(this is the archive of discussion on MurrayAltheim's page 2004 Q2-Q4)


I've spawned a PlatonicCategories page to begin a new discussion on wiki use of categories.

Here's a somewhat related article on the Semantic Web, blogs, natural language, the power of linking, etc.: More Noodling on the Semantic Web, by Earle Mardle.

(See response: CanTheSemanticWebWork)

LionKimbro and others here are working on some cool things that head off into the ozone, like OneBigSoup… something to keep an eye on.

The Wiki MIME Type

Where there's been talk of namespaces in terms of wiki content, I'm particularly interested in trying to denote wiki text markup as distinct namespaces. People get to create their own namespace, identifying the specific wiki language they're using.

The idea would not be an attempt to harmonize the hundreds of wiki text variants, it would be simply to devise a sniffable first line that could be added to wiki pages to allow them to be discerned from text/plain. People could decide to play or not to play, but having a universal identifier for wiki texts would enable computer systems to treat them as processable files rather than plaintext. And we'd not rely on some file extension scheme, which is fragile.

Borrowing an idea from XML Namespaces, this first line would include a key token followed by a WykiWord describing the wiki syntax variant followed by a URL as an identifier for the syntax. The WikiW serves as a shorthand or proxy for the URL (in the same way as an XML namespace prefix serves as a proxy for the URI, for those familiar with XML). For example:

    #!wiki CeryleWyki

would be the first line of wiki text files using my Ceryle Wyki language. The MIME header might look like:

   text/wiki; variant="CeryleWyki"; charset="UTF-8"

The MIME type registration would do little more than describe this first line and how the variant parameter is a local name for the URL. For more information, see WikiMime, WikiMimeComments, WikiTextMimeType, WikiMarkupStandard, WikiInterchangeFormat, WikiPageInterchange, and InterWiki.

Wiki Syntax Registry

This is nothing official, nor do I think it necessarily even reasonable to try this, but here goes anyway: I've set up a page where people can add their wiki syntaxes as part of an informal registry. More information is on the WikiSyntaxRegistry page.

InterWiki Markup Language

[from an announcement posted 2004-04-12 to the InterWiki-discuss mailing list:]

I've posted an initial distribution of IWML 1.0, an XHTML-based interchange syntax for wiki content. The link can be found just following the table of contents on the IWML page. The tar.gz file contains the normalized DTD, a catalog file and a README file. This is a conservative first draft that represents something fairly similar to XHTML 1.0 Basic, plus additional xml:space attributes on block container elements, and <meta> anywhere an inline is allowed, as part of the Augmented Metadata proposal. This version includes Basic Forms, but I'm thinking of removing the module (it's an easy switch on/off in the DTD). Forms aren't really part of interchanged content (I don't think). Tables are not included. More information can be found on the IWML page.

Feedback, comments, etc. of course very welcome.

Argh! I will have to modify Oddmuse's output format from HTML 4.01 Transitional to XHTML eventually. Actually, the Oddmuse output format should be as close to IWML as possible.

Alex, I don't want to give the impression that I consider IWML a finished project. It's just a first draft, and I will highly value implementation experience in what it lacks, or change that need to be there. As I said in the announcement, it's a fairly conservative schema right now, close to XHTML Basic. It is modular, so we can also add modules for specific reasons, creating sub-languages for any needs that arise. Often in industry, there'll be a number of related DTDs for various stages during a document life-cycle (loose for authoring, very strict for editing, addition features following post-processing, etc.). I envision perhaps two or three IWML derivatives for various wiki purposes, but I'm keen to hear how things shake out in actual use to see what these use cases are.


i've just started to look at the semantic web stuff, and i have one main question right off the bat: where are the linguists involved with the project? does anyone have links to linguistic analysis and viewpoints?

There aren't any linguists involved, to my knowledge. The major involvement is with a small group of Description Logics researchers that has of course grown with all the money being poured in. Any connections between the Semantic Web and linguistics will likely be to existing linguistics research that predate the SW. If you're interested in the subject, I'm not much of an expert, but you could try starting at John Sowa's page on Concepts in the Lexicon. There's a lot of "classical" stuff like Minsky, but I'm not up on the state of the art. It's also an enormous field, so it's kind hard to know what you're looking for.

hi murray. you are on the wiki tree (WikiFolks | WikiDe:BaumListing?). i hope, this is the solution of the problem between lion & you. he too is on the wiki tree and so am i … and … --sigi

Sigi, there is no problem between Lion and I, we're just having a healthy discussion. We get along fine. It's okay to disagree, and we're talking things through. We might not agree in the end, but there's nothing personal, AFAIK. See TheSemanticWebComedy if you're unsure. But thanks for adding us to the wiki tree!

Btw, I thought the XCL web page design was neat. How do you do that (I guess I'm kind of behind the times). Is there a specific program you used to put it together? Did you use CSS? Thanks

No magic at all. It's a hand edited XHTML document that uses a fairly simple CSS stylesheet, and was put together using vi (well, vim). So if anything, I'm behind the times… but thanks.

Even more amazing!

Well, I've never created either XHTML or CSS, by hand or no, so I think it's me :)

No, I am the sister of Don Fernando. ;)

btw. "The Green" is great. It's a relief to return here from many other wikis. It's a pleasant light forrest … The pics came first, but "The Green"'s great as that. Would still change on the old logo, but anyhow …

Btw, any chance that ClaiMaker? will be released as open source? I think that AugmentedArgumentation? will really take off someday, but only in the context of open standards, available open source components, etc. (see also TypedThreadedDiscussion).

Bayle, sorry to take so long to get back to you; I've been out of town. When I last talked with Simon Buckingham Shum (one of my advisors, and probably the one person most in "charge" of the project now), he seemed very interested in going open source with it, but there are issues with Verizon, who donated the code to KMi with specific provisos on its further distribution. For example, it may go open source, but only be distributed by KMi. If these kinds of things can be worked out, it might show up as open source code, but my guess is that it'll not be as open as people think of open source unless Verizon changes its restrictions. I certainly agree that any TypedThreadedDiscussion needs to be based entirely on open standards, and I think that while ClaiMaker? is an interesting development, it probably has a fair way to go before becoming completely open and based on completely open technologies. It would probably take a major reengineering based on a currently nonexistent representation standard to make this happen, but it's a great and necessary step in that direction.

Ack. Bayle, I got confused. You asked about ClaiMaker? and I was thinking Compendium. Both are projects here at KMi, and related by way of some of the same people being involved and being very highly related. ClaiMaker? is part of the ScholOnto project, and I'm guessing the part you're most interested in is the argumentation modeling aspects of ClaiMaker?. I've heard nothing about ClaiMaker? itself going public, but Al Selvin has just announced that the Compendium source code is now available publicly (see tools-yak for more info). They're not the same thing, but you might want to check out Compendium since it's in the same realm of hypermedia discourse, and KMi has just spent the past year or so improving the source code, with the addition of new features designed to better support computer-mediated argumentation.


Define external redirect: BaumListing AugmentedArgumentation ClaiMaker