John wrote:
The very thought of writing another makes me feel tired, not energetic.
It’s such a solitary pursuit, and the payoff comes only after so much labor.
The whole idea of books is looking a bit old-fashioned. A wiki, on the
other hand, is a communal enterprise. And, it’s fun from the very start!
This exactly captures my own feelings when I contemplate creative activity,
but there are many other reasons to prefer a wiki, some of which I obliquely
hint at below.
Generally speaking, we should expect the new site to develop in ways we have
not anticipated, so we should try to ensure at the outset that we are
prepared to deal with that gracefully. For this reason I advocated spending
some time discussing “wikigovernance”. I will suggest a very simple way to
minimize but not eliminate the time spent on developing a flexible but clear
“wikiconstitution” and also avoiding endless headaches thereafter.
I am very glad that both Blake and Garret have offered to help set up the
new website, and I am heartened to see that most of you seem to recognize
that this offers a wonderful opportunity to reconsider the reshape the cafe
into something completely different and hopefully far superior. I think it
is reasonable to expect the new site should immediately foster goals
including but limited to helping John develop his long-awaited next book.
But I am concerned that several posters are heading down the same path which
caused trouble with the current cafe.
I’d like to urge you all to think in terms of striking a wise balance
between conflicting desiderata, as you plan the new website.
My most important and least negotiable desideratum may unfortunately prove
most controversial:
I strongly believe, for a whole host of reasons, that edits to the new
website should be stringently restricted to a small and select “invisible
college” with new members occasionally elected by existing members, with the
original membership being the John, Urs, and David. This does not preclude
allowing other individuals from leaving comments or questions on certain
pages having especially lax permissions, provided that they are willing to
“register a comment account subject to our verifying your IRL identity”.
Nor does it preclude eventually constructing a more elaborate scheme of
permissions enabling a larger group to do more things at the website.
My (reluctant) insistence on this point will probably horrify many here, so
let’s just leave that hanging for a moment and immediately pass on to some
further examples of possibly conflicting desiderata:
Some things we want to ACHIEVE:
- easily learned markup (I dare say we all already know both MediaWiki and
LaTeX markup!)
-
easy and flexible inclusion of mathematical equations, tables, and figures
(all of which standard MediaWiki does superbly, IMHO, AFAIK).
-
easy and flexible internal and external links (compare pasting urls into a
wik and bracketing them with making a link in standard html)
- easy and flexible organization (MediaWiki categories are easy to learn and
use; I hope that future editions will enable graphic visualizations but we
could easily provide our own jpeg figure showing basic organizational
structure)
- site should be (mostly) world-readable but certainly not world-writeable
(I am told that MediaWiki allows for flexible global permissions)
- easy way for general public (but not IP anons) to leave comments in
“comment space” (I am told that MediaWiki allows for flexible permissions
applied page by page)
-
easy way for general public (but not IP anons) to “ask an expert” a
question on n-categories
- easy webform for applying to register an account, with clear rules for who
is allowed to register an account
-
easily found, clearly written, morally elevated but not unrealistically
idealistic (a) statement of purpose (b) privacy policy (c) wikiconstitution,
and (d) behavioral rules
-
easy editorial oversight, reasonably uniform and attractive
appearance/style of public articles
- easy maintainance
- foster cautious but responsible exploration of possible technical
innovations above and beyond standard MediaWiki (bearing in mind all
conflicting desiderata)
- stick close to the tried and true
Some things we want to AVOID:
-
allowing any anon edits whatsoever, in fact I would strongly prefer to
restrict registered accounts to entities whose true identity is known and
whose good judgement is well established; I feel so strongly about this that
I myself would be willing to forgo registering and account myself if others
would accept a rule preventing trolls or sockpuppets from messing with the
website, e.g. by stating that “verifiably identified persons only”, in fact
“tenured faculty only”, can request an account— my thinking is that it is
rare for people to attain professorship or even graduate studentship at
accredited institutions unless their academic credentials have been vetted,
or having their existence and unicity verified)
-
compromising security of the website itself (e.g. by trying to shoehorn in
some buggy code; anything which requires Javascript might be suspect)
-
any technical innovations requiring either a reader or contributor to
install something on his own computer, or to extensively change his browser
settings, or anything potentially compromising security of the user’s
computer, without a very good reason; indeed, the working assumption should
be that the default user has tight browser settings; ideally, anyone should
be able to read public pages with very tight browser settings, and if
someone who has registered a comment account and who uses Firefox to enable
“session only” cookies from the new website, should be able to comment in
public areas of the website; enabling javascript or popups or inherently
insecure code should not be neccessary for casual interaction with the new
website
-
anything at all tending to make it easier for trolls, vandals, or
well-intentioned but overenthusiastic ingoramuses to disrupt the website
-
anything at all which tends to create unneccessary technical obstacles for
either contributors or readers
-
chaotic mix of notation and style
-
time consuming maintainance, getting bogged down in technical issues
-
getting bogged down in developing experimental and possibly buggy code;
radical technical innovations should be avoided unless there is a strong
rationale for risking them,
- getting bogged down in “wikigovernance”
Let me expand on the issue of “wikigovernance”: IMO, the lamentable example
of Wikipedia shows very clearly that the single most important thing one must do to minimize the time spent on wrestling with wikigovernance is also
one of the easiest: forbid anon edits, period. That is, forbid any edits by
unregistered users, and require that applications to register an account
have verifiable identities (most easily achieved by requiring that they have
e-verifiable affiliations).
This is one reason why agreeing at the outset on a common and clearly
expressed vision at the outset is so important; I am not kidding when I use
the term “wikiconstitution”, a term intended to suggest a clear an
unambiguous framework for resolving unexpected future issues with a minimum
of fuss). I feel that initially restricting edits to a small and trusted
group would essentially eliminate any governance headaches, but it should be
easy to provide a world-writeable comment page (I’d advise stating that if
it is abused, the maintainers reserve the right to block anyone or even to
stop reading it if it becomes a timesink).
Another set of considerations to bear in mind: the new website would among
many other things comprise, would we will it or no, one more scholarly
response to the endemic and apparently incurable quality control problems at
Wikipedia. So let me add a few words about the prospect of competing with
Wikipedia, in some sense, for the attention of search engine spiders and
those seeking mathematical/scientific information.
I have long stated that IMO attempts (e.g. Citizendium) to compete
“head-to-head” with Wikipedia are unlikely to succeed. Rather, I have
advocated creating small highly specialized wikis, perhaps combining
exposition and research material, created and maintained by small groups of
genuine experts (typically, tenured faculty members) who know each other
personally and share a common vision. I have argued that the proper role of
massive “universal information resource” websites should be to aggregate
reliable information from such specialized encyclopedias.
At present, the best example of what I have in mind may be Scholarpedia.
Another model (also boasting Terry Tao as a contributor) which seems well worth examining is DispersiveWiki.
Note that the denizens of this wiki have taken advantage of wiki’s
flexibility to make annotated bibliographies, to write expository articles
which should help graduate students understand the importance of the subject
at hand, and so on. I think this website, while still very young, provides
a good hint of what I have in mind. (Unlike Scholarpedia, it seems that
DispersiveWiki follows the traditional wiki model of being both world
readable and world writeable, which I feel is a terrible mistake.)
So, one possible “spinoff” would be the creation of a small highly
specialized encyclopedia on categories, or even just on n-categories. This
could happen naturally via the creation of expository articles on topics
such as “gerbe”. Perhaps David Corfield could be persuaded to switch fields
and to create a small highly specialized encyclopedia on philospophy of
statistics. (No? Well, it was worth a shot!) Perhaps I could be persuaded
to try to resurrect my hoary old “Categorical Primer” in the form of a set
of expository wiki articles. Should these possibilities come to pass, it
would well worth taking the time to look into how one requests sites like
answers.com or wikipedia.org to auto-aggregate articles from the new site,
how one can get it listed at yahoo, and so on.
Coming back to the troublous question of who would be allowed to edit the
new website: unfortunately, limitations of time, energy, space, and a
certain desire for personal safety, all preclude me from attempting to even
sketch my concerns about manipulation of information except by setting up my
own “protected wiki” collecting links to on-line newspaper articles, case
studies using Wikiscanner, and so on. In lieu of such a sketch, I hope
everyone here will accept that as a highly experienced Wikipedian (at one
point I was in the top 500 most active contributors) who has also
participated in various other public forums for a long time, I am all too
familiar with the prolixous and sometimes subtle ways which nasties have
invented in order to promote a hidden political or financial agenda by
taking advantage of the scholarly tradition of Assuming Good Faith. Life
would be so much simpler if we could operate under that assumption! At the
risk of beating a deceased equine, I reiterate: by limiting editing of “best
articles” to a small group of known and proven-trustworthy contributors, the
new website probably COULD operate under that assumption, to a great extent,
at least initially.
One point I’d like to stress is that, if the maintainers are willing to
accept the risk of a timesink should this not prove straightfoward to
implement, it should be possible to eventually grant a range of privileges
to different registered users, and also to impose a range of restrictions on
individual pages at the new website. For example, I tend to feel that
“approved articles” should be world readable but only writeable by a very
select group, while internal discussions could be readable only by
registered members. My experience at other forums suggests that there will
be a need for private discussion of sensitive issues, but for a small and
highly focused website this could probably be handled by email among a core
group of maintainers who all have registered GPG keys.
I note that restrictions need not be enforced within MediaWiki but can and
probably should be reinforced by such simple steps as write-protecting
certain directories at the level of the operating system (it goes without
saying, in this crowd, that the operating system will be unix or linux). If
some of what I suggest above sounds like SE-linux, well, some way this is
much easier to use now than it once was, but there are alternatives which
are even easier. But as I stress again and again, simply enforcing settings
which forbid edits execept by registered users will go a very long way
toward protecting the new website from the timesink of reverting vandalism,
dealing with cranks and trolls, and so on (and on and on and on).
(If anyone here doubts that trouble invariably finds any well-regarded
website offering “good stuff”, ask just Blake about wikishilling PR flacks!
We all need to accept that, as experience shows very clearly, the more
highly regarded a website is, and the more respected its
information/judgements, the more likely it becomes that baddies will try
very hard and very persistently to infiltrate and subvert it.)
Making the new site appear friendly while insisting, gracefully but firmly,
on strict rules for who can do what, requires a rather delicate balancing
act. But you (we?) don’t really need to explain in detail why you (we?)
have strict rules, since it would be your (our?) site.
As some of you already know, I have not been able to participate in the Cafe
because the Cafe has violated several of the considerations outlined above.
I feel so strongly about the neccessity of imposing strict rules on who can
edit (at least, who can edit public pages the new site stands behind) that
I’d be willing to forgo my own participation if crafting clear and simple
“rules for participation” which would allow me but not trolls, vandals,
cranks, and otherwise disruptive users should prove difficult. This
wouldn’t preclude my sharing, as an exemplar, the style manual I created for
the defunct WikiProject GTR in Fall 2005.
I dare say that all of us, or almost all of us, are strong advocates of the
“Open Access” movement, and also of what has been called the “New
Enlightenment”. As such, we probably tend to strongly favor empowering the
ordinary citizen, and we may well be attracted by some elements of the
venerable tradition (in American politics) of populism. Nonetheless, at
Wikipedia (before my user pages were vandalized out of existence), I tried
to argue that scholarly ethics are unintentionally but ineluctably elitist
and thus in direct conflict with populist ethics.
But let’s not lose sight of the simple principle underlying all of my
suggestions above: knowledge is power. In the long run, the best way
academic experts can empower ordinary citizens is by ensuring that their
expert knowledge, advice, and judgement is made freely available to the
public, while also ensuring that their time is not wasted dealing with
trolls, vandals, kooks, wikishills, legal threats from malcontents, and
assorted other disruptive forces.
Lest any of these considerations should seem daunting or even discouraging,
I stress again that by far the simplest way to avoid headaches is to
restrict editing to a small and trusted group, at least initially. This bit
of wisdom runs directly counter to the original vision of the wiki, but I
trust the energy and enthusiasm of the core group will prove that wikis can
be usefully employed in a much less naive fashion to develop a truly
valuable public and scholarly resource.
Last but not least, John wrote “The Higher-Dimensional Wiki (or whatever we
call it)”. My own suggestion (and I’m surprised that no-one has beaten me
to it!) is “Weekipedia” :-)
jsMath
jsMath is great; Dave Cervone is a genius. It’s far-and-away the best option for embedding math in “ordinary” HTML pages. But I warn you (and Dave will assent) that it can be really slow on math-heavy pages.