## December 20, 2006

### Research Blogging

#### Posted by Urs Schreiber

Clifford Johnson ponders the advantages of research blogging in The Blog as a Sharp Tool for Research and now again in Research Blogging.

Related considerations were voiced by Craig Laughton early this year: Exploring the Blogosphere.

I don’t have much to add to that, except for noting that I used blogging in this sense, and almost exclusively in this sense, from the very beginning.

And I’d guess that, long long before my feeble writings, John’s TWFs served a similar purpose.

One big difference is that Clifford Johnson has a non-public blog for his research, which, as he writes #

[…] is also the place where everyone (including me) can say silly things and ask silly questions if we want to, without the whole world watching. That latter is a very important feature, in fact.

For some reason I have always felt like moving private discussions on technical issues out in the open. For me that’s a matter of increasing the reaction rate of research by increasing the reaction surface. And, looking back, it did work for me #.

With the esoteric stuff we are talking about, this seems more important to me than shielding away insights and hiding mistakes. The game here is not Bingo #.

But I am aware that most people feel quite different about this – and quite possibly for good reasons. But I cannot help it. On the other hand, I am fond of having found philosophical support # # from David Corfield.

While I cannot prove it, I think everybody would benefit from seeing more research-related communication done out in the open. The most valuable aspect of many conferences is the conversation one has in between the talks. And this kind of conversation is what I am after.

Posted at December 20, 2006 4:50 PM UTC

TrackBack URL for this Entry:   http://golem.ph.utexas.edu/cgi-bin/MT-3.0/dxy-tb.fcgi/1080

### Re: Research Blogging

Hi Urs,

Your comments about privacy vs non-privacy are ironic, given that in the original idea I tried to get going (linked in my recent post) I was advocating a public system with research conversations coordinated between all participating research groups, with different groups leading the conversation at different times, etc, etc - and this public aspect was the one that people most objected to. In the end, it seemed that nobody really wanted to play, so I implemented the smaller-scale private system just for my own lab. Please read the other post for more on what I was suggesting. Sounds like it would fit your wishes rather well.

Cheers,

-cvj

Posted by: cvj on December 20, 2006 6:33 PM | Permalink | Reply to this

### Re: Research Blogging

Hi Clifford!

and this public aspect was the one that people most objected to

I would think that the way to go is not to ask everybody beforehand and then set up a big system that tries to suit everybody’s needs.

Few people will have any good idea what that mysterious public-research-like-discussion-system would actually be like. I wouldn’t. It has no predecessors.

On the web, the interesting things tend to develop dynamically. Interconnection is more or less automatic (direct hyperlinks, automatic trackbacks, comments, etc.)

We are having an inter-blog discussion here right now. And I imagine if you posted more about your current research to your public blog, or at least about related things, I for one would be highly interested in that and follow it closely. Reading your entry might inspire me to post a related entry here, sending a trackback to yours. And so on.

So, I don’t think we are lacking the technology for a web-wide interactive research-related discussion forum. Currently we are lacking people who would consider contributing.

But more might join in when they see more big shots like you talking about serious research-related stuff (preferably clearly seperated from all those other topics, like politics etc.) on an existing weblog.

Posted by: urs on December 20, 2006 6:57 PM | Permalink | Reply to this

### Re: Research Blogging

One factor concerning privacy has to be the rank of the participants. A bunch of grad students may be much happier talking if only their supervisor gets to hear. And there might be genuine fears that ideas could be appropriated at a critical time as first papers are being written.

I don’t see that these considerations should stop sufficiently established creative researchers. John wrote:

I want to switch over to even more of a gift economy approach to research - just giving away lots of ideas, instead of officially publishing them. I already give away This Week’s Finds for free, and that’s paid back immensely. So, giving away more will probably make me even better off.

On how one understands this ‘better off’ depends his rationale for virtuous action.

Posted by: David Corfield on December 20, 2006 6:52 PM | Permalink | Reply to this

### Re: Research Blogging

genuine fears

Exactly, that’s the problem.

On the other hand, given - as it often happens - that I have some thought which I really need to share (a thought may need more than one mind to grow), then I feel much more secure with posting it to a weblog, where it appears complete with time stamp, electronic signature and all.

Posted by: urs on December 20, 2006 7:16 PM | Permalink | Reply to this

### Re: Research Blogging

Maybe what you’re working on is distinctive enough that priority could be easily established. The question of when two research ideas are the same is an intriguing one. Are there known priority disputes involving blogs?

But perhaps we should look at the larger picture of the totality of a person’s blog research. After all, John is not just one episode of TWF. In philosophy it’s easier to convey one’s personality though print, in a book at least.

Posted by: David Corfield on December 20, 2006 8:05 PM | Permalink | Reply to this

### Re: Research Blogging

Hi Urs,

Thanks. Comments (sorry it is long and less well-structured than it could be):

(1) I’m not a big shot. But thanks anyway.
(2) I already talk about research on my blog. And it is more or less properly category-filed, for ease of finding it. so people can already link to it, as you say. Sadly, nobody really seems interested enough in my corner of the research sphere to say much in response. That’s ok. I’ve so far chosen to blog about the less mainstream aspects. My choice.
(3) I’d love to see more people do it, indeed. You and the others here, at the coffee table, and at musings, set an excellent and well known set of examples, of course.
(4) There is more to come, now that I have LaTeX installed over there.
(5) Expanding on (2): I should probably be more explicit and have a category called “research”, a subcategory of “work”. Right now it is dumped in with “string theory”, and that category is indeed chock-a-block with other stuff, as you know. Overall though, I am not convinced (Jacques and I argued about this a lot) that the fully distributed system of ping/track-backs is the best way to go as the primary mode for this global conversation, but I could simply be wrong. There’s some inertia to clicking on trackback links and leaving one discussion to fly off to another. Best ot have all the discussion primarily in one place. But I could be wrong.

(6) The point about public vs private. I seem to recall that the private aspect was pushed a lot in that early discussion on my first post due to two concerns. The first was that young people don’t like to ask silly questions -that get archived forever- when the whole world is watching. This is less of an issue for old folks like me who don’t care so much about future hiring issues and/or already have an established profile (for better or worse). Since the private blog I spoke of in the second post was primarily for the training of my young charges, a major privacy component was paramount, or they simply would not feel free to talk as though we were all sitting around a table somewhere. The second was the priority issue…. which is a sad one to get hung up so much about, as I commented back then, but nevertheless it is a major factor. And yes, it is a concern for everyone…. but as I said back in the first comment thread, I think that there are two levels of conversation to be had.. one is where speculative stuff is thrown out by different people…. Unexpected results may come from putting together things from differnet sources that way. Then there is the more touchy level of revealing details of an already well formed idea, and idea that gets completed by an opportunist before you get the chance. I know these are hard to separate out, but the former type is the one I’d like to see more of in a global conversation, for the good of the field. I personally would not care (too much1) if a half-idea of mine combined with a half-idea of someone else to become a result worked out and written up by yet a third person -if it resulted in real progress for the whole field, and if those half-ideas would never had lef to anything on their own. This is exactly how things work right now when you have conversations after giving a seminar or during coffee at a workshop or conference. I don’t see why people are so afraid of having them on the web, in an archived and searchable way. It would be a great resource indeed.

(1An acknowledgement of which conversation inspired them to figure it out would be nice and maybe even useful… as is sometimes done now… with the Conversation in place as an archive of the discussions, they could even point to specific posts/comments! But it is not that much of a big deal in the scheme of things…at least speaking in the abstract about it here. :-) )

I see others in this thread have picked up on some of these issues again. Good.

Cheers,

-cvj

Posted by: cvj on December 21, 2006 12:01 AM | Permalink | Reply to this

### Coffee catalyzes Communication

This is exactly how things work right now when you have conversations after giving a seminar or during coffee at a workshop or conference.

I wholeheartedly agree with this point – as already indicated in the last sentence of my entry above – that the useful kind of research-related discussion that we could have on blogs is more or less like the one we all enjoy most anyway: namely that taking place over coffee.

That’s, after all, why both the String Coffee Table and the $n$-Category Café have the names they have.

Whether or not such discussion requires lots of equations will depend. There is a reason why good conference places have blackboards in the hallway, since there may be a need to write down a symbol or two to help coordinate the thoughts of those conversing.

(I don’t think you would usually want to write down entire computations into a blog, but you might crucially want to display the central equation that your discussion revolves around.)

Whether or not anonymity is relevant may also depend. On big conferences it often happens that, as people walk here and there and join this or that discussion group, not everybody at every instant knows who everybody else is.

I certainly prefer to know who I am talking to, but if my only choice is talking to somebody anonymously or not talking to him or her at all, I very much prefer the anonymous conversations.

From time to time I receive comments on weblog entries by private email instead of posted to the comment section here. That should not be so. Chances are that somewhere out there is somebody who has an important comment to make on that particular comment, but he will never do so when it rests hidden in my private mailbox.

Therefore, I strongly encourage everybody here to post comments anonymously if that’s the only reasonable alternative to not posting at all.

An administrative comment in this respect: the comment edit window of the $n$-Café asks you for a name, an email address and a web address.

The name is what will be displayed. Put a pseudonym here if you desire. It will be hyperlinked to the web address given. Specify no such address if not desired.

The email address, however, will not be displayed in any case on the public pages of this blog. It will be visible, though, to the hosts of this blog, which includes, as you know, John, David and myself (and I guess Jacques can see everything, anyway).

So, if you feel like the hosts should know who you are, give your email address and then choose if the public will get to know your true name or just a pseudonym. If you feel nobody at all should know the name behind your comment - very well - just give some bogus email address.

Posted by: urs on December 21, 2006 10:40 AM | Permalink | Reply to this

### Re: Research Blogging

…I’d like to see more of in a global conversation, for the good of the field.

Never far from the scene is the language of Aristotelian ethics, two of whose central components are the notions of goods and virtues.

An acknowledgement of which conversation inspired them to figure it out would be nice and maybe even useful…

It would also be just. A failure to acknowledge displays a lack of the virtue of justice.

If a fully-fledged Aristotelian account of a tradition of intellectual enquiry is to be considered here, we should want to think about the good of mathematics/mathematical physics, both internal and external, and the good of the individual researcher as a member of a research community and of a political community. E.g., what does it mean for John to adopt a ‘gift economy’ with regard to his research, when his daily bread is provided by a capitalist economy? Is there an incompatibility between why a modern nation state thinks it should fund mathematics and the reason mathematicians do it? Should I keep my one small idea private when I can see it’s ripe for development by someone with a different background, so that I may later gain more recognition?

Posted by: David Corfield on December 21, 2006 10:41 AM | Permalink | Reply to this

### Re: Research Blogging

In regards to your worry that young people might not want to be archived for hiring concerns: Its a legitimate problem however isn’t that one of the nice properties about the web, namely that you can take anonymity behind a pseudonym (take for instance my name).

On some of the research boards I post with my real name and university affiliation when I know (or think I know) something so thoroughly im not ashamed to write it down. On other more tentative places where my domain of expertise runs to a breaking point, I can take refuge behind a pseudo and still hopefully contribute something or ask a silly question that invariably helps me.

So ultimately I think its better to just put the knowledge out there, and let the dice fall where they may.

Posted by: Haelfix on December 21, 2006 12:36 AM | Permalink | Reply to this

### Re: Research Blogging

Re Privacy, which is one of my major concerns. I think more good people, who by definition have lots of demands on their time, would be willing to participate in a conversation only if they find it useful. For that to be possible the medium has to be much, much (much) less noisy than the public blog. I think by now this fact is experimentally established.

Don’t get me wrong- the public blogs do a great job exposing research to the public, that is a good goal, and the public blog is nearly perfect for that. Research is a much more exclusive activity, and the medium has to reflect that.

Posted by: Moshe on December 21, 2006 1:46 AM | Permalink | Reply to this

### Re: Research Blogging

Hi Moshe,

I suspect that this issue of public vs private is not the one you’re thinking of. What I’m thinking of (or was in my first post in 2005) was something that was public in the sense that registered users from many different research groups can read and post. Private would mean something like I blogged about in this more recent post: a smaller group of people collaborating together or close to that. Public was not meant to be like the technical discussions we occasionally have on Asymptotia where any fool can come in and yell stupid nonsense -and they do, sadly- and reduce the signal to noise. That is not the public I meant, and I’m guessing that Urs also had that more reduced definition of public in mind too?

Making anonymous posts (as Haelfix suggested) is then also ruled out since you want people to take responsibility for their contributions…. otherwise it again reduces to a pub brawl. Alternatively, you elect one person a moderator who knows who all the anonymous people really are….. this is less than ideal. There should be no such heirarchy for the Conversation, in my opinion.

-cvj

Posted by: cvj on December 21, 2006 3:14 AM | Permalink | Reply to this

### Re: Research Blogging

I don’t think I agree… The signal to noise ratio at sights like this, Musings and the string Coffee table is pretty good.

I think its b/c the second things are written down in a technical or esoteric form, right from the getgo, it has an immediate tendency to ward off crackpots who thrive in places where there is no math. All it takes is really one moderator to remove spam bots and annoyances like that, as well as the considerably rarer but brighter breed of crackpot.

Posted by: Haelfix on December 21, 2006 3:22 AM | Permalink | Reply to this

### Re: Research Blogging

Haelfix,

Some (if not most) of the truly valuable physics research discussions involve few or no equations. A high quantity of equations and appearance of technical complexity for its own sake does not make a physics discussion, and certainly is not a good basis for how to shut out crackpots. No, I think that we should model it after the kinds of conversations that are had in the coffee breaks during conferences, workshops, and the like - no anonymity, and free exchange of ideas, comments and suggestions between colleagues who respect each other and come together for because it is of value to all their respective research programs. People should register to contribute. See the different levels of participation I suggested in the original post. That could also be tried.

Cheers,

-cvj

Posted by: cvj on December 21, 2006 4:47 AM | Permalink | Reply to this

### Re: Research Blogging

Thanks Clifford, we are in agreement I think. What will probably happen is that many people (myself included) will try all kinds of models for using electronic communication as means of enhancing different aspects of research, and see what works. Your initial idea, in the original post, seems very good to me, but is likely to be more of a “second generation” development, after the initial, more modest, models yield some benefits for their users.

Posted by: Moshe on December 21, 2006 3:35 AM | Permalink | Reply to this

### Signal to Noise

Clifford wrote #, in small part:

[…] any fool can come in and yell stupid nonsense -and they do, sadly- and reduce the signal to noise.

Haelfix notied #

The signal to noise ratio at sites like this, Musings and the string Coffee table is pretty good.

I think its b/c the second things are written down in a technical or esoteric form, right from the getgo, it has an immediate tendency to ward off crackpots

What Haelfix writes is indeed a remarkable empirical fact:

a weblog entry that has no component in sociology, politics, personal stuff and similar issues tends to attract little to no undesireable comments.

On the $n$-Café, Jacques’ technology fights off automatically all the automatically sent spam that arrives irrespective of what we do here. What remains is a small tiny number of really undesireable comments.

I’d say the word noise is simply not applicaple to the comment sections on any of the three blogs under Jacques’ administration.

And all indications are that the simple cause of this no-noise phenomenon is that the entries are focused on technical issues.

Posted by: urs on December 21, 2006 11:06 AM | Permalink | Reply to this

### Re: Research Blogging

About private versus open science discussions. First let’s hope that the scientist has a choice.

If you map the blog to a face-to-face discussion, then I have a relevant story. A couple of years ago I helped organize a medium sized workshop in Italy for the NASA/ESA/ASI Cassini mission scientists. The timing was 4 months after Saturn orbit insertion when the sense of discovery was at its height (hasn’t dropped much since, tho), and the scientists we spoke with wanted very much to meet in a forum that was private with no press and where the atmosphere would be relaxed and conducive to working through the biggest questions that that their volumes of data were giving. The workshop was a success. I’m sure that if, at that time, the press and public were involved, they would have felt too exposed to say off-the-cuff speculations, while in the environment we provided, it was perfect to speculate with no inhibitions.

I think that each type of forum: open and closed, has its best time and place, and the scientists themselves should know what will work well for any given situation.

Posted by: Amara on December 21, 2006 11:29 AM | Permalink | Reply to this
Read the post Blogs vs Wikis
Weblog: Musings
Excerpt: Research blogs, research wikis and math.
Tracked: December 21, 2006 2:55 PM

### Re: Research Blogging

More about privacy: I would advocate not merely restricted write access, but also restricted read access. One issue is that in order to feel comfortable making public statements, which may be wrong, or foolish, or controversial, a person would like to know they are talking to friends who are going to take it the right way. People reading may respond in some other forum, and the blogosphere is interconnected, so restricting write access is not enough.

I also like the option, usually not presented in blogs, or retracting or editing comments to make your points clearer. This is a model of a friendly group meeting among restricted number of people, who typically all know each other and would like to get a seminar room with a blackboard, and then close the door. It does not come instead of other models of communicating, but seems to me a more efficient way for developing the particular kind of conversation we have when we meet at conferences etc.

Posted by: Moshe on December 21, 2006 5:58 PM | Permalink | Reply to this

### Re: Research Blogging

Moshe writes:

More about privacy: I would advocate not merely restricted write access, but also restricted read access. One issue is that in order to feel comfortable making public statements, which may be wrong, or foolish, or controversial, a person would like to know they are talking to friends who are going to take it the right way.

This depends a lot on ones personality. Perhaps most people feel the way you describe. But when I make a statement that may be wrong or foolish, I like to make it in front of the largest audience possible — so I have the greatest chance of being corrected and learning interesting things! That’s why I’ve spent so much time sci.physics.research, writing This Week’s Finds, creating expository webpages, teaching courses online, and posting to the n-Category Café. I’ve learned a huge amount from all these activities, usually by people correcting wrong or foolish statements of mine.

When it comes to controversial statements, I don’t mind making them in a publicly viewable place. But, I avoid making them in a forum that permits easy feedback, because overall I don’t find the feedback useful: when people are hostile, they tend to stop providing corrective information and start providing personal attacks.

So: personality types vary. Instead of taking a specific fixed position on issues of privacy, a good online research environment will provide different options for different personality types. The challenge is helping collaborators with different views on privacy issues work together. Negotiations among the collaborators will always be necessary — but a clearly laid-out, broad menu of options would help a lot.

Posted by: John Baez on December 21, 2006 6:42 PM | Permalink | Reply to this

### Re: Research Blogging

The challenge is helping collaborators with different views on privacy issues work together. Negotiations among the collaborators will always be necessary — but a clearly laid-out, broad menu of options would help a lot.

What might be a concrete implementation of that idea, software-wise?

What comes to my mind is this:

Label every entry/comment on a weblog with two pieces of information:

a) a specification of the useres allowed to read the post

b) a specification of the useres allowed to reply to this post.

Posted by: urs on December 21, 2006 6:49 PM | Permalink | Reply to this

### Re: Research Blogging

the greatest chance of being corrected and learning interesting things

I seem to recall a quote by Feynman, which I liked a lot. Paraphrased, it said something like:

Doing research is not about making no mistakes. It is all about finding the mistakes one inevitably makes as quickly as possible.

Posted by: urs on December 21, 2006 7:50 PM | Permalink | Reply to this

### Re: Research Blogging

Thanks John, that makes complete sense. As I said earlier, probably lots of models should be attempted, I am not suggesting replacing current models with others. The experimental fact I alluded to earlier is that extremely small number of my colleagues participate in the various conversations as of yet, and based on private conversations the privacy issue seems to me an important one.

Posted by: Moshe on December 21, 2006 8:23 PM | Permalink | Reply to this

### Re: Research Blogging

I would like to use blogs and/or wikis as part of a new ‘open-source’ method of collaboratively writing expository papers, a bit like Linus Torvald managed to get lots of people to contribute work on Linux.

I have occasionally had people take it upon themselves to convert some of my webpages into TeX, or otherwise improve them.

But, I was really delighted when with no prompting on my part, Apoorva Khare took it upon himself to TeX up the Fall course notes on Quantization and Cohomology.

I’m also delighted that Blair Smith has TeXed my course notes on Classical Mechanics — more on that later.

These notes may eventually be published, and I’ll gladly make anyone a coauthor who contributes a significant piece of work to a publication like this. That should be sufficient incentive — especially since people are starting to do this work even without that incentive.

But the real trick would be make it easy for people to find out which pieces of work are available to be done.

How do they do this in the Linux project?

Posted by: John Baez on December 21, 2006 6:03 PM | Permalink | Reply to this

### Re: Research Blogging

I now feel like a fool. Having asked (via e-mail) whether the classical mechanics course notes had been put into LaTeX, I switched over to my Firefox window and discover the answer.

I know very little about the innards of the Linux developer community, but on Wikipedia, a significant part of the communication seems to happen via the WikiProjects, such as WikiProject Physics, and their associated talk pages. There is also a great deal of highly decentralized discussion on the talk pages of individual articles and of individual users.

One important way that people know a new article is needed is the redlink. If one article contains a link to a wiki page which does not yet exist, that link shows up red instead of blue (in the default skin). This might also work if somebody set up a course-notes wiki.

Posted by: Blake Stacey on December 21, 2006 6:24 PM | Permalink | Reply to this

### Re: Research Blogging

happen via the WikiProjects

in case you missed it, Jacques has, meanwhile, discussed that aspect here.

Posted by: urs on December 21, 2006 6:28 PM | Permalink | Reply to this

### Re: Research Blogging

This might also work if somebody set up a course-notes wiki.

As long as one doesn’t want to hold on to too much authorial control, one can always set up a Wikibook.

Posted by: Toby Bartels on December 25, 2006 1:37 PM | Permalink | Reply to this

### Re: Research Blogging

I’m not a Linux contributor, but as a computer science guy I can make some comments about Linux. Three important points that maybe don’t apply to research are:

(1) “I need this” – people mostly start working on things because they either need some new functionality, or what’s currently there doesn’t work well for them. Unless you’re signed up to teach a course on something, what you need is more likely to be research based rather than expository.

(2) “Partially objective success criteria” – Whilst there’s a certain amount of review criteria which use subjective things like “long term maintainability”, part of the judgement of quality comes from objectively running the computer code on real machines; if it has problems it clearly isn’t suitable. But suppose I write a review of some subject area, unless I get things really wrong your view on it will be primarily based on subjective criteria like “did I introduce ideas in what you think is the wrong order”, is the style “not to your liking”, have I written about a (valid) approach which isn’t your preferred one? If the feedback you’re sending to a web-collaborator that you don’t really know is of mainly of this form it’s likely they’ll give up. (Indeed, most of the friction in Linux development comes when things technically work but there’s loads of subjective “change requests” that aren’t objectively either right or wrong.)

(3) Lots of the advances in Linux are accumulated small refinements to existing things (maybe even just 2-3 line changes) because these are easier for people to complete and move on from in short amounts of time. In expository text would you be
happy with a document that slowly evolved from a text about basic arithmetic to the distinctions in cardinal and ordinal counting over the course of a couple of years?

As a side remark, note that whilst Open Source coders produce amazing code they aren’t as good at documentation! (Even I haven’t documented my one released program nearly enough.)

I think more collaborative work in science and engineering would be a great thing and Open Source computer software is a good place to take inspiration from.
But I think we all need to think about the analogies and differences between code and exposition.

On another note about blog collaboration, one of the problems I find with talking about my work (in computer vision) is the “critical mass” problem: I’m not particuarly interested in “vague conversations where no-one really engages enough to say anything thought provoking” (to be honest because I find it time consuming without being intellectually rewarding). On the other hand, discussing interesting stuff requires more than a fleeting conversation but a real engagement. That’s why I don’t really find quick conversations at conferences of much interest; spending a couple of hours in a pub or getting stuck next to someone on a train is much more interesting to me. The connection with blogging is: the text of even long blog conversations would take 5-10 minutes to speak. What sociological features would promote longer interactions in the blog/general internet arena? Eg, JB and DC have clearly spent a lot of time doing various 2-geometry stuff: other than the intrinsic interestingness of the subject why did that become a long conversation and other things don’t?

No answers, just food for thought.

Posted by: dave tweed on December 21, 2006 7:21 PM | Permalink | Reply to this

### Re: Research Blogging

dave tweed says:

Lots of the advances in Linux are accumulated small refinements to existing things (maybe even just 2-3 line changes) because these are easier for people to complete and move on from in short amounts of time. In expository text would you be happy with a document that slowly evolved from a text about basic arithmetic to the distinctions in cardinal and ordinal counting over the course of a couple of years?

The analogous problem occurs at Wikipedia, as well as the subject-specific wikis with which I am familiar. The majority of edits will by nature be minor ones, affecting and being affected by only their immediate surroundings. Typically, a small group of editors works very hard to improve an article in which they have an interest to a local maximum. After they move on, other people come along, adding bits of this or that — everyone mentioning their favorite Calvin and Hobbes strip, say, or tossing in the latest breathless pop-science article about the String Wars. These edits, each one of which is made in good faith, collectively bury the article’s organization under geological strata of cruft. Then someone notices the problem and puts in the effort to clean it up, and the cycle continues. Overall, the motion may be forward, but it can sure be a Brownian path getting there.

This problem might not matter so much if the intent were to produce collaborative course notes or co-author a textbook. In that case, each page might be intensively “workshopped” for a month, and then the whole thing is exported to a PDF. Instead of continual revision, with its inherent hazard of continual degradation, the system would operate more like a process of successive editions.

Posted by: Blake Stacey on December 21, 2006 8:16 PM | Permalink | Reply to this

### Re: Research Blogging

I think this is implicit in what you’ve said, but one of the things I had in mind was: if you’ve written a text on basic arithmetic that you use for a semester, basic arithmetic might be what you want. Coming back a year later, whilst all the stuff about ordinals might be correct and well written, it might not be suitable for what you wanted the document for. (Of course you’ve hopefully saved an old copy of the document.) Slow-change-of-purpose in software generally goes hand in hand with people upgrading their hardware so it’s rarely a problem.

Posted by: dave tweed on December 21, 2006 9:05 PM | Permalink | Reply to this

### Re: Research Blogging

Thank you for elaborating.

It’s worth noting that MediaWiki saves all the old revisions of each page, so you can in principle retrieve an old version of your wiki document — if you can find it! If the page has been heavily edited, extracting a particular previous draft begins to resemble an archaeological expedition.

This is one reason why the Wikipedes have started tagging the talk pages of their “Featured Articles” with a template which specifies the version which was first “uplifted” to Featured status.

Posted by: Blake Stacey on December 21, 2006 9:17 PM | Permalink | Reply to this

### Re: Research Blogging

Dave Tweed listed some reasons why people contribute to Linux:

(1) “I need this” – people mostly start working on things because they either need some new functionality, or what’s currently there doesn’t work well for them. Unless you’re signed up to teach a course on something, what you need is more likely to be research based rather than expository.

Indeed academia is quite different from the Linux community, so the comparison is a bit strained.

“I need this”: what all academics need is more publications on their resume. This is especially true of young folks: grad students and postdocs. Young folks also especially need to learn lots of stuff — or at least, they’re more aware of this need than old fogeys.

So, when it comes to getting people to help me turn my lecture notes into publications, I’ve mainly been successful with young folks. My new idea is to advertise opportunities to help out world-wide, using this blog, instead of just asking grad students and postdocs here at UCR. It seems to be working!

(3) Lots of the advances in Linux are accumulated small refinements to existing things (maybe even just 2-3 line changes) because these are easier for people to complete and move on from in short amounts of time. In expository text would you be happy with a document that slowly evolved from a text about basic arithmetic to the distinctions in cardinal and ordinal counting over the course of a couple of years?

That would be okay — but I’m very fussy, so I’m rarely satisfied with anyone else’s ability to explain things. This is why I personally want a lopsided, somewhat dictatorial collaboration where I give a bunch of lectures, then someone else writes it down, then maybe someone else turns it into LaTeX, then maybe someone else adds pictures, and then I edit the result until I’m happy. Young folks are often perfectly willing to go along with this lopsided approach, since they know they’ve never written a book before.

Lots of lecture notes are already available in handwritten form here. The challenge for me is to push this stuff further down the pipeline to publication. There are even half-written books languishing in this pipeline. The problem is that I always prefer explaining new stuff.

Mind you, I expect that lots of very different forms of collaborative exposition are possible! Right now I’m just talking self-centeredly about my own issues. But I should mention something much more important:

When it comes to “I need this”, what the mathematics community desperately needs is a good free calculus textbook with lots of homework problems. Right now big publishing companies are sucking up way too much money from students! Students often don’t read these books; they just buy them to know what homework problems have been assigned. Solution? How about a really good Calculus Wikibook!

Of course, it’s not just calculus where the books are too expensive — the same problem arises in other subjects. We need free Wikibooks on all academic subjects! As a math professor, though, it’s calculus where I see the problem most vividly on an everyday basis.

Posted by: John Baez on December 23, 2006 12:11 AM | Permalink | Reply to this

### Re: Research Blogging

[Disclaimer: I’m an outsider of both the OSS and Wikipedia communities, but have been using both for a while. So what follows is considered opinion but should be treated cautiously.]

I think there’s a risk here of confusing Wikipedia and OSS (Open Source Software). Both are based on the same idea: that there is more wisdom in the crowd than in any subcrowd. However they differ in how to extract that wisdom.

The difference comes about because they have different tests for distinguishing “good” wisdom from “bad”. The aim of Wikipedia is to produce accurate information. That of OSS it to produce useful software (and, often, fun and silly software. Sometimes all three at the same time).

The Wikipedia model is to attempt to create a monotonic sequence of articles converging to a completely accurate one. For most articles, this works; perhaps we could say that accuracy is almost a total ordering. Occasionally this goes haywire and the sequence goes oscillatory or divergent; then the admins have to step in and add a forcing term.

The OSS model, on the other hand, is more evolutionary. Though, at the risk of starting a flame war, it has a touch of intelligent design about it as well. The notion of “useful” is harder to pin down than that of “accurate” so open source software tends to branch and fork as people modify it and add features that they themselves find useful; these are then “released into the wild” and everyone sits back and watches to see what people find useful. Thus there are often many versions of a particular program available and one can pick and choose. This isn’t often apparent to “mere” users as the selection has already been done by the distribution. But if you disagree with that selection, you are always free to install your own version - for example, I use a version of emacs modified to allow a transparent background.

The “Intelligent Design” part comes in because an open source program generally has an official version maintained by a small group of people. “Maintained” but not necessarily written as they can add patches from the other variants as they like.

The OSS model is probably more what you (John) are looking for. You have the control, but you can allow others to work on sections and then merge them into the main branch if they meet your approval. You can be highly selective in this (much more so that I think you can be in the Wikipedia). As far as I know (which, admittedly, is not all that far; oh, about two ells) a Wikipedia article is not owned by anyone and so no one person has any more rights to its authorship than any other.

It is true that in both Wikipedia and OSS the majority of changes are minor (bug fixes, one might call them). But if you produce a useful set of lecture notes then you will get these anyway. The difference is that in an open source model the person telling you that ‘the word “that” on page 100 should be a “who”’ will, hopefully, also supply the patch so that all you have to do is accept or reject their suggestion and you don’t have to search through lines and lines of TeX to find the right occurrence of the word to change. However just because the majority of edits will be minor is no reason not to do it. In amongst all of those will be gold.

In fact, OSS actually has exactly what you want. It’s called the Google Summer of Code. In brief, Google sponsor students to work on projects suggested by, essentially, the open source community. If you could get the AMS (or other) to do something similar then you’d be done.

How about it? The “AMS Summer of Lecture Notes”. Lecturers (and their students) submit proposals to typeset lecture notes to the AMS and the AMS selects those it feels are worth doing (and possible to do) and sponsors students to do the work. The AMS gets a nice set of lecture notes and the student gets a chance to learn some mathematics, and how to write it as well.

Of course, if the AMS won’t do this then you can always launch it on your own. The problem then is recruiting people. The open source community is, I suspect, much larger than the mathematics one and even there the majority of programs only have one or two people working on them. Still, if you offer an incentive (such as, “You won’t get a degree unless you typeset my lecture notes.”) someone will probably sign up. I once spent a break typing something for a bottle of sherry and I don’t even drink the stuff! (I also thought it was time to learn TeX, and it was a problem set not lecture notes, but this is a blog not a Wikipedia article so accuracy isn’t the main criterion).

HTH

Posted by: Andrew Stacey on March 31, 2008 10:26 AM | Permalink | Reply to this
Read the post Lectures on Classical Mechanics
Weblog: The n-Category Café
Excerpt: Lectures notes on classical mechanics - now available in printed form!
Tracked: December 21, 2006 7:39 PM

### Re: Research Blogging

Hello urs!

Your post was one of the reasons I finally started my own blog. Not really a research blog, but rather learning blog.

I hope very much that learning (and possibly doing research as well) is possible in a little bit “cafe-like” manner. However, few years ago I told one very young, succesful postdoc that I like to learn math with others, talking, drinking, etc, and he answered something like “Well, it is not this way, Luke. My feeling is rather that doing math is like being a lonely old wolf”. He said it more poetically, though :-).

The address is sirix.wordpress.com, I just started so for now there is only 1 article - report from lecture.

Posted by: sirix on January 4, 2007 6:28 PM | Permalink | Reply to this

### Re: Research Blogging

Just because you’re a mathematician, doesn’t mean you need to be a lonely old wolf. Cafés have always been popular among mathematicians. But now that the internet, web, and blogs exist, you can learn a lot and have a lot of fun talking to people around the world — without even going to conferences! I would have been really bored in Riverside without the internet.

So, good luck to you and your new blog!

Posted by: John Baez on January 5, 2007 5:51 AM | Permalink | Reply to this

### Re: Research Blogging

[“Always” above redirects to information about the Scottish Book]

One of my teachers here, in Poland, told me several apocryphal stories about the Scottish Cafe (coincidentally, I’ll see him tomorrow, so I ask him for any references), which is sometimes perceived by me and my colleagues as an ideal of the mathematician’s life ;-):

1. Contrary to information in a link above, it was not Banach who bought a notebook which became a Scottish book, but his wife, on the request of cafe waiters, because Banach and friends were writing on table-tissues and tables.

2. It is known that Banach was an alcoholic. One of his friends, possibly Mazur, used to come to Banach every evening (or earlier), make a mock Indian-like bow to Banach and say “Chieftain, lead me.” And they were gone to the Scottish Cafe for a drink and mathematics.

3. Amounts of alcohol drunk in Scottish cafe are legendary. Somebody points that there were few 3-day-without-a-break periods of heavy drinking / doing research and at least one 7-day such period. After the latter one, the story says that the whole company knew that they were doing things incredibly important, and proved theorems of great beauty and prominency. Unfortunately, nobody could recollect what exactly they were doing.

Posted by: sirix on January 5, 2007 10:10 PM | Permalink | Reply to this

### Re: Research Blogging

New Scientist on “Finding the best science blogs”.

Posted by: Thomas Riepe on March 2, 2008 1:23 PM | Permalink | Reply to this

Post a New Comment