Wikipedia & Google: Free Schooling?

[Chris Hillman]

Flamebait?

Well, raolduke, there are many things you probably haven't thought about:

1. There are profound and deeply troubling ethical issues involved in trying to track edits by individual identity.

2. There are challenging technical issues involved in such tracking which you obviously haven't yet considered, and which I think we neither need nor wish to discuss here.

3. The scary thing is that those who know a little bit about this stuff assume the technical issues can't be overcome, but those who know more know that given sufficient effort and resources (available to, let us say, Google, Choice Point, or maybe even the government), know that to a great extent they can be.

4. These rather inflammatory issues are but a small part of a much larger social issue involving the troubling implications of the very large scale tracking/recording/monitoring the existence of individual citizens by private and governmental agencies. For a non-hysterical non-technical study, see No Place to Hide, by Robert O'Harrow, Free Press, 2005. But I can tell you that the situation is actually rather more grim than he describes.

In my deleted user space edits I argued that the very first step to addressing WP's quality control problem has to be eliminating edits by unregistered users, and officially proscribing socks (so that those who create socks at least know they are breaking rules and can be summarily evicted if caught).

Jimmy Wales appears to be moving ever so slowly in that direction, but there is a profound internal conflict in WP culture between (1) WP as a vast MUD for "anon blogging" (2) WP as a public information resource. I argued for splitting off "wikispeech.org" from "wikipedia.org" as a first step permitting rational discussion of policies including behavioral rules (wikilaws) appropriate for these two very different functions. Obviously, for wikispeech.org, secure anonymity (very hard to achieve, as it turns out!) is of paramount importance. For wikipedia.org, very different considerations take precedence. I argued for a "constitution" specifically designed to avoid the need for "user monitoring" on the vast scale you envision. I actually proposed something not dissimilar to a karma system, but it is clear that this would be tolerated only by users deeply committed to the ideal of volunteer service in working toward a common goal, users willing to sacrifice considerable amounts of their privacy while working on the project.

All of these issues are so inflammatory within the WP community that rational discussion of them has proven perennially impossible. And even if discussion were possible, changes are almost impossible given the fact that, as I already pointed out, the WP procedure for discussing policy changes is hopelessly impractical. I feel this stunning inefficiency is one of the main reasons why, as Stacy Schiff put it, "Wikipedia has become a regulatory thicket".

(An irony: Stacy Schiff has written a book about V. and V. Nabokov; many of Vladimir's novels center around the ambiguity of personal identity, so it is no coincidence that her analysis of WP is so perceptive.)

 

[mathwonk]

I am in math, and I have often been able to learn something apparently correct in my field from wikipedia.

In my career as student I have encountered roughly three levels of expertise. In high school teachers it was lowest, they often did not understand what they were teaching and knew very little.

In college there are two simultaneous levels presented, that of the textbook, and that of the professor. The professor at a good school usually knowing more and understanding more than than what is in the average calculus text. But in an honors course at a good school, one often meets a book as good as a good professor.

Wikipedia artciles seem, in comparison to that spectrum, often at or above the level of the average text. I.e. not incomptetent, usually quite knowledgeable, seldom at the level of a good prof at a top school, but sometimes.

Unfortunately sometimes the article on a math topic, say tensors, seems to have been written by someone who uses them and thinks they understand them, but not at the level of a mathematician. so the discussion is about how to manipulate them rather than understand them.

presumably other areas suffer the same incursions from posters who do not realize they are not experts. still their aeticles offer something. Perhaps this is less dangerous in math, where there is a test for correctness of every statement, namely proof.

I.e. since in math we do not take anything on faith, we are insulated from harm by crank opinions, or non expert discussions. I.e. they just don't help much, but they do little harm. so in the flawed tensor discussion alluded to, one can still learn something, just not everything.

 

[raolduke]

I thought I added something about it being costly but I believe it can be done.. I agree with you... Its not practical in the slightest but if there are people who really want to have wikipedia as a credible source of information then what has to be done will be done.

 

[Chris Hillman]

Trying to ward off potential confusion

Hi, mathwonk,

I agree that at present math articles at WP are much less likely to be cranky/incoherent/generally awful that physics articles, and I also enjoy reading them. However, regarding your somewhat delphic comment "in math we do not take anything on faith, we are insulated from harm by crank opinions, or non expert discussions", I must demur.

First of all, I assume you meant "in research mathematics we are insulated". Ideally, that would be true and I dare say this ideal was a crucial aspect which initially attracted us to mathematics, since in other endeavors one can hardly make this claim even as an idealization.

Second, I think this remark, even if it had been less ambiguous, is potentially misleading in the context of this thread, since it distracts attention one of the crucial points I and others made above: while trained mathematicians have mastered many "bull**** detection" skills and are capable of quickly spotting many mathematically invalid claims, most students consulting WP have no reliable way of telling which articles are (in the version they view!) honest and reliable accounts and which are cranky. This is unfortunately just as true for math articles as for articles in other highly technical subjects. Over the years there have been, unfortunately, many cases of putative "math articles" at WP written by cranks/ignoramuses which have been discussed at WikiProject Mathematics. In some of these cases it was clear that students could in fact have been fooled. Unfortunately I did not succeed in getting permission from those I asked to quote from emails from individuals willing to confess in private that they are students who were in fact fooled by ludicrous math-related WP articles. The stunning aspect of these confessions was that they concerned articles which you or I, after a mere glance, would consider to constitute "obvious nonsense". One of the points which is difficult to express but which needs to be made is that I feel that experts can often find it very difficult to appreciate/recall how naive untutored but mentally "normal" youngsters can be.

[Edit: oh, that's nifty. I've been here for quite a long time and I never knew until now that PF autobleeps certain words, or in this case, a suffix.]

Third, while discussing this would get OT (start a new thread?), the assertion that mathematicians take nothing on faith while performing their research duties is questionable, and I think most mathematicians will conclude, after a bit of reflection, that they take quite a bit on faith. (I've had discussions with several mathematicians about this very issue, hence my confidence regarding "most mathematicians".) I'd add that at the very least, the suggestion that mathematicians (or Mathematics?) is magically "insulated from cranks" requires qualification.

 

[Chris Hillman]

I repeat, NOT a good idea!

I thought I added something about it being costly but I believe it can be done.. I agree with you... Its not practical in the slightest but if there are people who really want to have wikipedia as a credible source of information then what has to be done will be done.

Last year, in internal discussions at WP, I attempted to frame my conclusions as a binary choice:

(1) split wikispeech.org off from wikipedia.org and create wikiconstitutions, wikicitienships, wikilaws, wikiexecutives, and wikicourts appropriate to each mission (an appalling propspect indeed, but every highly complex endeavor requires regulation),

(2) implement elaborate monitoring of individuals (correlating sockpuppet accounts and anon ip edits with individual users).

Unfortunately, I and others have found that even attempting to discuss the technical possiblity of (2) at WP itself, even in the course of trying to argue that (1) is the only reasonable choice, terrifies many Wikipedians who post under an "anonymous" handle, including some very prominent Wikipedians, who either hadn't known their vulnerability, or had known but didn't want anyone else to know, sometimes with good reason One of the subtexts which may or may not be apparent from discussion with WP of various internal scandals is that what goes on "behind the scenes" tends to be much more disturbing than the genial public face of WP as represented by the charismatic character of Jimbo Wales. (This might be a good place to recall that Wales himself tolerated my user space essays; so while I have questioned his suitability for the role of God-king of wikipedia.org rather than wikispeech.org, I don't question the sincerity of his tolerant libertarian/populist ideals.)

The sad irony is that WP has always been moving toward (2) and since I left, has moved further in that direction. That's ironic because the monitoring (both internal and external) is not on a scale sufficient to deal effectively with the abuses, but is sufficiently capable and intrusive to raise many of the grave ethical concerns I alluded to above. IOW, WP is suffering most of the social costs of user monitoring but receiving little of the desired benefit for the (allegedly) encyclopedic goals of the project.

 

[mathwonk]

you are quite right that not only do we mathematicians take some things on faith, but that our students often have not developed the ability to test statements by proof. so i was indeed over optimistic as to the paradise we supposedly live in. so it is rather an ideal we strive for but often, or even usually, miss.

still it makes a good talking point, until someone perceptive punctures it. and indeed in puncturing it illustrates the power of logic and cogent argument.

 

[Chris Hillman]

Thanks!

The (virtual) pen is still mightier than the stilletto

 

[mathwonk]

well i just had an interesting experience. I started to edit the very minimally useful, and arguably inaccurate Wikipedia article on the Johns Hopkins days of Oscar Zariski, a man and a mathematician whom I knew as a teacher, and whose field is my own specialty, and whose collected works I received as a wedding present.

Moreover almost anyone could improve on the description there of the Zariski topology, which contains essentially no information on that topology and a very questionable negative assertion as to its usefulness.

Indeed none of the assertions there satisfy a mathematicians desire for a logically defensible, fact based presentation.

In spite of this, and in spite of the fact the page is already flagged as marginal in quality, I could not bring myself to set up as an anonymous expert and alter even one word.

At least here my statements and opinions are labeled as my own.

 

[Chris Hillman]

Zariski topology "not very useful"?

Hi, mathwonk,

almost anyone could improve on the description there of the Zariski topology, which contains essentially no information on that topology and a very questionable negative assertion as to its usefulness.

Hmm... I ran right over and grabbed this permalink to the version which you must have examined: http://en.wikipedia.org/w/index.php?title=Zariski_topology&oldid=128708007

Hmm... I'm not seeing the "very questionable negative assertion as to its usefulness", but neither is the first paragraph suitable for beginning graduate students, much less undergraduates, much less the janitor.

[EDIT: That was because, as mathwonk points out in post #81 below, he was in fact looking at a different WP article.]

I noticed that one of the major recent editors says he's a second year grad student at Harvard and adds that he's particularly proud of this article(!): http://en.wikipedia.org/w/index.php?title=User:Ryan_Reich&oldid=127458158 My guess is that he's a deeply involved in learning about Grothendieck and got carried away. A moment with Google gave www.math.harvard.edu/~ryanr/ so there is some reason to think this user is not making up his academic credentials out of whole cloth, but this is a possibility that WP readers should always bear in mind: http://en.wikipedia.org/w/index.php?title=Essjay_controversy&oldid=128378085

 

[mathwonk]

the assertion that the zariski topology is only useful for regular maps is absurd. the statement about the necessity of considering valuations as a method of studying blowups is very dated in my opinion.

grothendieck's etale topology is merely a variation and refinement of the zariski topology, as would be clear if the poster explained anything about either of them.

 

[mathwonk]

you examined the wrong article. the one i referred to was the biography of zariski, in the section on his hopkins years.

 

[mathwonk]

the article you refer to on the zariski topology, although very detailed and lengthy, is still a bit tedious to my taste and makes a big deal out of a trivial concept.

namely an algebraic set is one defined as the zeroes of polynomials. in the zariski topoogy, the closed sets are simply the algebraic sets. that's it. so on a curve, the proper closed sets are finite sets. on a surface the proper algebraic sets are unions of finte sets and finite sets of curves.

grothendieck's etale topology is merely the refinement where one uses the same open sets but considers them as maps from open sets into the space which need not be inclusions. then there is no need for the maps to be injective, and one allows coverings of zariski open sets to be called open "sets".
but it is obviously still based on the zariski topology.
thats all.

wikipedia belabors this for two pages or more.

 

[matt grime]

[Edit: oh, that's nifty. I've been here for quite a long time and I never knew until now that PF autobleeps certain words, or in this case, a suffix.]

Yep. Not just suffices either. In my hasty typing I have frequently misspelled function by reordering the letters and had it asterisked out (think about it).

Anyway. I've found Wiki quite useful as a tool when I don't want to write out an explanation here of something well known, though I usually scan the article first to make sure it contains some relevant information. This seems reliable on well known topics, e.g. if some asks 'what does lagrange's theorem state?'

But that is just for maths, where opinion matters less, and we're not talking the latest fashionable material as I imagine physics students might read.

However, I try to use planetmath or wolfram as the first line of reference on the web rather than wiki since they tend to be more wide reaching and more reliable respectively.

 

[mathwonk]

Ok I read the page on the zariski topology in wiki, and it is essentially correct, but it reads as if it were written as a school project by a student. It has the officiousness of an article by an older authority, but it is tedious, overlong, and not as insightful as one by a knowledgeable person would be.

It has a bit the tone of, I am telling you something arcane and wonderful, but I would rather ballyhoo how cool it is than just tell it to you. I.e. it tries to make it look fancy rather than simple. The account in any textbook seems preferable.

And it also seems to me to be filled with pejorative implications about the importance of the zariski topology that i find inaccurate. I mean who cares if it is hausdorff or not? as shown in any beginning book, the key point of hausdorffness of Y is that 2 maps into Y should agree on a closed subset of their domain.

This also holds for the zariski topology of affine and projective space. there is nothing wrong with the zariski topology. what is wrong is our outdated habits of focusing on the form of definitions, instead of on the esential mapping properties that come out of them.

Indeed grothendieck emphasized this, but this emphasis only reveals the fact that the zariski topology does have most of the good properties one wants for varieties. the key point is that although the zariski topology of a variety seems weak, the zariski topology of a product is in fact much richer than the usual product topology would be.

this is made very clear in mumford's famous redbook on varieties, but not even mentioned in wiki.

 

[mathwonk]

i am going to conjecture that these articles read as they do because they were mostly written by the founders of wiki, in an attempt to get the project off the ground. I.e. suppose I want to start an encyclopedia without plagiarizing, so i do a little research on a lot of topics and write articles on them based on what I have been able to learn quickly.

Then I invite more knowledgeable people out there to come in and edit and improve them, hoping this will take care of necessary shortcomings of their origin. so theoretically the project will only work if the open editing process works.

This way you start an encyclopedia founded not on the expertise of the authors of the project, but of that of the audience. so those of us criticizing it are actually the ones expected to do the work of improving it.

 

[Chris Hillman]

the assertion that the zariski topology is only useful for regular maps is absurd. the statement about the necessity of considering valuations as a method of studying blowups is very dated in my opinion.

grothendieck's etale topology is merely a variation and refinement of the zariski topology, as would be clear if the poster explained anything about either of them.

I hope everyone realizes that mathwonk have provided here several excellent examples of the phenomenon I was talking about: in many cases only an expert can spot nonsense.

I think I'm converting mathwonk to my view that it is important that faculty speak out on the dangers of using Wikipedia naively. I am probably much more dubious than he that WP has much use for the average citizen seeking reliable information, but I think we all agree it ought to be good for something.

One thing Wikipedia is good for is a fishbowl for sociologists, psychologists, ethologists, and anthropologists (not to mention network engineers, systems theorists, and so on) who wish to study humans interacting "in the wild". This is because edits from user accounts and IPs are fully logged and timestamped, and this data (and much more) is publically available to anyone interested. More and more academic papers on WP (the arXiv so far mostly offers CS papers, but search history, sociology and psychology elsewhere) in fact study various sociological phenomena using WP as a rich source of data, as indeed it is. Of course, this reflects that fact that WP is not simply a website offering a putative information resource; at least nominally, it is also a utopian social experiment which provides a technosocial environment for a large scale collaboration. In practice, this seems to mean that Wikipedia functions as a large scale role-playing MUD.

Another thing WP is being used for is market research. Private information brokers like ChoicePoint keep vast amounts of data on individuals. Allegedly, currently about half of all persons now living, including virtually everyone in "the first world", are the subject of individual dossiers which document biographical data, medical records, financial/banking/insurance data, phone records, travel data, lifestyle choices, etc.---dossiers maintained by ChoicePoint and other information brokers, for purposes of selling this information to essentially anyone willing to pay. WP offers a treasure trove of data on the interests of said individuals--- often including interests they might not wish their family/boss/government to know about, but which information brokers think they have a right to know-- and to sell, e.g. to companies with marketing or personnel departments, or to politically/religiously repressive governments (see BBC for specific examples). The U.S. government has more or less acknowledged that the resources of such information brokers, Google, etc., outclass its own computational resources (which are often fragmented, outdated, etc.), and has persistently tried to gain complete access to these databases, on the grounds such access is (they claim) needed for national security, mostly meeting (it is said) with little resistance.

One of the points I tried to make in my deleted user space essays at WP--- unfortunately I was shouted down there--- was that WP and other MUDs have a responsibility to make it clear that "anonymity" is mostly a myth, that companies like ChoicePoint have the resources to track individuals across different user accounts. Currently information brokers are effectively unregulated in terms of what information they can collect and how they can use it. Anyone who assumes that their social security number or medical records are confidential should definitely read the book I cited, which is actually the least hysterical account I've yet seen. (Before someone asks--- yes, ChoicePoint has sold dossiers to people who turned out to be identity thieves. Information brokers not surprisingly insist it is not their job to vet their clients, and it may be plausible to assume that most thieves prefer to buy the data they need to steal money from other thieves rather than from a "legitimate company".) And all netizens should know about "cyberSLAPP lawsuits" http://www.cyberslapp.org/about/page.cfm?PageID=7

There is also much good information available from organizations like the ACLU http://www.aclu.org/privacy/index.html
the EFF
http://www.eff.org/
and EPIC
http://www.epic.org/
Also, BBC News
http://news.bbc.co.uk/
frequently runs stories on major compromises of private and governmental databases (e.g. someone looses a laptop containing millions of records containing sensitive personal information--- this happens all the time, and typically no-one affected is told what happened). See "Breach Report 2007" at http://www.idtheftcenter.org/artman2/publish/lib_survey/Press_Release_-_2007_Breach_List.shtml
for some examples (largely independent from the incidents described in BBC news items, which you can search for using their search tool or using Google).

Returning to possible academic responses to the rise of the Wikipedia: I've been mulling arguing that the leaders of American mathematics consider suggesting to the AMS that the Society suggest to department chairs that they consider requiring their graduate students to write mathematical content for the WP for a two credit course in "mathematical exposition" or something like that. But now I am reconsidering the wisdom of that, even if I could persuade said leaders. (Happily, quite a few of them are already interested in mathematical exposition and mathematical education at all levels generally.)

Two-credit course: at one point the UW Math Department experimented with such a course. I don't know if they still offer such a thing, but quite apart from possible WP assignments, I think such courses are an excellent idea. In fact I think writing expository articles is something of a professional duty, which--- again Halmos himself stressed this point--- is sadly undervalued by the academic rewards system.

it reads as if it were written as a school project by a student.

Well, maybe someone at Harvard had the same idea I had. As I pointed out above, the version of the article on "Zariski topology" we are looking at was mostly written by a current Harvard graduate student. It sounds like we are all agreed it "isn't encyclopedic", as Wikipedians put it.

I am going to conjecture that these articles read as they do because they were mostly written by the founders of wiki, in an attempt to get the project off the ground.

Not sure if you are serious, but the "history" tab shows who has contributed to a given article, and sometimes user pages offer information which can be verified. I haven't looked at the wikibio of Zariski, but the distinct article in the specific version I glanced at appears to have been mostly written by the graduate student I mentioned.

 

[mathwonk]

Thanks Chris for the tip about the hostory link. I am obviously a beginner at WP.

The zariski topology article has evloved from a few sentences that to me were almost more useful, into a lengthy but naive student level article, mostly definitions, with little insight.

Just the sort of thing a beginning student would write indeed.

How do you tell where the student is from? Becasue he mentions Joe Harris by first name?the article is not so bad, once you know the source of it, its just obvious it was not written by say a fields medalist.

Is it typical that grad students write these articles?, as they usually do not write books.

 

[mathwonk]

ok a brief browse on wiki reveals that apparently one guy, named charles matthews, with a phd on gauss sums 10 or more years ago, and no other math experience since then, has taken it upon himself to write, begin, or edit, hundreds of articles on math, including the ones i object to as inexpert on zariski and his topology.

fortunately this information is of public record, but makes it quite curious anyone would take these articles as seriously authoritative. he may not be a wiki founder but he does fit exactly the model i conjectured of one person, assuming the job of trying to learn and write up large numbers of articles.

this guy is apparently not an expert in virtually any of these topics.

 

[Chris Hillman]

Ferreting out authorship information

How do you tell where the student is from? Becasue he mentions Joe Harris by first name?

In this case, the student in question stated these facts at his Wikipedia user page http://en.wikipedia.org/w/index.php?title=User:Ryan_Reich&oldid=127458158 and Google's first hit on the obvious search was www.math.harvard.edu/~ryanr/ which makes it likely, in my experience, that User:Ryan_Reich is indeed the Harvard Math grad student he says he is. As I pointed out, one should never assume that WP users are telling anything remotely close to the truth about themselves on their user pages without making some attempt to verify the information provided there.

More generally, in my experience, honest WP users generally provide some factual information about themselves, and avoid registering sockpuppet accounts (unfortunately, many Wikipedians believe sockpuppetry is legitimate for various purposes, but again these rationales arise from conflating the "wikispeech.org" function with the "wikipedia.org" function). But as is the case at PF, I know of worker bees valued in the community who jealously guard their anonymity for reasons which they are generally reluctant to share. As at PF, WP users with few contributions who appear to know WP well are almost certainly sockpuppets, but I don't think we want to get into more sophisticated approaches to authorship identification (although the mathematics is quite interesting).

Everyone should be aware that authorship identification is an extremely inflammatory issue in the WP community (which overlaps with the PF community)
because so many use this website as an "anonblog"; see http://www.aclu.org/privacy/anon/index.html
http://www.cyberslapp.org/about/page.cfm?PageID=7
for some legitimate concerns about the growing suppression of free speech in the U.S. (where it is nominally protected, but increasingly, in practice discouraged because the personal consequences for speaking out can be so damaging).

But I stress that while these concerns are legitimate, they arise from the "anonblog" function of WP, which as I said above I have long felt should be split off into "wikispeech.org", and they directly conflict with equally legitimate concerns arising from the scholarly nature of an encyclopedia project. See http://www.aclu.org/privacy/science/index.html for some scholarly concerns arising from contemporary assaults on privacy, incidently (I know of at least one case in which a cranky company tried to force an editor to divulge the names of its referees, with the stated purpose of suing them--- see again the cyberSLAPP link.)

As I said above, a huge benefit of splitting up wikipedia.org and wikispeech.org would be that the consistent and declared author e-identity desirable in any project seeking to produce a reliable information resource (because professional identity coupled with personal accountability is such an important part of the scholarly discourse) would become much less controversial.

the article is not so bad, once you know the source of it, its just obvious it was not written by say a fields medalist.

OK, let's not beat up on a second year grad student :-/ I don't know this particular student but my default assumption about Harvard Math grad students is that they can learn to write :-/

Is it typical that grad students write these articles?, as they usually do not write books.

That's a good question. My own sense is that many of the active editors of math-related articles over the years have been math grad students. However, an amazing proportion of the work is due to a handful of extremely dedicated Ph.D. mathematicians who somehow find the time to produce a vast output, such as Charles Matthews http://en.wikipedia.org/wiki/User_talk:Charles_Matthews/About_me
Indeed, CM contributed to an earlier version of the article you read, as did Axel Boldt
http://en.wikipedia.org/wiki/User:AxelBoldt
and my former colleage Walt Pohl (who is now studying economics or something in Dallas, but who remains interested in pure math)
http://www.arsmathematica.net/
WikiProject Mathematics asks its members to provide some information about their academic credentials:
http://en.wikipedia.org/wiki/Wikipedia:WikiProject_Mathematics/ParticipantsI
I don't know anyone who suspects that any members have told any lies on that page, but no-one tries to systematically check. To be sure, neither does the Mathematics Geneology Project http://www.genealogy.ams.org/

There are many "edit counters" available, some which run at WP itself and some which run at external sites, and some of the more sophisticated ones make some attempt to provide quick visual clues showing how much of the text of a given version was written by a particular editor. (As I said, WP is a treasure trove of data on all kinds of things, including how a single author writes, since many articles are only edited by a single user for a long period. This data can be fed back into author ID tools, which can in principle use almost any characteristics of user behavior to unmask socks, etc., to track an individual across the web.)

ok a brief browse on wiki reveals that apparently one guy, named charles matthews, with a phd on gauss sums 10 or more years ago, and no other math experience since then, has taken it upon himself to write, begin, or edit, hundreds of articles on math, including the ones i object to as inexpert on zariski and his topology.

fortunately this information is of public record, but makes it quite curious anyone would take these articles as seriously authoritative. he may not be a wiki founder but he does fit exactly the model i conjectured of one person, assuming the job of trying to learn and write up large numbers of articles.

this guy is apparently not an expert in virtually any of these topics.

Well, no, he never said he is an expert on all the stuff he has written about "in the scholarly sense". Other Wikipedians would characterize him as an expert on those subjects "in the WP sense". Everyone who consults WP, or teaches students who might consult WP, should be aware that Wikipedians use the term "expert" to mean "someone who has studied standard textbooks" (and hopefully mastered the material therein), whereas scholars use that term to mean someone who has written a standard textbook, a much stronger criterion. More generally--- sorry to be the bearer of bad news--- WP has become so important that it is incumbent upon educators at all levels to familiarize themselves with the specialized vocabulary used at WP, not to mention extensive wikiskills.

Anyway, about CM: he has clearly continued to (try to?) learn after earning a Ph.D., which I think is laudable.

Hmm... actually, aren't you going to the opposite extreme now regarding math-related articles at Wikipedia, or at least Wikipedians? Running a Cauchy sequence in reverse?

I think you are being a bit harsh on CM; if nothing else you certainly have to grant his volunteer service at WP has been an impressive labor of love. I confess I tend to think many of his articles suffer from failure to implement the Halmos/Baez model (one paragraph for the janitor, followed by...), but given the vast scope of his output, I expect that many members of WikiProject Mathematics would counter: "let's see you try to write about a similar array of topics!"
http://en.wikipedia.org/wiki/User:Charles_Matthews/New_-_mathematics
It might help you assess his contribs to use the Way Back machine to sample Wikipedia math pages circa 2002---when CM joined the project, WP was a pretty poor thing. Quite a few Wikipedians think CM has been one of the key figures in the astounding growth of WP, since his efforts appear to have played a role in recruiting hundreds more math "experts" (in WP sense) over the years, with the result that many mathematicians find the coverage of contemporary math in the WP quite impressive. (As you know, I am less sanguine, and I think you are coming to agree with my doubts.)

Many members of WikiProject Mathematics are also adherents of the "open information movement", which believes that scholarly books and scientific research papers (maybe even scientific data) should be freely available to anyone. I tend to feel that way myself, although it is clear that the success of this movement (which appears not unlikely) will create new social problems; for example, just imagine a legal requirement that NASA provide a "professional conspiracy theorist" like Richard Hoagland http://skepdic.com/faceonmars.html with all their Martian data. Bamboozlers will benefit, initially, as much as serious students. The perversion of Wikipedia by guerilla marketeers, wikishills, cranks, political "dirty tricks" operatives, and other individuals who manipulate information presented at WP in order to pursue some hidden agenda, also illustrates that seemingly beneficial innovations are inevitably quickly exploited by persons seeking personal gain at the expense of the wider community.

 

[mathwonk]

i am indeed impressed that CM has been able to learn and write about so many topics in math. i myself could never do that. on the other hand i prefer to confine myself mostly to commenting on things i actually think i understand, not just things i can parrot the definitions of from a text.

I am just beginning to learn how low the standards are for a WP author. If anyone with a phd in an unrelated math field is considered particularly well qualified to comment on hundreds of other math topics, then wikipedia is, as you have suggested, nowhere i want to go for information, at least not about math.

i do not mean to beat up on these people for what they are contributing, but i do want to make it clear i agree that no one should consider these contributions anywhere near the quality of the content of essentially any standard textbook on the subject.

on the topic of zariski topologies, we all have the option of reading mumford or hartshorne, or indeed zariski himself, and artin or grothendieck on etale topologies, so what does a graduate student have to offer of value on this subject?

in the case of CM, he has apparently not at all understood what he has read about Zariski, his topology and the theory if algebraic surfaces, and some[possibly his] contributions to that biographical article are not only useless but partially false.

it might be of help if authors used the time honored practice by scholars of describing their intended audience. e.g. my notes on the RRT, begin with the statement that i am novice, i wrote the notes for myself and for anyone else beginning its study who may find them useful.

the default assumption about an "encyclopedia" author is rather more ambitious.

 

[mathwonk]

the article on riemann roch theorems also apparently by matthews was much better, reflecting his phd background. it was still in my view what one could glean from books, rather than original sources such as riemanns and rochs on works. i have read those, but still do not consider myself expert on the topic. Fulton is an expert on RRT, and Hirzebruch, and Grothendieck. And all these people have written books on it too.

 

[Chris Hillman]

For benefit of those who tend to use Wikipedia uncritically, let's contrast these "before" and "after" statements:

I am in math, and I have often been able to learn something apparently correct in my field from wikipedia.

I am just beginning to learn how low the standards are for a WP author. If anyone with a phd in an unrelated math field is considered particularly well qualified to comment on hundreds of other math topics, then wikipedia is, as you have suggested, nowhere i want to go for information, at least not about math.

I'm beginning to suspect this back and forth is mostly a rhetorical device, mathwonk, but assuming you really did revise your opinion from (if I might paraphrase) "the average math-related WP article appears reliable and even a professor of mathematics such as myself can learn from articles in my field of research" to "WP is nowhere i want to go for math-related information", this would provide a striking illustration of my contention that the biggest danger posed by Wikipedia (other than the Wikipedia-Google feedback loop) is that almost anyone can be seduced into using "information" found at WP without adequate fact checking.

And who the bleep wants an encyclopedia which needs to be fact checked every time you consult it?!

Really, it's utterly absurd, but everyone is doing it anyway. Why? Because it is so convenient to ignore one's scholarly instincts (if one has them--- most readers do not). So I ask again: is Wikipedia really a Good Thing for the World? I argue that it is not.

Now, even someone as smart as John Baez currently disagrees with me about the "average utility/reliability" of math-related articles in the WP. I feel that this reflects the fact that he has not been anywhere near as active in the Wikipedia community as was I (from 2005-2006). So I wish to stress the point is that as mathwonk has learned more about "who writes Wikipedia", his opinion of its reliability has (if one takes his statements at face value) dramatically decreased.

I am extrapolating quite a bit here, but I would like to suggest that this is consistent with my feeling that those mathematicians who praise WP as a valuable innovation simply haven't enough knowledge of how WP articles are actually created, and by whom. Or sufficient appreciation of the way in which information can be, has been, is being, and increasingly will be subverted in the Wikipedia by individuals hiding behind sockpuppet accounts or anon IP addresses who are pursuing some hidden agenda, to the detriment of the readership.

on the topic of zariski topologies, we all have the option of reading mumford or hartshorne, or indeed zariski himself, and artin or grothendieck on etale topologies, so what does a graduate student have to offer of value on this subject?

Getting back to my idea that one way to write a better Wikipedia might be to organize graduate departments around the world to give their second year grad students editor accounts and have them write (signed) articles on a topic in their area of intended specialization for credit. Part of that idea is that while an encyclopedia written by students may be of limited value for readers, it could be of great value for the writers!

Nonetheless, my own credo at WP was that writing for WP is a volunteer service, not a right, and that all questions regarding WP policy should ultimately reduce to answering the question: which choice best serves the interest of the readers of the Wikipedia?

In my deleted user space essays at WP, I sketched a "ring model" for a better Wikipedia. Think of a Venn diagram. Initially users with minimal credentials (pass the written exams at an accredited university, say) are granted an account with minimal privileges enabling them to perform limiting editing of particular articles on particular subjects. As they demonstrate increasing expertise (breadth and or depth of knowledge) and as they hone their writing skills, they can be granted privileges enabling more extensive editing, and/or editing privileges in a larger "ring". In my model, articles would be signed by the principle authors, and the history pages would give a clearer view of how much text each author contributed.

Wikipedians will be quick to object that one of the features which turned many novice WP editors into fully fledged Wikipedians is the ability to contribute to articles on all topics of interest to them, regardless of their actual knowledge or expertise. Here too we see the essential conflict between the "anonblogging" and encyclopedic functions of Wikipedia as it currently exists. In addition, in my experience many oldtimer Wikipedians are still stuck in the ancient days when no-one (including myself) believed that volunteer labor could build "something often called an encyclopedia" (thus Sanger) with two million articles, and are fixated on the issue of recruiting volunteers. I feel that we are living in a changed world; the example of Wikipedia, however inadequate as an information resource it might be, has acquainted a sizable fraction of the world's population with the idea that volunteer labor for a project like WP can be fun. And of course I am suggesting here a model in which significant labor would be provided by individuals "rewarded" with modest academic credit, and for the real experts on the faculty, hopefully, some kind of credit for professional service within their professional organization (in this case, the AMS, and maybe their department).

Administering a ring model wikiproject is a bit like administering SElinux--- it would be a lot of work, so every effort would be needed to make this burden bearable by the "executive staffers".

Let's not forget that I have also suggested that head-on competition with Wikipedia is hopeless; rather, visionaries should focus on creating specialized "reliable microwikipedias", perhaps employing something like a ring model. My idea is that such "locally intelligent wikis" (world readable but with sharply defined write privileges) would be readily autoaggregated into something much bigger, i.e. I propose to split up the linking functions. Innovations like Greasemonkey http://en.wikipedia.org/wiki/Greasemonkey suggest how linking from a closed "algebraic curve wiki" (for example) to a "mathbio wiki" in an article where Cramer is mentioned could be implemented. The effect would be similar to surfing Wikipedia, but the leadership of the individual wikis would be academics whose reputations would be on the line in terms of running them well in order to ensure maximal accuracy and minimal bias. (Indeed, there are already some scripts running at Wikipedia which autocreate internal links, which demonstrates my contention that aggregation can be done by "dumb" software; content creation within a subject requires intelligence and indeed wisdom.)

Wikipedians may object that one of the much touted advantages of "open wikis" is that they can be quickly updated, virtually in real time, to stay current . I'd counter that within the world of algebraic curves, only experts are really in a position to judge what constitutes A Significant Development. Give them the software tools and the profesional credit, and the natural desire to raise the profile of their own field within their academic discipline will do the rest.

Some might object "but then someone like mathwonk, local god-King of 'algebraic curve wiki' would be tempted to overemphasize his own interests, the contributions of his Ph.D. students, etc."; I'd reply that this danger already exists in current institutions such as hiring committees, journal editorships, and so on. And we've tolerated it fairly well for quite some time.

it might be of help if authors used the time honored practice by scholars of describing their intended audience. e.g. my notes on the RRT, begin with the statement that i am novice, i wrote the notes for myself and for anyone else beginning its study who may find them useful.

One of the reasons I gave up on my last-ditch attempt to reform WP from within was that WP has no facility--- even though this would be technically easy to implement--- allowing WP users to write signed and write-protected essays for policy purposes. Indeed, while Wikipedia users have their own directories (see the "prefix search" function under "special pages") , unlike what unix users familiar with large computer systems in academia would expect, these are world writeable, which leads to considerable chaos. So vandals can and do trash essays expressing views they dislike, which is just one more absurd barrier to even proposing major policy changes, much less implementing them.

"Authors": one of the more disastrous malapropisms which has become entrenched in Wikipedia culture is the use of "editor" to denote a role similar to the role of an "author" in the Brittanica model. This seemingly small mistake is, I suspect, partially responsible for many of the problems discussed by McHenry; a linguistic barrier to recognizing the distinction between the encyclopedic roles of author and editor fatally obscured the importance of providing "technosociopolitical mechanisms" to promote orderly structure at all scales.

Returning to the issue of how my ideas differ from the original "open wiki" model for building an information resource, I think the fundamental flaw with the "open wiki" model is the notion that checks and balances are unneccessary, because (according to the wikifaith) a wiki will be automatically attracted to a state of utopian perfection. I argue on the other hand that checks and balances, while a huge pain for everyone, are neccessary, and that software tools and thoughtful design of appropriate "wikiconstitutions" can lessen the pain to the point where the construction of "flawed but viable" microwikipedias are feasible. These can then be autoaggregated with little fuss.

Incidently, I been eyeing Puppy Linux for some time. In the short term, I think the MediaWiki software platform is valuable as a personal website authorship tool, and I envision a live CD tailored to the purpose of setting up a "closed wiki" for the purpose of conveniently creating world readable but unwriteable web content. Here, the math formatting abilities of MediaWiki are highly attractive to prospective authors of webpages with mathematical content.

 

[mathwonk]

i must distinguish two types of statements in my communiques" 1) factual ones, 2) attitudes.

the statement that i have actually learned something at WP is a fact.

the statement that i am disinclined now to look for expert info there is an attitude.

my attitudes vary depending on my mood, the latter having been expressed during a slightly cranky and indignant one.

still both are sincere, if naive. So I am not impressed with the credentials of the WP authors, but do confess to having learned something from them at times.

for example, even though charles matthews is probably not an authority on the atiyah hirzebruch spectral sequence, still he probably knows more about it than me. but i would probably not consult him to learn about it anyway.

i have mostly the luxury of being able to distinguish the false from the true information there in my own field, and am now more suspicious of the information I find there in other fields.

But many people have told me to be suspicious of it so I am not easily fooled by false information, being naturally sceptical. I also am an old man, and have noticed the need for scepticism in all fields over my life. Moreover I have kind of a false information detection device built into my brain after all these years.

Some articles on the front of the NY Times are also false.

 

[Chris Hillman]

Thanks for the clarification, mathwonk!

I also am an old man, and have noticed the need for scepticism in all fields over my life. Moreover I have kind of a false information detection device built into my brain after all these years.

Some articles on the front of the NY Times are also false.

True, true. As you have probably already guessed, my attempts to shout from the rooftops "Wikipedia reader beware!" is only a continuation of previous efforts to try to promote the education of students to think critically in all endeavors, including reading the newspaper. Part of what I am trying to get at here is that while many high school and college teachers are familiar with biases and so on which can creep into newpaper reporting, or with the implications of something like the Jayson Blair scandal http://en.wikipedia.org/wiki/Jayson_Blair these same educators are so far only very rarely aware of the vastly more complicated problems posed by Wikipedia, which has doubled or trebled the things they need to warn their students about. The fact that, as recent surveys confirm http://en.wikipedia.org/wiki/Wikipedia:Wikipedia_Signpost/2007-04-30/Statistical_profiles a large percentage of college students are already using the Wikipedia on a daily basis, lends added urgency to this mission.

(Before anyone complains that I am linking to Wikipedia articles while complaining that everyone is using this "thing that is often called an encyclopedia" as if it were the real McCoy, well sure, I am being ironical. My point is that if I didn't link, as said surveys suggest, anyone who missed the Jayson Blair scandal would head right to WP anyway to find out what I am talking about! And that's a problem.)

 

[chroot]

I'd just like to mention that Wikipedia currently does not claim to be an sort of authority on anything. It has roughly the same number of errors per article as other professionally-written encyclopedias, yet is many times larger and easier to use. Despite the fact that, today, it is not a substitute for fact-checked textbooks, it is a work in progress, far, far from its conclusion. It is possible that if Wikipedia continues to evolve for another hundred years, it may well become essentially error-free, aside from random acts of vandalism, and it may well render many textbooks obsolete and change the fundamental model of the educational system in a beneficial way.

In short, I think it's an incredibly valuable (and interesting) project, the likes of which the world has never seen before. I'm wholly in support of it, yet caution anyone who uses it to check it against other sources. This is, however, general advice given to every student after the fifth grade: no single source, no matter the author, can ever be considered absolutely inerrant. Taken in that context, Wikipedia has accomplished something truly remarkable in only its first few years of existence. Taken in its appropriate nascent context, it is a stunning achievement already.

- Warren