ansible at xnet.com
Sat Nov 18 17:17:20 CST 2006
Mark Miller wrote:
> On 11/17/06, Martin Scheffler <wooyay at web.de> wrote:
>> Now when I find out something new it would be great if there was a place to dump that
>> knowledge for future newbies. I know there is a wiki page (linked on the erights.org
>> homepage), but that does not seem very useful.
>> What about setting up a new wiki? Or should I extend the c2.com wiki?
> Thanks for raising this. We could use a wiki for many good purposes.
> At first, I was enthused about the c2 wiki
> <http://c2.com/cgi/wiki?EeLanguage>, until I got this complaint:
> [But also see WalledGardens. E language is very cool, but keep the
> E stuff on
> this page, or start an E wiki elsewhere]
> I have assumed without checking that this statement represents the
> consensus etiquette at c2. Frankly, I don't understand this objection,
> but I don't want to post pages where they're not welcome.
I've never been impressed with C2. They are using ancient, crufty wiki
software. And worse yet, the community standards weren't very strong at
the start. As a result, you have people in the habit of scribbling
whatever they think, anywhere, without much consideration of the overall
system. It is the broken windows theory, all over again.
For an example of an well-run wiki for a programming language, visit:
I've probably talked about this before... At any rate, here are my
recommendations on running a wiki:
1. Good software. MediaWiki (used by Wikipedia), TWiki, or something
comparable. Must haves:
Sensible article title support (none of the EeLanguage crap).
History of edits. Easily revert changes.
Ability to easily ban users and IP addresses (for persistent
Talk pages (optional but strongly recommended).
2. Good starting point. Setting up a blank wiki isn't very effective.
It is necessary to 'seed' the wiki with a decent chunk of good content.
You have this already; large chunks of erights.org can be wikified
> Brian Marick offered to host an E wiki rooted in
> <http://www.testing.com/cgi-bin/e-lang.pl>. However, as you can see
> from that page, it got spammed. I don't understand how it is that some
> high profile wikis, like wikipedia, manage to survive this threat. We
> could use CapWikis, like Waterken's, Sandro's
> <https://www2897.ssldomain.com/higherlogics/www/Wiki.ashx/About>, or
> Wideword <https://wideword.net/> as an access controlled wiki.
> Some of the things I'd like see develop on some kind of wiki-like medium are
> * A "security patterns shootout", where various small security
> problems are proposed by advocates of various different paradigms, to
> be taken as challenge problems by the advocates of other paradigms. We
> have a draft of some text towards such a shootout using Waterken's
> CapWiki at
> However, for this one purpose, it's probably a strategic mistake to
> host this on a CapWiki, as it looks like we're unfairly biasing the
> playing field towards our own paradigm. Other existing possibilities
I think it would be fine to use CapWiki... though I haven't reviewed the
I could potentially host a wiki, but it would be at the far end of a
puny ADSL line.
> * Existing explanatory material, especially, as you suggest:
>> Maybe the walnut could be imported as a starting point, split into a number of chapters
> Also, much of erights.org, my thesis converted to html, several of our
> other papers.
> * I would like to start developing an informal but precise and careful
> spec on Kernel-E, starting with the non-reflective subset of the
> Kernel-E special forms. As a spec, I am not comfortable allowing the
> open editing characteristic of wikis. I think a CapWiki would be good
> for this.
This is all good stuff to work on.
> OTOH, the erights.org site is already checking into Subversion
> <svn://svn.synchrona.org/erights/doc>, so people can check it out and
> submit changes to be approved and posted. I'm unclear on the tradeoffs
> between wikis and simply editing a version-controlled website.
Much, much lower barrier to entry with a wiki. Requiring a SW download
filters out 99% of the people that might want to contribute.
There are some papers that discuss how and why wikis work. I can
provide pointers if needed.
More information about the e-lang