These are quick notes taken by Sean B. Palmer on the Semantic Web, Python and Javascript programming, history and antiquarianism, linguistics and conlanging, typography, and other related matters. To receive these bits of dreck regularly, subscribe to the feed. To browse other months, check the contents. This file was generated from plain text source, for convenience of posting, so apologies for all the in-your-face URIs.
<d8uv> With careful consideration, Pepsi > Coke
On #swhack, mumbles asks for Phenny to be backported to Python 2.3... I say that will only be done as a consequence of $$$$$$ (because it's quite pointless and I don't want to fuel 2.3-usage; it's gone already! get over it people! 2.4 has been around for two years and a day!), but he says he'll just wait until I'm bored. That's an annoyingly good strategy.
Cf. http://swhack.com/logs/2006-12-01#T14-22-59
"Consider the whole thing to be preceded by a big "In my opinion" and all the
generalizations to be preceded by "In most cases". I could qualify and nuance
everything, but... in my opinion, and in most cases... that leads to
wishy-washy, unreadable prose."
- http://www.zompist.com/predic.htm
"the triumph of 19C (= "19th century") liberalism"
- ibid., zompist.com
I was taught C19, not 19C. I guess 19C makes more sense, but the former looks better. Oh, aesthetics vs. practicality...
Oh man, this is great stuff...
"As a linguist, though, I consider definitions to be artificial constructions;
most words are really generalizations from prototypes."
- ibid., zompist.com
"Chomsky suggests that there could be no word "similar to 'limb' except that it
designates the single object consisting of a dog's four legs". As a
counter-example, Sampson suggests French rouage, which can be used to refer to
(just) the wheels of a vehicle (p. 109); I would also suggest the baseball term
battery, which refers to the catcher and pitcher. (And could we really not
invent a term 'limb system' to refer to the dog's limbs?)"
- http://www.zompist.com/langorg.htm
Cf. ...ah crud, it's lost. My conversation with Kragen Sitaker over whether Polonius' advice in Hamlet was good advice or not, which kinda morphed into whether it was good advice at the time or not.
(manxome was running out of disc space, I guess loggy.py got in a blue funk over it).
Wiki idea: a page only exists as a set of dated versions, where they can split and branch and recombine as though they were in some revision control system, and then the admins choose a version to use as the main page.
http://en.wikipedia.org/wiki/William_Haughton
- Elizojacobean playwright, wrote with Chettle and Dekker
Can't exactly recall where he was mentioned, but it's likely that I came across him whilst reading Philip Henslowe's diary. The diary and the stationer's register are great reads (and online, too!).
n3s -> token traverser -> N3 editor -> browser, a la Tabulator ...
http://extra.shu.ac.uk/emls/iemls/shaksper/files/SONNETS%20TAG1609Q.txt
- Good example of a "public domain" thing where commercial use is forbidden.
Cf. http://www.library.utoronto.ca/utel/ret/shakespeare/1609.html
http://www.library.utoronto.ca/utel/ret/shakespeare/1609coco.html
Also, I really like i Harrey iv ij.i. None of my versions gloss it adequately
though; it's one of the better candidates for heavy glossing.
http://nfs.sparknotes.com/henry4pt1/page_51.epl <- nice design
I've been thinking about using Joanna:
http://en.wikipedia.org/wiki/Joanna_%28typeface%29
But the only thing that I could found typeset in it online, as a PDF, looked pretty gnarly. I should get my hands on a copy of Gill's Essay, which is also set in it--and has nice pilcrows strewn about the place.
10:47 <sbp> woah! newsflash on Slashdot! 10:47 <sbp> "twofish writes to tell us that Sun's Tim Bray (co-editor of XML and the XML namespace specifications) has posted a blog entry suggesting RELAX NG be used instead of the W3C XML Schema." 10:47 <sbp> - http://it.slashdot.org/it/06/12/05/0025208.shtml 10:48 <sbp> quick people! it looks like it's time to migrate! 10:48 <darobin> rofl
Bwahahaha. From #swig
I started thinking about a trivial RDF application and realised that I'm still interested in making a better RDF API. I was thinking about a current decent API for the sorts of things that I want to do. It involves using nested formulae and rules and such a lot.
Embedding Graph and Formulae instances into triples... is there any reason why Graph needs to exist at all? A Formula with empty universal (and maybe existential over different scopes) bindings is presumably equivalent.
cwm, rdflib, pyrple: a unified API accessing the best of these three would be very interesting, though rdflib now has C compiled parts.
APIs are rather close to programming languages at this level. It might be a thing to see if there could be some simple format for interfacing a language independent API. Writing bits of RDF inside a language should be easy; indeed, N3 is already a kind of language, but declarative. It'd be interesting to make it procedural (people have tried before, with not much success).
The use-case I had in mind was pretty printing an ontology. I think there's been a little work on RDF templating too.
Okay, you know what would be cool? A medium length page explaining how to do things with rdflib. There's an example bit of code on the rdflib site somewhere which is kind of useful, but having a kind of "examples reference" would be even cooler. I'm noting this down because I've been looking back over these notes from time to time as inspiration for stuff to do.
If you want to do it for me, of course, that would be great.
Last night I made http://inamidst.com/proj/sdoc/ -> Schemadoc, a tool to get XHTML from RDF Schemata. Today I'm thinking about Semantic Web link properties again:
It's important to link data on the Semantic Web, but how is this achieved? Since there are several ways to find new data, each of these ways can be used deliberately to link:
- By dereferencing the fragmentless racine of a URI. Find information about http://example.org/ont#Term at http://example.org/ont
- Follow well known generic link properties such as rdfs:seeAlso
- Follow application specific link properties (like what?)
So, which properties are well known for linking to new RDF data? That's one thing that a huge search engine for the Semantic Web might be able to solve descriptively: it could expose a list of properties whose objects tend to be HTTP URIs that 200 OK resolve to some application/rdf+xml or one of the other RDF MIME types.
What about application specific link properties? What do FOAF crawlers and Tabluator follow?
Link Ontology
What about finding new generic and application specific links? TimBL suggests triples like foaf:made link:listDocumentProperty foaf:pubs to say that foaf:pubs is a property whose range is a class of RDF documents that contain foaf:made as a property. This touches on GraphSchemata.
I'd be interested in making such a link ontology, since it has that link to the kind of graph schemata thing I was working on: http://infomesh.net/2003/graphsl
Datestamping files with their modification times, on OS X/Darwin:
for fn in * mv $fn $(stat -f '%Sm%N' -t '%Y-%m-%d.' $fn)
Using zsh
>>> sorted((v, k) for (k, v) in things.iteritems(), reverse=True)
SyntaxError: Generator expression must be parenthesized if not sole argument
This seems odd because, in obligatory caps: IF YOU CAN GIVE ME AN ERROR EXPLAINING THAT THEN YOU CAN BLOODY WELL PARSE IT CAN'T YOU?
Cf. http://swhack.com/logs/2006-12-10#T17-29-43
Just used the word "tatterdemalion" on #svg, reaching for a kind of word for trash or uninformed dreck. Ironically, I meant to invent a word and ended up pulling out a real one. I like it when that happens.
I've been reading bethisad.com and some other conlang stuff. When I was young, like many young intelligent people I made conlangs but grew out of it when I realised that to do the task properly is overwhelming. But that's thinking of conlanging and conculturing as a science rather than an art.
Alternative cultures give you a framework within which to work on interesting things. Alternative horology? Alternative caligraphy? On Ghyll we tried to keep away from transparent earth parody, though allowing very deep or complex earth parody, such as that of John Cowan's Wikipedia parodies that were hard to notice even *after* he'd pointed them out.
Alternative science is boring, and doesn't need any "alternative". That is to say, if it's alternative small-scale science, reinventing the car or whatever. Conjecturing alternative sciences on the large-scale is science fiction. This too is boring to me, though not to many. Perhaps I'm conflating the two. It seems that Tolkien got it right because he was good at both art (poetry, if you like), and science (philology).
Using literature for alternative art, however, is exceedingly difficult for the layperson even if it's possibly valid. Literature and poetry is the lessermost of the arts. Perhaps one should start with the haiku: Syldavian and Lupine being good examples of the conlang equivalents of haiku.
To really exercise alternative culture as an art, in literature, perhaps one should start with an alternative language, and only write in that language?
The Gaillimaufry of Whits is rather like an IRC channel... I paste in gobs of crap and I get a URI out of it that I can pass around to people. But though it has the benefit of being less cluttered, it has the disadvantage of being much less interactive.
http://www.dangermouse.net/esoteric/petrovich.html
- A (Bayesian?) DWIM OS
I've thought of doing something similar for editors...
http://swhack.com/logs/2006-12-10#T19-32-43
- Research on when Infobot was written (spoiler: 1995)
<alienbrain> Now I'm getting the feeling of that ".. historian" bit ;-)
http://www.zompist.com/dfc404.htm
- A peek at the slush pile
So... what caption got used?
http://www.zompist.com/dfcindex.html#400
-> 404 http://www.spinnoff.com/dfc/archive/404.html
-> http://web.archive.org/web/20010415190315/
http://www.spinnoff.com/dfc/archive/404.html
(Broken image URI)
-> http://www.google.com/search?q=%22DFC+Archive%22+404&oe=UTF-8
-> http://dfc.furr.org/archive/404.html
-> Tsk, they pick several. Fine.
New Subsite! http://inamidst.com/topic/
Ideas for future topics: * Excerpts of Drummonds Conversations with Ben Jonson * Hints and Tips on Creating a Language * Link Properties on the Semantic Web
Cod8uvy is thinking about making a Whits-a-like, based on Markdown instead of plain text. He's not thinking about making an Atom feed for it, but he will support Atom if I send him code for it, which I'll probably do.
It also got me thinking about HyperAtom. In other words, an HTML profile (a la Microformats and GRDDL) that can be transformed into Atom using XSLT or some other transformation mechanism. Aaron came up with such a profile for RSS:
http://www.aaronsw.com/2002/rssXP
And there is an Atom Microformat already, but it sucks so I'm not linking to that. Aaron links to Barger's post predeicting the rise of an RSS-a-like done in a similar way, which is really worth reading:
http://www.robotwisdom.com/web/parsing.html
Pretty amazing stuff for 1998. Anyway, I'd just like to be able to get some Atom blood from the XHTML stone. But then that shifts the trouble of creating Atom to the trouble of editing XHTML...
http://inamidst.com/topic/edithtml
And as you can see from the article above, I'm not making a great deal of progress on that. And that's why the Gallimaufry of Whits is plain text.
Jim Hendler just wrote about the Dark Side of the Semantic Web:
http://www.mindswap.org/blog/2006/12/13/the-dark-side-of-the-semantic-web/
Via Chime, on #swig just a moment ago: http://chatlogs.planetrdf.com/swig/2006-12-14.html#T15-22-18
Jim is talking about the practical side of the Semantic Web, which links very closely to the kind of rants that I've been doing mainly in semi-private on the subject:
I should really get cracking on writing up some of those things. I've been thinking that the first step is to rewrite my Introduction to the Semantic Web, http://infomesh.net/2001/swintro/ - it's really out of date now. One way of writing a decent article is to start it off as an email to a friend, so I have a draft of this addressed to a certain Mr. Javier Candeira in my drafts box, but I haven't got around to writing it yet.
I've also been shirking (ho) away from even outlining the basic problem areas of the Semantic Web. Perhaps that's something that I can work on here; indeed, I already am since what Jim's identified here is one of the things that I've been yammering on about to anyone that'll listen. The Semantic Web should be a cool remixing zone for people who like to hack about with data. There's a three tiered system:
Big companies, big data Enthusiasts and hackers, remixing the data Consumers, using the data
I'm in the middle, but current business models tend to completely exclude the middle zone. What the Semantic Web does is really to enable that middle tier, and until people understand that, the Semantic Web won't reach its full potential.
Sigh, now raxor just posted this: Validation Actively Harmful:
http://www.coactus.com/blog/2006/12/validation-considered-harmful/
It's by Mark Baker, of XHTML 1.1 Basic fame, and it's about
future-extensibility in languages.
The basic idea seems to be that validation as it's performed at the moment is way too rigid, and especially in that it doesn't account for future changes that might take place. Formats need to be extensible on the Web.
TimBL has written a little about this:
There is a common requirement for the design of a language on the web that it should allow for extensions, but it must allow a clear declaration as to whether understanding of an extension is a requirement to understanding of the document or whether it may be ignored.
Basically, this means you have two kinds of extension. MAN (mandatory) and OPT (optional). This is what the short-lived HTTP Extensions thing had. User Agents are to just ignore OPT things that they don't understand, but have to raise errors if they see a MAN thing that they don't understand.
I'm not sure how this would work in HTML extensions. Obviously migration to RELAX NG is one step, and the kind of MNG thing that Masayasu Ishikawa was working on is the next one, but this kind of stuff has to be really simple. Moreover, I'm starting to do weird things like writing HTML with XHTML style elements in it... I think that we need to be able to account for all of this in a not-too-tag-soup-ish way.
MarkB says that he'll write more in future but hints that the tags of the article give the answers. The two tags are "webarch" and "semanticweb". Hmm.
The homepage for Whits is a simple shell script, and I wanted some shellish way of formatting a number nicely, for the article wordcounts. I came up with this:
function format() {
echo -n $1 | rev | sed 's/.../&,/g' | rev
}
$ format 1234567890
1,234,567,890
Hooray for simple solutions!
Oh, you have to lstrip commas on three digit numbers...
function format() {
echo -n $1 | rev | sed 's/.../&,/g; s/,$//' | rev
}
Much better.
http://inamidst.com/topic/bugfree
- Bug Free Programming!
Being a new topic. I'm starting to wonder about having a feed detailing recent changes in the topics directory... I think it'd be possible to have a feed of some kind, but. Hmm. I've been wanting something that makes an Atom feed of any changes in any given directory on inamidst, or perhaps even to any given pattern. So if I wanted to subscribe to a feed of all changes for phenny modules with a vowel in them, I'd subscribe to
/changes/.../phenny/modules/*[aeiou]*
or whatever. I'm not sure how to make the URIs pretty, and perhaps they don't even need to be, particularly. Another idea would be to have something like the W3C comma-tools, so /topic/,feed would link to a feed for all of the items in that directory.
Oh, also, the whole point of Topics and Ideas is that it's kind of like a personal wiki. So it's not serial, like a weblog, and pages can be gone back to and revisited as much as is necessary. I wonder if it'll play out that way in practice. It's also meant to be content-oriented, with basically no style. It will cover technology, history, and other stuff, but it's just technology at the moment.
So in other words, it's a kind of replacement for /eph/, which I rather miss, only it doesn't let other people edit it, which is a great shame. The challenge of the technology required to do that would probably be too big... unless I can rig up some kind of diffbot on IRC. Cf. http://inamidst.com/topic/typodiff
The Flog: http://d8uv.org/flog/
So Cody Woodard decided to take Whits, add somewhat of Markdown, a soupcon more Shell, take away all the Python, add much as is good Style through CSS and JPG magics, and er... well the results look awesome.
I'd steal his improvements back, so it could become this huge cycle of stealing, but really all of the improvements circle around the use of Markdown instead of plain text. Perhaps it's time that I learned to use Markdown naturally, or worked on Avocet: http://inamidst.com/topic/avocet
I think I'll also try smushing the Whits text as it is now through Markdown just to see if it comes up with anything useful. It probably won't, in which case I'll have to puzzle over something else. I've actually been planning to write some massive descriptive parser to convert this all into HTML, but man... I dunno.
Anyway, good work d8uv!
The Markdowning of Whits did Not Go Well.
http://d8uv.org/flog/2006-12#t1166286479
- Cody mentions Whits mentioning The Flog on The Flog; I have now mentioned The
Flog mentioning Whits mentioning The Flog on the The Flog on Whits. He thinks
that this will end up being an infinite loop that destroys the world, but just
as Python simply goes "RecursionError" when it runs out of stack rather than
melting your computer, I know that Cody's brane will have now blown, saving the
world utterly.
http://d8uv.org/flog/2006-12#t1166346447
- Oh yeah, also this. Cody replies to my write up of his lamentation:
http://inamidst.com/topic/headings
So my reply is merely to turn this:
"ON WANGS" by Cody Woodard
- The different varieties of wang - The short and stocky - The Long and Lean - The massive, otherwise known as "The Cody" - A day in the life of the average wang - 10:18, Wake up - 10:20, *censored* - 1:11, *censored* - 6:23, *censored* - 6:32, Urination Duty - 11:23, *censored* - Dongs vs. Wangs - Dongs are more dongly - Wangs have the propensity to wang a lot
Into this:
"On Wangs" by Cody Woodard
- All about wangs
- The different varieties of wang
- The short and stocky
- The Long and Lean
- The massive, otherwise known as "The Cody"
- A day in the life of the average wang
- 10:18, Wake up
- 10:20, *censored*
- 1:11, *censored*
- 6:23, *censored*
- 6:32, Urination Duty
- 11:23, *censored*
- Dongs vs. Wangs
- Dongs are more dongly
- Wangs have the propensity to wang a lot
The main change is that there's a single top level h1 which commands the whole rest of the hierarchy. Why this is important is that it lets you, in practice, place say an introductory paragraph and a table of contents into the document whilst having a single structurally and visibly labelling the whole document.
I have to think about this some more though. Occasionally someone puts up a document that has no headings at all, and I tend to think they look awesome.
The best option is to have <section> as in XHTML 2.0, of course, so that the hierarchy is really explicit. That way you can do:
Level two
Level three
Return to same level two
Which you can't do in HTML at the moment. On the other hand, nesting is very hard to keep track of. Also, I don't really like:
<section> <h>Heading</h> ... </section>
Which is what XHTML 2.0 does at the moment, I believe. It might be nicer if the heading were an attribute, though that'd mean it couldn't be styled. I wonder how I did it in XNote... probably the element way. Oh well.
Ah, it was indeed: sec, and hd. Cf. http://infomesh.net/2001/09/XNote/
I just wrote an heuristic Whits source parser:
http://inamidst.com/stuff/whits/parser.py
- Experimental Whits Parser
It's... so complex and bizzare, some might say "baroque", that it's probably sentient, and if so, it's also likely to be slightly angry at the world. It uses heuristics because obviously I don't adhere to any syntax here; I just type whatever I want into a text window. But HTML is not that complex, and generally parsing is just a case of finding which of just a few categories of elements the input would best fit into.
So eventually I think I might hook this up so that:
In that way, when someone requests, e.g. /2006/12, it'll call the CGI if there's no 12.html, or return the baked 12.html if there is. If someone requests /2006/12.txt, the 12.txt source is returned.
But for that to happen, I'd really need to rewrite the heuristical (you have to use the secondary adjective because the primary one is so nominalised now) parser so that it doesn't slip up and produce crazy things. In other words, to make it more robust and easy to extend.
The robustenising of the Whits Parser has gone rather well, so I hope to transition possibly for New Year's Day 2007. I might also move the whole website to /whits/, putting in redirects so as not to break things.
I've also written feed.cgi for Cody, so The Flog should have an Atom feed rather soon! I'll link to it when he publishes it.
I'm wondering about writing a Topic on a basic overview of the kinds of things that I'm interested in. I'm listening to a lecture from Berkeley about the "18th Century Public Sphere: Early Newspaper, Coffee Houses, Political Discourse", and though it's not all that good, it's the kind of topic that I salivate over... the history of knowledge and all that.
I'm mainly working on language at the moment. I haven't taken any notes down (I should do!), but I'm studying basically everything that contributes to the makeup of a language, including:
I've found out some interesting things. For example, I worked out by myself that determiners were a kind of adjective, and then found that Wikipedia agreed with me, which I was rather pleased about. I also learned that "adverb" is a catch all term for various kinds of word, and that switching between Ergative and Nominative marking of verb arguments is called Austronesian alignment.
I'm currently thinking about parts of speech and prototypicality... it seems that words don't very closely cling to conventional descriptions of types of speech: much like phonemes, they sit on a spectrum. When we verb a noun, we're just shifting the domain that it exists in on the spectrum, but even in grammar there are reflections of whereabouts on the spectrum a word sits. For example, there are lots of tests for adjectives, as some adjectives only pass some, but the more prototypical ones pass more.
What I wonder is: when you invent an adjective form of a noun, say, does that help the mind to remember that the noun form sloops around the noun area of the part-of-speech spectrum? Perhaps it doesn't... most people would think of heirarchic as an adjective, but it's pretty nominalised. I don't think people would only use "grey" as a noun if there were a word "greyic", because grey is perhaps one of the better examples of a word existing in a domain that crosses two classical part-of-speech classes.
Lots of interesting stuff to consider.
Oh, I'm also regetting into art. It occured to me that I've been using computer science as my art for a long time now: I'm always trying to design new projects and write beautiful code. But that's kinda silly when there are other kinds of art like poetry, painting, music, sculpture, dance, and whatever. I'm by no means interested in all of those, but I've dabbled in one or two...
Javier is interested in computer art, that is art generated by computers according to some human generated rules, but I'm not so keen on that; but I wonder if there's some way of looking from it in the other direction. Art that generates computers? Well, the Turing Complete demonstration of Conway's Life is probably close to that. Hmm. I like those old paintings that use perspective really amazingly and have lots of allegory and stuff encoded in them. That sort of use of science in art I can really dig.
So three potential Topics:
The second is quite a heavy subject, and I'm only really interested in it at a hobbyist's level, so perhaps I shouldn't bother. The first is quite compelling.
I've been thinking about organisations, rewards, incentives, recognition, that sort of thing. The SCA start all their members at "minor nobility" which is quite funny, but the interesting part is that everyone is privileged... it's like how the Quakers abolished the laity to make everyone clergy. But in that case, what sets you apart? Why not retroactively apply privilege to everyone? Well, the distinguishing factor is that you sought out and joined these groups, and there's a "membership" process for both. It's a small barrier, but any barrier at all is intellectually a big one.
I was thinking about a group that you can only join by finding something in a particular cave. I guess it could be an intellectual cave. Perhaps there are lots of intellectual caves already: I just outlined two.
Cf. the old "By reading this sentence, you are a member of ..." idiom.
http://d8uv.org/flog/feed
- The Flog's Atom Feed
Whits and The Flog are a bit like a binary star system. I'm not sure what that makes the subscribers... comets? Asteroids?
Cody mentioned Whits mentioning The Flog on the The Flog. I mentioned The Flog mentioning Whits mentioning The Flog on The Flog on Whits. Now Cody has upped the stakes:
"The Flog mentioning Whits mentioning The Flog mentioning Whits mentioning The
Flog on the The Flog on Whits in today's issue of The Flog."
- http://d8uv.org/flog/2006-12#t1166402177
There's also some stuff about Vishnu and the apocalypse. So as far as I follow, I've now mentioned The Flog mentioning The Flog mentioning Whits mentioning The Flog mentioning Whits mentioning The Flog on the The Flog on Whits in today's issue of The Flog on Whits. If Cody mentions that I mentioned The Flog mentioning The Flog mentioning Whits mentioning The Flog mentioning Whits mentioning The Flog on the The Flog on Whits in today's issue of The Flog on Whits on The Flog, then I'll have correctly predicted that he mentioned that I mentioned The Flog mentioning The Flog mentioning Whits mentioning The Flog mentioning Whits mentioning The Flog on the The Flog on Whits in today's issue of The Flog on Whits on The Flog.
Refuse to lose that, Mr. Woodard!
Here I am demonstrating how Whits works to Javier.
New Topic:
http://inamidst.com/topic/drummond
- Jonson's Conversations with Drummond
'excerpts from an 1842 edition of "Notes of Ben Jonson's Conversations with William Drummond of Hawthornden" (January, M.DC.XIX)'
"100 ciphers are still a cipher" - Pierre Villey, via Hugh Grady on SHAKSPER - http://www.shaksper.net/archives/2002/2334.html
Anthony Grafton cites the following log passage from an early
navigator's account of reaching the equator:
"What could I do then but laugh at Aristotle's /Meteorology/ and his philosophy? For in that place and that season, where everything, by his rules, should have been scorched by the heat, I and my
companions were cold."
(Originally posted at 2006-12-18 19:54).
Yanyuwa might be the coolest language ever:
http://en.wikipedia.org/wiki/Yanyuwa_language
Notable features:
There's also an archaic form of the language whose words are known as the "old bugger words". See "Dialects in the Yanyuwa Language", p.52 in this:
http://eprint.uq.edu.au/archive/00000072/01/yanyuwatotal.pdf
- Yanyuwa Dictionary and Cultural Resource, Bradley and Kirton (1992)
Some background colours:
#ffffcc (yellow) http://members.aol.com/gulfhigh2/words7.html
#4a525a (grey) http://daringfireball.net/
#e3ffe3 (green) http://robotwisdom.com/jaj/
#33ccff (turquoise) http://www.kisa.ca/finnish-phrases.html
#ddddff (blue) http://www.zompist.com/syldavian.html
#eeffee (green) http://www.w3.org/DesignIssues/Notation3
#ddffdd (green) http://www.w3.org/DesignIssues/LinkedData
#ccffff (turquoise) http://www.tokipona.org/
Cody asked what the background colours thing is all about:
<d8uv> sbp: http://inamidst.com/stuff/whits/2006/12#N20-1515 is way too vague.
Well it's like this, see? I was thinking that I ought to try out using a background colour in a design sometime, but I wasn't really sure which out of the 256 ** 3 to choose from, so I figured that I'd note down the details of pages that stick in my mind for having particular background colours.
Actually, it'd be quite funny to take an average... let's do that.
>>> hex((0xff + 0x4a + 0xe3 + 0x33 + 0xdd + 0xee + 0xdd + 0xcc) / 8) '0xba' >>> hex((0xff + 0x52 + 0xff + 0xcc + 0xdd + 0xff + 0xff + 0xff) / 8) '0xde' >>> hex((0xcc + 0x5a + 0xe3 + 0xff + 0xff + 0xee + 0xdd + 0xff) / 8) '0xda' >>>
So, #badeda, which you can see in action if you copy the following broken URI out and paste it back together in your address bar:
http://inamidst.com/data/%3Cbody%20style%3D%22background%3A%23badeda%22%3E %3Cp%3EHello%2C%20world%21%3C/p%3E%3C/body%3E
It turns out that the perfect background colour is teal.
http://swhack.com/logs/2004-12-10#T04-20-57
- n3p was named here (by Cody)
Last night I worked on n3p and N3 Grammar stuff a little with DanC and TimBL, trying to use the new n3-selectors.n3:
http://chatlogs.planetrdf.com/swig/2006-12-20.html#T22-18-44
TimBL tried to browse http://inamidst.com/list/n3p/?rdf in the Tabulator but found that it wouldn't work because http://crschmidt.net/ns/list is served as text/plain rather than application/rdf+xml. When I wrote Pyrple, an RDF API, I used the following regexp to sniff the difference between N-Triples/Turtle/N3 and RDF/XML:
r_xml = re.compile(r'^[\t\r\n ]*(<[?!]|<[^ >]+ )')
It isn't foolproof because people might not be using sane encodings, but it should be able to tell the difference between iso-8859-1 and utf-8 N3 and RDF/XML documents. Once you do that test, you can parse accordingly: even if the input turns out to be garbage, at least you don't have to try several parsers on it.
Okay, a dilemma (which is easier to spell than quandary). Cody's latest move in the Flogwhitsomachia is rather a good one:
http://d8uv.org/flog/2006-12#t1166497652
See, I predicted [1]. Cody somehow got his branecell working and realised that if he did as I predicted then I would gloat, hence winning, and if he didn't then he'd lose the game. But then he employed his *other* branecell, one I didn't even know he had, and it turned out that he played a supermove: [2].
So, [3]!
[1] The Flog will mention Whits mentioning The Flog mentioning Whits mentioning The Flog mentioning Whits mentioning The Flog on The Flog.
[2] He predicted that Whits will gloat over predicting that The Flog mentioned Whits mentioning The Flog mentioning Whits mentioning The Flog mentioning Whits mentioning The Flog on The Flog.
[3] Before I post the answer to whether I on Whits will gloat over predicting that The Flog mentioned Whits mentioning The Flog mentioning Whits mentioning The Flog mentioning Whits mentioning The Flog on The Flog, we're in a quantum superposition of Whits gloating over predicting that The Flog mentioned Whits mentioning The Flog mentioning Whits mentioning The Flog mentioning Whits mentioning The Flog on The Flog and Whits not gloating over predicting that The Flog mentioned Whits mentioning The Flog mentioning Whits mentioning The Flog mentioning Whits mentioning The Flog on The Flog. Since I am saying no more than that, that junctive operation cannot collapse, meaning that Cody both cannot gloat back, and cannot say that I have lost the game, since the states involved don't yet have a truth value. If this were a game of chess, this move would be annotated with "!".
Just published a new Topic:
http://inamidst.com/topic/haikulang
- An Haiku Language
This is a summary of some ideas I've had about constructed language, something that I've been working on recently. The basic idea when I started writing the article was to talk about how to *actually* make a language, but then I got sidetracked into giving details about how language worked, and then eventually I realised I wanted to do something else instead.
But what I'd written was pretty interesting anyway, and crystalised a kind of point where I was a day or two ago, so I decided to publish it as a topic with a colophon explaining why it's only kinda half finished. It's still, ironically, the biggest topic that I've written so far, at over 2,300 words, so there's a fair bit of information.
So TimBL asked crschmidt to put some RDF/XML at his /list/ns# ontology namespace, and crschmidt replied that he would if someone made it for him. Eventually the question came down to whether it was his responsibility or mine, but I couldn't remember who made the namespace originally.
It took a heck of a lot of grepping on three different machines to find out the original place. I actually managed to find it via my keylogs, but ironically the place where it all happened turned out to be Swhack:
http://swhack.com/logs/2005-01-31#T21-44-48
- Directory Listing Script discussion
I just hadn't had the right term to search on: I'd been using "ns/list", but it hadn't been mentioned in the logs, only "/list#". Anyway, to cut a long story a little bit shorter, it turns out that it's my responsibility.
So now we're back to my own problem that I hate writing RDF, even in Turtle form. I should probably just go ahead and do it in emacs, but it's more fun to think about writing an Ajaxified RDF editor online. I was thinking about how to export you could post to some magic binaries/filename.rdf URI and it would give you some application/octet-steam data back, which in most browsers means that it prompts a download. The cool thing is that since the filename.rdf part could be chosen in the interface, it wouldn't give you some crappy filename that you then have to change. I like little bits of UI goodness like that.
I suppose I'd better document list.rdf since I made the terms.
http://lists.w3.org/Archives/Public/www-archive/2006Dec/0023
- List Ontology; Thu, 21 Dec 2006 17:35:28 +0000
Contains some lovely goodness for Chris:
Oddly enough, Schemadoc <http://inamidst.com/proj/sdoc/> came in handy for transforming the RDF/XML to XHTML. Not that it wasn't an optional step, but it's nice to show it off.
Editing the schema with Turtle in emacs turned out to be okay, though I needed to read up about datatypes in the OWL reference:
http://www.w3.org/TR/owl-ref/#Datatype
- Datatypes, OWL Web Ontology Language Reference
Incidentally, this is awesome:
http://ebiquity.umbc.edu/blogger/100-most-common-rdf-namespaces/
- 100 most common RDF namespaces
I used it to fill in the namespaces at the top of list.n3, which is the master file of course. Here's an example of where the list ontology is used:
http://inamidst.com/list/phenny/modules/?rdf
- Phenny Modules Directory List, in RDF/XML
Now once Chris posts the ontology documents it should work in Tabulator, which will be cool. I wonder if Tabulator can follow the autodiscovery links from the non-RDF index?
The Dude is Fast:
http://swhack.com/logs/2006-12-21#T17-45-13
- 17:45, Chris deploys the documentation
Ten minutes is a good response time. Anyway, it still doesn't work in Tabulator so I suspect that Tabulator doesn't add 'Accept: application/rdf+xml'. Works in curl:
$ curl -Is -H 'Accept: application/rdf+xml' http://crschmidt.net/ns/list HTTP/1.1 200 OK Accept-Ranges: bytes Date: Thu, 21 Dec 2006 17:52:55 GMT Content-Length: 4918 Content-Type: application/rdf+xml Server: Apache/2.2.3 (Debian) DAV/2 SVN/1.4.2 mod_fastcgi/2.4.2 PHP/4.4.4-8 mod_ssl/2.2.3 OpenSSL/0.9.8c mod_perl/2.0.2 Perl/v5.8.8 Content-Location: list.rdf Vary: negotiate,accept TCN: choice Last-Modified: Thu, 21 Dec 2006 17:43:51 GMT ETag: "6fc03d-1336-dfd063c0;e00d6cc0"
Ah, it works on refreshing; must've just been cached. Excellent.
Another new Topic:
http://inamidst.com/topic/songwriting
- Principles of Songwriting
I owed DanC this from back in January 2006 when he wrote a little essay about songwriting. It turned out to be a 4,500+ word essay, which is why I haven't written as much in Whits today.
So, given that Ebenezer Scrooge eventually reformed his ways, why do we tend to call people Scrooge in the bad sense? Kinda odd.
Since Whits posts actually get published in the feed the day after I write them, Merry Christmas everyone! Please write in if anyone sees a snoobird!
An heavy Boxing Day rambling: metameaning.
The great metaphor of the book that opens, that one pores over and reads in order to know nature, is merely the reverse and visible side of another transference, and a much deeper one, which forces language to reside in the world, among the plants, the herbs, the stones, and the animals.
- Michel Foucault, The Order of Things
I was thinking that when people use words like polyvalency, not only are they tersely encoding some bit of knowledge to make their communications run smoother but they're also forcing the reader to supply the background context, so it means they must be initiated in some respect into the writer's world in order to be able to keep up.
When William said that "the clutter is inherent to the organism", I guess that's the other side of the coin from the "irreducible complexity" of nature:
The idea of the Book as a totality, as a complete and unified structure that galvanizes the polyvalency of language into a single, determinate meaning, was (and perhaps in a residual sense still is) the model for our understanding of the world itself, for the way in which people try to make sense of the irreducible complexity of what we call nature, or reality.
As Tim says in Weaving the Web, knowledge is the connections between things, so the meaning of meaning is the meaning of connections. That's the philosophy that the Semantic Web, and even the World Wide Web, is founded on. Hypertext was a bit of a strange subject before the Web demonstrated the viability of it, and now we don't consider it too much because it's so transparent, so simple. But hypertext can have any of many realisations... the actual Web as it now exists just happens to be a very practical one.
In E.M. Forster's Howard's End, Margaret Schlegal has a sort of motto, "Only connect," which she lives by. Her pleasure, her human function, is to build a "rainbow bridge" of meaning between scattered, contrary things.
- Stan Dragland, AirWave DreamScapes
It's strange in a sense that language started off as being very ideogrammatical and went on to being grammatical. Languages go from symbols representing concepts to symbols representing sounds augmented by symbols representing concepts to symbols just representing sounds. On the other hand, these latter usually have enough texture to still be considered the representations of ideas, so it's all a matter of encoding; what is the optimal encoding.
Language is odd as a portrait of the world, because it's no different in a sense to constructed fantasy worlds, or paintings, or whatever; it doesn't even necessarily try to maintain any connection to the world that it describes. I can talk about unicorns for example, and I can use language to paint a fantasy world. Language exists for communication, not to be a true model of the universe, and yet inasmuch as it exists within the universe and does only what it does, it is true.
The planter, who is Man sent out into the field to gather food, is seldom cheered by any idea of the true dignity of his ministry. He sees his bushel and his cart, and nothing beyond, and sinks into the farmer, instead of Man on the farm. The tradesman scarcely ever gives an ideal worth to his work, but is ridden by the routine of his craft, and the soul is subject to dollars. The priest becomes a form; the attorney, a statute-book; the mechanic, a machine; the sailor, a rope of a ship.
In this distribution of functions, the scholar is the delegated intellect. In the right state, he is, Man Thinking. In the degenerate state, when the victim of society, he tends to become a mere thinker, or, still worse, the parrot of other men's thinking.
- Ralph Waldo Emerson, The American Scholar
Hmm, upgrading from Python-2.5c1 to Python-2.5 hosed the bsddb module for some reason, so I've installed the latest Berkeley DB and reinstalled Python-2.5. Hopefully that'll sort it out.
Okay, that didn't work, but that's hopefully just because I installed Berkeley DB 4.5.20. I rooted around in the Python 2.5 documentation, and it says:
"The bsddb module requires a Berkeley DB library version from 3.3 thru 4.4." - http://docs.python.org/lib/module-bsddb.html
So now I've installed Berkeley DB 4.4.20 and reinstalled Python, which is handy anyway since last time I forgot to use the --enable-unicode=ucs4 flag to ./configure which is needed for rdflib and n3proc:
http://chatlogs.planetrdf.com/swig/2006-12-22.html#T23-23-19
If this doesn't work, perhaps I should try reinstalling Python-2.5c1 or something.
Hooray, BDB 4.4.20 works! And all of the diffs have appeared in
http://inamidst.com/changes/ retrospectively. Awesome.
Just in case anybody's Googling for this with the same problem (Google is really unhelpful at the moment on this), here's the traceback I got:
File "/usr/local/lib/python2.5/anydbm.py", line 82, in open
mod = __import__(result)
File "/usr/local/lib/python2.5/dbhash.py", line 5, in <module>
import bsddb
File "/usr/local/lib/python2.5/bsddb/__init__.py", line 51, in <module>
import _bsddb
ImportError: No module named _bsddb
And a summary of how I fixed it:
If Python-2.5 had already been installed, reinstall over it, but make sure that you do the BDB installation first. This was on Mac OS X 10.4.8.
Again, since these posts go in the feed the day after they're actually written... Happy New Year!
"308 Factor This Huge Number Into Primes Before I Give You The Real Content" - From a conversation about request frequency in crawlers, and server loads
Sean B. Palmer, inamidst.com