain” soapbox: I do a lot of this work on a voluntary basis. I enjoy much of it. That doesn’t mean that we can’t talk about how easily “research-related activities” can drive out actual research if we let them, or what we do to maintain a balance. Nor does it have to stop us from speculating on how the system might evolve and which way we would like to push it.

Service to the department and the university. This includes mandatory service on departmental and university committees, attending faculty meetings, undergraduate advising, interviewing job candidates, reading and evaluating graduate applications, various kinds of reporting, and more. Some committee assignments are light; others carry a heavy workload, for example the hiring committee reads and evaluates hundreds of job applications. One departmental calculation estimates this part of a full professor’s workload as roughly equivalent to teaching one course.

Applying for research grants. I’m going to restrict this one to Canadian mathematics; other countries and disciplines of science will have other funding systems. Our main source of funding is NSERC Discovery Grants, for which we apply every 5 years. This is the money that pays the research stipends for our graduate students and postdocs, our travel expenses and other similar items. (Unlike in the U.S., we cannot use our grants to pay ourselves a summer salary.) Once in a while, we apply for funding (from NSERC or other agencies) for a specific purpose, e.g. equipment grants or support for conferences. We sponsor our students and postdocs for various awards. We may also be asked to support grant applications on behalf of institutions (in 2006/07, I was a co-applicant on the Fields Institute application for NSERC funding).

If you are interested to find out more about the NSERC application process, here is the information page for applicants for NSERC Discovery Grants. NSERC, unlike NSF, does not estimate the time required to complete the application, from collection of information (e.g. looking up the current positions of our past students and postdocs) to actually thinking about our research plans, but the corresponding number at NSF is 120 hours. The NSERC average is probably similar or higher.

For an especially work-intensive example, here is the guide to applying for a Banting postdoctoral fellowship. (I have not sponsored a Banting application, but I have served on a committee to screen them.) A Banting application is a joint project between the applicant, the proposed supervisor and the host institution, with the supervisor preparing some of the documents and presumably helping out with the rest. The institutional part alone (in practice, written by the supervisor and then edited in consultation with the research office) includes “Institutional Synergy and Letter of Endorsement”, “Supervisor’s Statement”, “Research Environment”, and “Professional Development”, addressing criteria such as “Describe the institution’s documented strategic directions and illustrate the synergy between that strategy and the applicant’s proposed research program”. (Remember that we’re talking about postdocs, not distinguished research chairs.) And that’s for an award where the chances of success are infinitesimal: there are only 70 Banting fellowships awarded each year, for all of Canada, in all fields of science.

Having postdocs and graduate students. I love working with postdocs and students, but it does add to the administrative load. This includes making postdoc offers, balancing the budget and making sure that everyone gets paid, writing letters of recommendation, helping out with proposals and applications, candidacy exams, final doctoral exams (including logistics and paperwork), advising students on administrative matters, and so on.

Not all of it is straightforward. My current postdoc is supported by a combination of an NSF postdoctoral fellowship, NSERC grant money, and teaching. It took about 3 workdays just to work out the logistics of the appointment so that it would be acceptable to everyone and compatible with each agency’s requirements. (That would be the actual time spent on it. The whole process was spread out over 2 weeks or so, including an email exchange with NSERC.) There is a staff member who helps with this, but we still end up having to know arcane rules such as where the funds for a postdoc’s benefits (extended health etc.) can come from depending on how much of their salary is processed through UBC payroll. Didn’t think that this was part of a research job? Sorry.

Then there are times when things aren’t moving as smoothly as they should. Students and postdocs have families, health problems, personal circumstances. Some don’t do well in the program, for various reasons. Some want to reconcile graduate school with exploring other options. We can’t sort out their personal issues, but we do have to look up policies on leaves, deferments and conditional admissions.

Service to the community: organizing conferences and institute programs (including applying for funding), editorial appointments, refereeing papers, service on grant selection panels and award committees, assessing promotion and tenure cases at other institutions, etc.

This is generally done on a voluntary basis and we can decline any particular request if we don’t have the time. Nonetheless, we are expected to pull our weight overall, and there are consequences if we keep turning down everything. Promotion and tenure committees expect to see some level of activity on the national and international scene. If you expect your own papers to get refereed and published, or your grant proposals to get reviewed, you owe it to the community to do a proportional share of refereeing. Some assignments (editorial appointments, membership on award selection committees or institute boards, etc.) are also considered to be marks of professional recognition, so that your willingness to do it (or not) reflects on other aspects of your career, from merit salary increases to the quality of the graduate and postdoctoral candidates you can attract.

There are many reasons why you may want to support and promote your area of research, for example by organizing conferences or summer schools. In the long run, it contributes to having a healthy and supportive community, encourages young researchers to enter the area, and makes them feel welcome once they’re there.

Also, this is often the most interesting and educational part of admin work. A standard invitation to serve on a grant selection panel usually says something along these lines (actual quote):

By participating in this panel you will have the opportunity to influence and shape current research in these areas. Many panelists have also found that their panel experience has provided them with new perspectives and exposure to recent developments on a variety of topics close to their own research interests.

… and this is actually true. You can learn a lot from serving on panels, award selection committees, meeting program committees, and other similar bodies. Organizing a successful conference or program can inspire a sense of accomplishment, of having done something for a group of people that you care about.

It does take time, though. Refereeing can be quite time-consuming, given that some papers can take weeks to read and understand. (On the other hand, I would not referee one of those unless I had research-related reasons to read them anyway.) Organizing a CMS or AMS meeting session is relatively simple: you are only responsible for choosing the speakers, inviting them, forwarding meeting information to them, and coordinating the schedule. Other conferences can get more complicated, especially when it involves funding applications.

Organizing institute programs should really be in its own special category. I organized the 6-month Fields Institute thematic program on harmonic analysis in Winter 2008. This was a long process that started back in January 2005 (no, not a typo) and involved multiple proposals and funding applications, postdoctoral offers coordinated with several universities, course buyouts, long-term and short-term visitors, graduate courses, distinguished lectures, two workshops and a major conference. I had a 1-course buyout from Fields to attend the program, but the total amount of work involved was far greater than that (probably the equivalent of 2-3 courses spread over the 3.5-year period), not that I regret any of it. I’m not going to try to actually calculate it (too much work), but in case you’re interested, I still have several mail folders with a total of 894 1078 program-related emails. (I missed a couple of folders the first time I counted.) I have no idea how many I’ve deleted.

My own service workload. For those curious how much this adds up to and how it breaks down, I’ve tried to estimate how much time I have spent on administrative activities in 2011, and this is what I came up with.

• NSERC Discovery Grant proposal: 120 hours. (The official NSF estimate; looks about right to me.)
• Chairing the Putnam committee: 120 hours. (Preparing my proposed problems and solutions, solving and ranking problems proposed by other committee members, compiling the rankings, organizing and attending the committee meeting, proofreading problems and solutions, various correspondence.)
• Departmental graduate affairs committee: 70 hours. (More than half of that time was spent on graduate admissions and ranking of scholarship applications.)
• Administrative work on behalf of students and postdocs: 70 hours. (Arranging a postdoc offer, one student’s graduation, financial support issues, providing advice on administrative matters, and more. Letters of recommendation are included in a separate category.)
• Editorial work and refereeing: 60 hours.
• Letters of recommendation: 45 hours. (Includes evaluations for promotion, tenure and senior appointments.)
• Promotion and tenure cases in the department: 20 hours. (Attendance at faculty meetings, service on promotion and reappointment committees.)
• Undergraduate advising: 20 hours. (Answering email questions, some meetings in person.)
• Banting screening committee: 15 hours.
• Miscellaneous: 50 hours. (Faculty meetings not included above, organizing the harmonic analysis seminar, making arrangements for research visitors, updating individual and group webpages, faculty mentoring, AMS Current Events Bulletin committee, miscellaneous email and correspondence, probably other things I’m forgetting about.)

This adds up to 590 hours, or 14.75 40-hour weeks, or about 3.5 months of working a full-time job.

I came up with these numbers by trying to remember at least the main items in each category, recalling how much time they took, and estimating how long it would take me to get through all of them in one stretch. In reality, some items were in fact done in long blocks of time (NSERC, solving and ranking Putnam problems, ranking graduate scholarship applications), but most of it was done in small increments throughout the year. I wanted to have an actual number assigned to each category, rather than a range. I also (obviously) rounded up the numbers, some upward and some down. Overall, I have tried to stay on the conservative side.

This is an estimate of how long it took me to do this work. It’s quite possible, for example, that the Putnam assignment would have been less time-consuming for someone who coaches a Putnam team and is better versed in the contest lore. On the other hand, I try not to cut corners, especially with things like grant proposals, recommendation letters, or opinions on tenure cases. (Technically, I suppose it would be possible to “complete” the NSERC forms in 2 days or so, not that it would get me any grant money.)

The estimates also include the “learning time” – whenever we do something we haven’t done before, which is often, we have to spend some time learning the procedures, policies and regulations, and finding out how it actually works in practice. It helps to have a colleague or staff member who has done the same thing before, but we can’t generally count on that. In my case, I was the only active harmonic analyst in the department for 6 years, and I’m still the only full professor in the group, so I’ve had to figure out a lot of it on my own (and still do). It does get easier when you do the same thing the second time around.

This is fairly representative of my administrative workload in recent years. The two main items, NSERC and Putnam, are not annual occurrences. The Putnam committee was a 3-year assignment (2 years as member, 1 year as chair) that has just ended, and NSERC applications are made every 5 years. On the other hand, last year I did not organize any conferences or institute programs, or serve on funding panels or award selection committees, all of which I have done in the past. I also had fewer letters of recommendation to write than in previous years, and I had to decline most refereeing requests due to the Putnam and NSERC workload.

We’ll be talking more about it later: how it interacts with our research and teaching, how much is enough, what could be cut (realistically or not). For now, I’ll just leave it there for information purposes.

Meanwhile, if you need a break (and who doesn’t?), here are a couple of photos from last weekend. (Click on the picture for the full-size image.)

I would have a different story to tell, and then a mental exercise to suggest.

There are indeed many of us who have made research our lives’ work. We’re in it for the challenge and the pleasure of discovery, for the outlet that it provides to our creativity. That’s why we ended up in academia in the first place. We’re good at research and we’ve demonstrated this to everyone’s satisfaction.

And yet, a full professor at a large research university in the U.S. or Canada often has to work a full-time job, 40 hours a week or close to it, before any research gets done. We spend that time on teaching, writing grant proposals, supervising graduate students and postdocs, serving on committees, attending faculty meetings (here, attendance at promotion and tenure meetings is mandatory), editorial work, refereeing, writing evaluations and reference letters, and more. Much of it (calculus teaching, many committees) is completely unrelated to research. Some tasks (refereeing, various evaluation exercises) call for the scholarly expertise we’ve gained in the course of our work as researchers, and some (grant proposals) are quite directly related to that work, but writing about research and actually doing it are two different things.

The time we have left for research while classes are in session amounts, essentially, to evenings and weekends. (The actual schedule may vary: we may carve out a weekday afternoon or two for research meetings, then prepare for classes after hours instead.) Maintaining a quality research program requires a good deal of time, preferably in long uninterrupted blocks. From the time classes start to the end of the final exams, we’re, for the most part, dead to the rest of the world.

It would be unrealistic to expect that more than a handful of us could make a living purely on research. Traditionally, academic jobs have always combined research with teaching. But the same tradition also defined university teaching in terms of making advanced knowledge available to students, with emphasis on the students’ own agency in partaking of that knowledge. That’s what I remember from my own undergraduate days. The university professors of old were not evaluated on their ability to inspire interest in otherwise indifferent students, nor did they have to teach the addition and multiplication of fractions. They were not under constant pressure, either, to use clickers, classroom technology, or innovative teaching techniques.

This is not to defend bad teaching. Boring and pointless lectures are boring and pointless. Better teaching requires more time and effort, though, and so does maintaining the current quality of teaching when classes get larger and students come in less and less prepared. This rewrites our job description, increasing our teaching workload in terms of hours spent on it even as the number of courses we teach remains unchanged. Such work can be worthy, challenging and satisfying; but, in the present quantity and combined with everything else we do, it is barely compatible with research, the activity that (as I said already) attracted many of us to academic jobs in the first place.

It’s not just teaching, either. The demands placed on us get ratcheted up across the board. NSERC has effectively made “HQP training” mandatory for all scientists who seek Discovery Grant funding. Projecting from current data, it may well become next to impossible to get a research grant 10-15 years from now without some form of outreach activity and/or engagement with the industry. New administrative tasks get added on a regular basis. Promotion and tenure committees have been known to raise questions if there is some specific type of activity that we did not engage in (why did this person not teach large classes?).

Research still gets done. We still have the summer (the “4-month vacation”, in the popular opinion), we learn to guard our time, and some semesters are lighter than others in terms of teaching and administrative loads. And we still have sabbaticals, although with the multiple postdoc jobs at the beginning of the career it can take a very long time before we’re eligible for one. (12 years, in my case. But at least I didn’t have to teach throughout graduate school, as many students do nowadays.)

But the cracks are beginning to show. I’ve known talented and accomplished young researchers who quit academia so that they could spend evenings and weekends with their families. There’s often a slump in a researcher’s productivity around the time the administrative and HQP duties really kick in. People burn out and give up. The plague of “least publishable” incremental papers is due not just to the constant pressure to publish, but also to the lack of time to work out something more substantial.

More importantly perhaps, the mindset required by the ever expanding non-research part of our job – entrepreneurial, businesslike, competitive, often adversarial, well suited to running multiple tasks in parallel on tight schedules – is not at all what it takes to do mathematical research, a quiet, focused, intense pursuit, sometimes playful, sometimes contemplative, almost always unstructured. Some of us combine the two well enough, be it through natural predisposition or learned skill. Others don’t.

True creative talent in mathematics is rare. It should be nourished and encouraged wherever it can be found. And yet we’re apparently willing to discard it unless it’s coupled with a passion and gift for teaching, a comfort with the pressure cooker of paperwork and deadlines, and any number of other skills completely unrelated to research per se. We saddle top researchers with more and more duties that just barely (if at all) touch on their particular exceptional talent.

Now, here’s the mental exercise. With all the above in mind, what other hypothetical models of an academic career might be feasible that would give us more research time and attract those potential researchers who may be discouraged by what they see now?

I, for one, would be very interested to see a world where teaching is highly valued and well rewarded, where being asked to teach three courses instead of two is a mark of prestige, where financial compensation for teaching is higher than research pay. That would make teaching more desirable for those who wish to go that way, but by the same token, it could provide opportunities for others to scale back or opt out. We’d earn less money and miss out on teaching prizes, but we’d have a more manageable workload and more time available for research. Some of us would find it to be a reasonable trade. (I also wonder how gender dynamics might work in that world. It could well turn out that women are only good at research, and only interested in research careers, but never as good as men when it comes to teaching. But I digress.)

(And yes, this would cost money. Teaching, at every level from kindergarten to college, is demanding and difficult work that has long been underpaid, especially in the U.S. Want better teachers? Pay them better.)

Universities may have to become more flexible in offering workload-reducing arrangements. This is already being done on a small scale, from personal and family leaves to sabbatical leaves to teaching buyouts. Part-time appointments might become more common, not just for health or family reasons, but also for those of us who are willing to accept a lower salary in return for a lighter schedule with more research time.

Some European countries (e.g. France) have government-sponsored research institutes that offer research appointments with no teaching. The system varies from one country to another, but there’s usually some trade-off involved, for instance lower pay or lack of job security. Perhaps someone better informed could fill us in on that?

And why not combine a part-time research appointment with a non-academic job? We could make our living working elsewhere, say at a financial institution, in the tech industry, or at a start-up, and combine that with a university affiliation through which we could supervise graduate students or apply for grants. No, really. Think about it. Many employers are already used to part-time work arrangements, and a grant system similar to the NSF summer salaries could supplement our income. Most of the possible objections from our individual point of view – the other job has little to do with our research specialization, it would take time and effort to learn it, and so on – apply equally well to teaching. Much of the administrative overhead might just disappear. NSERC is encouraging us to “engage” and “interact” with industry – what better way to do so than actually working there? There must be institutions or start-ups out there that would be happy to have some of our time.

Well, that’s enough speculation for now. But please feel free to add more in comments.

Who needs Christmas lights when there’s sun and raindrops.

In the wake of the tragically premature demise of Steve Jobs, it seems appropriate to ask: What can Canada learn about innovation from the career of Steve Jobs? I think there are two important lessons that we could take away.

The first lesson is that commercial success and impact is more about innovation than about invention. Invention is the creation of some new-to-the-world technology, molecule, material, or formula. It is typically the product of the curiosity of a scientist. It can be pretty earth-shattering when it is electricity or insulin. But it can be pretty irrelevant when it is a technology in search of a user.

Rotman’s recommendation, then, is to overhaul the traditional K-12 curriculum and “become the first nation on the planet to have universal education in innovation by explicitly and clearly teaching innovation in the primary and secondary school system.”

Well… let’s hear from Mr. Jobs himself, shall we? From the Stanford commencement address (emphasis mine):

And 17 years later I did go to college. [...] And much of what I stumbled into by following my curiosity and intuition turned out to be priceless later on.

Let me give you one example. Reed College at that time offered perhaps the best calligraphy instruction in the country. Throughout the campus every poster, every label on every drawer, was beautifully hand calligraphed. Because I had dropped out and didn’t have to take the normal classes, I decided to take a calligraphy class to learn how to do this. I learned about serif and san serif typefaces, about varying the amount of space between different letter combinations, about what makes great typography great. It was beautiful, historical, artistically subtle in a way that science can’t capture, and I found it fascinating.

None of this had even a hope of any practical application in my life. But ten years later, when we were designing the first Macintosh computer, it all came back to me. And we designed it all into the Mac. It was the first computer with beautiful typography. If I had never dropped in on that single course in college, the Mac would have never had multiple typefaces or proportionally spaced fonts. And since Windows just copied the Mac, it’s likely that no personal computer would have them. If I had never dropped out, I would have never dropped in on this calligraphy class, and personal computers might not have the wonderful typography that they do. Of course it was impossible to connect the dots looking forward when I was in college. But it was very, very clear looking backwards ten years later.

Again, you can’t connect the dots looking forward; you can only connect them looking backwards. So you have to trust that the dots will somehow connect in your future. You have to trust in something — your gut, destiny, life, karma, whatever. This approach has never let me down, and it has made all the difference in my life.

(Full transcript here in comments.)

I guess there are all kinds of important lessons to be learned from this.

I’m old enough to remember the days when most math departments had pre-printed postcards with requests for journal offprints. (“Dear Professor [fill in the blank], I would be most grateful if you could send me an offprint of your article [fill in the blank] that has appeared in [fill in the blank]“. That’s what the offprints were for, mostly. They also looked much better than a manuscript typed on a mechanical typewriter, with handwritten math symbols.) Scientific journals actually served to disseminate information back then – checking new issues in the reading room was an important part of keeping up with recent developments. Ah, the good old times.

Dissemination is in our own hands now. I usually check the arXiv every day, but it’s been years since I last bothered with the current journals in the library, other than to look up published versions of papers that I’d already seen as preprints. Of course we will want to take ownership of the rest of the publishing process: the record-keeping, the peer review with its twin goals of debugging papers and evaluating their merit. These are functions that are worth keeping. I do use the library on a regular basis for older articles; I’d rather cite a stable, debugged journal article (where possible) than a preprint that could get replaced or pulled down tomorrow; and, as inaccurate as it can be to judge papers by the journals they appear in, I’d rather have such (approximate) marks of the quality of my work in place than leave it to each year’s departmental committee on merit pay increases to try to figure out all over again what I’m doing and why it’s supposed to be important.

It’s clear enough that any alternative publishing model will likely be internet-based, with interactive components possibly similar to Math Overflow or blog comment sections. It has also been noted that women have significantly less visible presence on MO than they do in research mathematics overall. One might ask, therefore, whether switching to an internet forum-based model of publishing could have the side effect of alienating women mathematicians and driving them out of the field.

In this regard, most discussions of why women don’t participate in something or other tend to follow the same three-step outline:

1) Men come up with a list of reasons why they think women don’t participate.

2) Men decide that these are not valid reasons.

3) Therefore, women should just get over it and participate! Problem solved.

But when I started thinking of what I’d want from a future publishing and peer review model in general, and what would make such a model women-friendly, I came up with pretty much the same principles from both angles: fairness, professionalism, adherence to merit-based procedures, having error-correcting mechanisms built into the system. These are all good for women, but they’re also good for mathematics. The list below isn’t meant to be complete; feel free to add more in comments.

Decentralization. I would not be in favour of one huge, centrally designed website for all our math publishing and peer review needs, possibly with one numerical scale to rank us all. To be clear: I think that arXiv is great. I don’t hang out much on Math Overflow, but I’m glad that it’s there and I would have likely become a heavy user if it had been introduced a few years earlier. What I would not welcome is the mixing of different functionalities and rankings, the assigning of numerical scores to apples and oranges alike. If you don’t see why this could be a problem, I might suggest reading this first:

Amos and I once rigged a wheel of fortune. It was marked from 0 to 100, but we had it built so that it would stop only at 10 or 65 One of us would stand in front of a small group, spin the wheel, and ask them to write down the number on which the wheel stopped, which of course was either 10 or 65. We then asked them two questions:

Is the percentage of African nations among UN members larger or smaller than the number you just wrote?

What is your best guess of the percentage of African nations in the UN?

The spin of a wheel of fortune had nothing to do with the question and should have had no influence over the answer, but it did. “The average estimate of those who saw 10 and 65 were 25% and 45 respectively.”

So, if you’re ranking on a scale from 1 to 100 a paper written by someone ranked in the 98th percentile of frequent commenters on the site, in an area of mathematics whose “hotness” is 90%… OK, I’ll leave it there.

Mathematics is an ecology that would not benefit from having everyone compete directly against everyone else all the time. Emerging research areas often grow in small niches first before hitting it big time. A less fashionable area might carve out a bit of room for itself – and thus live to see the day when it becomes fashionable again. (Additive number theory comes to mind.) A centrally planned megasystem might recreate the complex environment that we have now, or it might not. And if mathematicians in some area felt that they were getting the short end of the stick… why, they might even come up with the idea of having a specialized journal for themselves! That’s why I wouldn’t really expect such a centralized system to last very long, even if people could be convinced to buy into it in the first place, which is not likely.

Second opinions. This is related to what I’ve said already about decentralization, but important enough to be mentioned separately. Right now, if my paper is rejected by A Prestigious Journal (good but not good enough, etc.), I can send it right off to Another Prestigious Journal. Chances are that the second journal will use a different referee and their opinion will be independent of the first one – and different from it. (And my paper will find a happy home.) This doesn’t mean that either journal acted in bad faith. There can be genuine differences of opinion as to how interesting or significant a given result might be.

But let’s say that my paper gets posted instead to a peer review website and the first review is negative. Every follow-up reviewer now has to address that. It’s tempting to think that we’ll just all be fair to each other, that any misguided or unfair opinions will be banished through the magic of downvoting. I wouldn’t count on it.

First, if a strongly dissenting opinion gets posted subsequently, there’s a good chance that the discussion will devolve into a shouting match, possibly on an unrelated subject. The comment page for my paper will become an unpleasant place to visit and most colleagues will just stay away, regardless of the actual merit of my work.

More likely to happen, though, is what happens in real life. We’re all human. I’ve seen sloppy and biased professional assessments, and I’ve seen some that can only be explained by unbridled partisanship if not downright malice, and I’m sorry to report that there isn’t always a long line of colleagues waiting for their chance to correct the record. People are busy enough already. They’re not always willing to risk their relationship with someone they know by speaking out on a subject where they don’t see themselves as unquestioned experts. And, well, they don’t want to be the impolite person who derails the discussion and makes the whole place unpleasant, so they maintain the decorum and tone down their disagreement.

This aspect of decentralization is especially important for women. Social change happens in small increments, one department at a time, one hire at a time. We’ve all heard the story of how David Hilbert convinced the Gottingen mathematics department to hire Emmy Noether. Well, if he’d had to run a Europe-wide referendum on that, chances are that Noether would have never had a paid university position in Germany.

Controlled anonymity. Much has been made of the anonymity of the journal referees. The editors, however, are not anonymous. They select the referees, make judgement calls, sign their names on acceptance and rejection letters. It’s their responsibility to try to send my paper to someone conscientious and knowledgeable enough in my area of research. It’s also their responsibility to arbitrate fairly any disputes between authors and referees and to step in when either side takes too many liberties.

The anonymity of the referees must be preserved. There’s no other way that people will submit frank and candid feedback. But anonymity can also lead to abuse and flame wars, usually “won” by those with the most time on their hands, as happens again and again in online comment sections. There must be an editor or moderator in between who chooses or at least approves the referees, knows their identities and does not allow Gawker-style comments from the peanut gallery.

Which brings us to the last point for now.

Moderation and civility. Mathematics blogs and websites, including Math Overflow, tend to be quite light on moderation. This works well enough when the discussion is about the finer points of the theory of C^* algebras. When the topics get more contentious, though, the thread often gets hijacked by a few commenters with strong opinions; even if everyone else disagrees with them and says so, that still does not undo the damage, which is that it becomes impossible to have an interesting, informative or enjoyable discussion in the same space. (I could link to any number of threads on women in math here.)

All political or general interest bloggers with good comment sections know this and moderate their comments heavily and aggressively. This doesn’t just mean deleting the most blatantly offensive and threatening comments and allowing everything else. Most major newspapers have policies like that, and their comment sections are awfully depressing to read. Compare this to Juan Cole’s comment rules, for example. Guess what? He actually gets good comments. On Middle Eastern politics, not exactly an uncontroversial subject. Many other blogs have similar policies, although they don’t always state them as explicitly.

Women get targeted particularly often for trolling, harassment and violent threats, including on academic blogs. Any math publishing site with a comment section will have to deal with this. It will also have to deal with lesser offences such as mansplaining. This isn’t usually considered serious enough to warrant banning a commenter, but nonetheless women will avoid (if they can) any websites where such behaviour is common.

I moderate comments on this blog and I’m not going to get into topical discussions on any sites where comments are not moderated. Of course, when I say that X is mansplaining, X might think that he’s just stating obvious truths that I’ve failed to understand. That is why I don’t see how websites with comment sections might breach contentious issues and discuss them productively without an aggressive moderator who, moreover, has to take sides from time to time. And that is why I don’t think that we can have genuinely contentious but still productive exchanges on general purpose websites where no one in particular takes responsibility for the content. A good discussion section on a blog is probably the best we can hope for.