Page Title

, images without alt texts, 02:27:32 images with rubbish alt texts, most kinds of ECMAScript garbage, 02:27:34
used to indent text, images used as spacers, non-relative 02:27:36 units used in stylesheets for all output modalities,
 used to
02:27:38  layout text, excessive use of 
, weird markup idiosyncrasies such
02:27:40  as bad source code layout, animated GIFs,  not being used
02:27:42  where it should be, totally proprietary markup like , ,
02:27:44  and (shudder) , transitional DTDs being used on strict and vice
02:27:46  versa, lack of CSS presentation and illustrations to help the
02:27:48  cognitively disabled where these could easily be provided (erm... I'm
02:27:50  probably guilty of that one), lack of support for i18n, people
02:27:52  thinking that _{and ^{are presentational elements, links with
02:27:54  "click here" text, and pages with large images or scripts. There's
02:27:56  probably some more stuff as well.
02:27:58  ]]]
02:28:00  woah! bigger paste than I expected... sorry
02:28:23  heh, interesting that the first thing on the list is what the new validator.w3.org page is guilty of
02:31:50  Hmm, what's the device-independent way to refer to web page visitors?
02:31:57  Visitors, I guess?
02:33:28  or 'users', which doesn't assume that they're visiting
02:33:53  Is someone who reads a printed copy of my website a user?
02:34:20  possibly... are they visiting it? I think not
02:34:42  it depends how you think of a page
02:34:57  I suppose so. Users just implies computers to me...
02:35:04  * sbp wonders what on earth Aaron and db are babbling about
02:35:23  Is it a place you can visit? Is it some text that gets processed into something else?
02:35:23  * AaronSw decides on "Don't say "click here": not everyone will be clicking"
02:36:02  ah, heh
02:37:55  for the record, coming from Mr. Pedantic... er Palmer, I think that "visitors" is just fine, because it implies online visitors. The act of visiting is analogous to browsing to a site via HTTP on the World Wide Web
02:38:10  But these people aren't browsing, they're reading dead trees.
02:38:18  sure, I like 'visitors'
02:38:27  exactly, so they're not incrementing the page counter
02:38:31  Ahh, Jutta to the rescue: http://kbs.cs.tu-berlin.de/~jutta/ht/writing/words.html
02:38:49  even better: http://kbs.cs.tu-berlin.de/~jutta/ht/writing/text.html
02:40:47  Who's Jutta?
02:41:09  Created Halfbakery.com, wrote a nice hypertext writing guide
02:41:32  er... I guess I should have asked, "what's her name?"
02:41:53  Jutta Degener
02:42:03  see http://www.cs.tu-berlin.de/~jutta/index.html
02:42:45  Dwarf kangaroo?
02:43:03  Huh?
02:45:29  - http://kbs.cs.tu-berlin.de/~jutta/me/whois.html
02:46:12  Heh: "Portrait of the artist as a young alligator"
03:22:11  sbp has quit
03:40:44  gerald has quit
03:41:04  gerald has joined #rdfig
03:51:12  sbp has joined #rdfig
03:51:21  thanks for the pointer
03:51:30  to http://www.w3.org/TR/REC-html40/struct/text.html
03:52:12  Hmm, Tim doesn't link to "buying a mower" in the example!
03:52:57  [[[
03:52:58  Good link text should not be overly general; don't use "click here." Not only is this phrase device-dependent (it implies a pointing device) it says nothing about what is to be found if the link if followed. Instead of "click here", link text should indicate the nature of the link target, as in "more information about sea lions" or "text-only version of this page".
03:53:14  ]]] - http://www.w3.org/WAI/GL/WD-WCAG10-TECHS/html-techniques#links
03:54:05  You want to email that one or should i?
03:55:12  I will...
03:56:05  done
03:56:14  there's a more up-to-date link
03:56:57  I sent a couple more things to www-qa about this stuff
03:57:34  Heh, "use  for titles"
03:57:54  Hmm, that came out a bit wrong, didn't it? :-)
03:58:04  Sheesh, such bad examples in WCAG10-TECHS
03:58:10  [[[
03:58:11  Sheesh, such bad examples
03:58:20  Oops:
03:58:25  You forgot to say *why* using  at the top level is bad
03:58:43  send an update!
03:58:50  I don't have to explain myself. ;-) You send one.
03:58:53  Why is it bad again?
03:59:12  Bad example: [[[
03:59:12    My document is available in HTML,
03:59:12    PDF,
03:59:12    plain text
03:59:12  ]]]
03:59:35  why is that a bad example? Except that it forgets the media type attributes
03:59:43  and it's not XHTML
03:59:58  and there's no "and" before "plain text"
04:00:08  Nor a period.
04:00:36  And "My document is available" shouldn't be linked, which is what I was thinking of.
04:01:38  Although, perhaps that's style.
04:02:21  * sbp sends the  addenda
04:02:30  no, I think you're right
04:02:41  and in that case, a title needs to be added to the first link as well
04:03:09  and for these titles to be useful, "my document" should be defined more clearly
04:04:06  that's the best reason you could come up with? tree structure?
04:06:12  that's the only reason
04:06:24  Surely, there's a better reason...
04:06:28  isn't it? are there any more?
04:06:36  well, default styling is affected
04:06:56  it's disorientating
04:07:01  I think it simply sets bad prescedent.
04:07:08  yes
04:08:58  I sent my "XHTML Tips" rant to www-qa as well
04:09:10  whilst we're bombarding it with stuff :-)
04:09:25  oh, you sent a reply already
04:09:51  use acronym: hooray! I was hoping someone would point that out
04:09:55  Another good style tip: link to the printable version of stories when possible ;-)
04:10:34  actually, there are two points there: expand acronyms, but not for obvious acronyms, and realise that support for  is at best terrible
04:10:50  My browser supports !
04:10:57  so does mine
04:11:04  well, one of mine
04:11:04  Hey, you should do a Quality Tip, then they'd have three and that'd be enough to launch. I'm sure you could do hundreds -- you're quite good at it.
04:11:10  Which one?
04:11:15  Ns6
04:11:28  IE5/Mac
04:11:42  Hmm, you're quite good at it. I'll have to take you on my "Semantic Foundations of the Web" book tour.
04:11:57  Ns6 is actually quite good, it even puts a little question mark on the cursor when you hover over an  or  element
04:12:02  book tour?
04:12:35  Yeah, I hope to write a Semantic foundations of the Web book.
04:12:45  A book on Web Architecture, really.
04:13:03  question mark: ooh, nice
04:13:12  They clearly took CUAP very seriously.
04:14:32  Yeah :-)
04:15:09  we're really quite good at these tips
04:15:29  Yeah... to bad we're the few...
04:15:41  yeah...
04:15:49  Saw It's funny, I see people saying things like "use  instead of " and I see them as kindred souls!
04:16:01  Saw JDarcy saying that the other day.
04:16:08  "Only do that for the first instance in the document, though" why?
04:16:52  because it's pointless doing it all of the time. If you have the acronym "SW" linking to the Semantic Web activity homepage 500 times in a document, it quickly becomes tiring for both the author and the user
04:17:23  I'm assuming in this situation that they'd be linked automatically (like in a Wiki). WHy is it tiring for the user?
04:18:22  For the user: because, like it or not, links disturb the flow of text. They should only be used when necessary. I mean, you could take it to the extreme and link every word to its dictionary definition, couldn't you?
04:19:00  Although this is getting close to the realm of personal opinion... some people like documents where almost every word is linked. DanC seems to fall in that category
04:19:42  :Linker rdfs:subClassOf foaf:Person; rdfs:comment "person who likes linking every second word or so in a HyperText document" .
04:19:58  I like linking every third word or so :-)
04:20:07  :MildLinker ...
04:20:15  Aha, here's the blog I was talking about http://www.platypus.ro/diary/
04:21:40  Can you point to something where DanC links every other word?
04:22:03  Now Joe Clark's piece, that was good -- he had almost every word linked. I think it was 7-hundred something altogether.
04:22:13  I can't seem to find it though
04:23:36  every other word: http://www.w3.org/People/Connolly/pubs
04:24:13  Oh, come on -- it's a links page!
04:24:26  Perhaps
04:24:41  but all of Dan's pages are links pages
04:24:49  Not true
04:24:57  Hmm, good point:
04:25:04  [[[
04:25:04  The MENU element should work with stylesheets
04:25:04  ... to represent the following idiom: 
04:25:05  Search | Index | Products | Services
04:25:05  ]]]
04:25:24  it is true, and I'm certainly not condemning it
04:25:40  where'd you get that from?
04:26:13   is deprecated
04:26:22  that page: http://www.w3.org/People/Connolly/pubs
04:27:22  You can do it with  and CSS... let me do an example
04:31:12  * sbp wonders how to do this so that it will gracefully transform...
04:31:53  ta da:-
04:31:53      xml:lang="en" >
04:31:54  
04:31:54  Page Title
04:31:54  
04:32:10  
04:32:12  
04:32:14  
04:32:16  [Search]
04:32:18  [Index]
04:32:20  [Products]
04:32:22  [Services]
04:32:24  
04:32:26  
04:32:28  
04:32:30  ignore all of the other accessibility problems with the page, such as no DTD, title, context, weird links with no titles, no address...
04:32:38  Yeah, but then you have:
04:32:54  [ Search ] [ Index ] [ Products ] [ Services ]
04:33:00  Not:
04:33:06  A | B | C | D
04:33:07  no, you have: [Search] [Index] [Products] [Services]
04:33:16  Oh, right
04:33:22  ah, but the NLPCs have to be part of the structure, not the style, for a start
04:33:29  NLPC?
04:33:32  and to do that, you'd be doing ... |
04:33:38  non link printable character
04:33:41  Hah,  attack again.
04:33:46  yeah :-)
04:34:19  NLPC
04:34:32  happy now?
04:34:38  heh, thanks
04:34:42  :-)
04:35:40  the problem, as ever, is in maintaining backwards compatability. If old UAs didn't suck, there'd be no problem
04:35:49  or: less of a problem
04:39:06  Wow, this is... er... "different": http://www.w3.org/TR/REC-DSig-label/
04:40:37  Umm, how so?
04:41:21  Other than that style of many early W3C publications which can only be described as: "Huh?"
04:42:22  Wow, Brian LaMacchia was an author... I coulda been working with him in RDF.
04:44:17  Neat
04:44:52  So what is "different" about that draft?
04:46:55  Speaking of the "Huh?" era of W3C style, I love this: http://www.w3.org/TR/REC-DSig-label/pics
04:47:17  It rambles on a bit, and with RDF and HTTP-ext, I'll bet it would be done a *lot* differently today. IT's also probably one of the lowest implemented Recs ever, although I certainly can't verify that slanderous claim
04:47:34  I think it has some competition. ;-)
04:47:45  yeah, that's a neat image
04:48:01  I love how it's blown up 2x in the document, leading to severe pixellation.
04:48:23  That goes right up in the history with Timicons
04:48:25  yes...
04:48:28  *that* is why we need SVG :-)
04:48:30  Heh!
04:48:53  Timicons?
04:49:12  SVG... love it, but when are the implementations coming? hello? implementations? I mean, on the same 95% scale that flash is deployed on
04:49:58  Timicons: a term i just coined for the cute little icons such as http://www.w3.org/Icons/WWW/activity_48x48
04:50:18  Rumored to be from Tim, and perhaps Henrik
04:50:59  Those icons have some real gems.
04:51:42  heh...
04:51:44  [[[
04:51:44  Oct 95: Infuriated by finding documents containing obsolete icons, I have moved them into "unused" in the hope that the MOMspider will catch them and the miscreants will be plagued by messages. - Tim 
04:51:54  ]]] - http://www.w3.org/Icons/WWW/
04:51:57  Yeah, I loved that.
04:52:12  Better living thru metadata: http://www.w3.org/Icons/WWW/pics_48x48
04:52:12  "(This is mostly to try and prevent Henrik from making any more stupid icons) " - ibid.
04:52:46  * AaronSw wonders who added that.
04:53:24  I do like http://www.w3.org/Icons/WWW/WWW48x though
04:54:08  Hmm, why doesn't RDF have a "better living" icon like that?
04:55:02  I wonder what the copyright on those icons are?
04:55:16  Iwonder what sort of psychoactive drugs led to  being the icon for OS. ;-)
04:55:31  [[[These icons are under the copyright of the W3C. They can be used by others, but only if they are used in referring to the appropriate part of the W3C web site. ]]] - ibid.
04:55:59  oh yeah
04:56:09  * AaronSw will have to ask Henrik about that one someday
04:56:13  yeah, weird
04:56:25  Earth in a crystal ball?
04:56:41  I like using ruby now as WWW symbol
04:56:50  ruby?
04:57:03  i mean, how would you use ruby?
04:57:12  World Wide Web over WWW?
04:57:26  like the one on: http://infomesh.net/sbp/?HomepageNotes
04:57:55  Hmm, I think I will have to blame Henrik, since the icons seem to appear mostly on Henrik-signed pages.
04:58:04  yeah
04:58:05  Unless it's a set up...
04:58:21  "This is a PURL or permanent URL." - ?HomepageNotes
04:58:30  You keep making this mistake. It's Persistent URL.
04:58:35  bugger
04:58:41  I had to fix it on the SWAG site.
05:10:03  sbp has quit
05:19:21  http://www.ics.uci.edu/~fielding/talks/webarch_9805/ 
05:19:21  B: http://www.ics.uci.edu/~fielding/talks/webarch_9805/ from AaronSw
05:19:34  B:|Web Architecture in 40 slides or less
05:19:34  titled item B
05:19:41  B:By Roy Fielding
05:19:42  commented item B
05:20:08  B:for those that couldn't/wouldn't catch [the original|http://www.ebuilt.com/fielding/pubs/fielding_dissertation_2up.pdf] (PDF) here's Roy's Dissertation in handy slide form
05:20:08  commented item B
05:20:49  B:Summarizes Roy's view of Web Architecture, but leaves out the connections to "A Pattern Language" that (for me) made it all click
05:20:49  commented item B
05:21:07  B:I'll summarize (since he didn't):
05:21:07  commented item B
05:22:22  B: [A Pattern Language|http://www.amazon.com/exec/obidos/ISBN=0195019199/coolbooks02] talks about architecture (real buildings) but is so good it has been used as a guide for life, programming patterns, etc.
05:22:22  commented item B
05:23:12  B:It's argument is that architecture is like poetry, the more meaning (symbolized/identified by the patterns) you cram into a smaller space, the better the space.
05:23:12  commented item B
05:23:48  B:Roy extends this to Protocol Architecture, showing how HTTP crammed all the other paradigms into one.
05:23:48  commented item B
05:24:06  B: The end result, I think most would agree, was pretty cool. Even if it didn't go quite as planned...
05:24:06  commented item B
05:31:20  B:Good quote: "An XML representation is an object turned inside-out, with behavior-by-reference"
05:31:20  commented item B
05:31:44  B:([cite|http://www.ics.uci.edu/~fielding/talks/webarch_9805/tsld034.htm])
05:31:44  commented item B
05:43:00  oierw` has quit
05:43:00  oierw` has joined #rdfig
06:07:55  AaronSw has quit
07:01:47  AaronSw has joined #rdfig
08:13:59  danbri_ has joined #rdfig
09:18:47  danbri_ has quit
09:56:58  oierw` has quit
11:27:40  jang has quit
12:09:23  rreck has joined #rdfig
13:00:53  dmiles has joined #rdfig
13:54:52  dmiles has quit
14:09:01  dajobe has joined #rdfig
14:22:34  AaronSw has quit
14:23:24  AaronSw has joined #rdfig
14:52:05  jang has joined #rdfig
15:04:39  A:
15:04:39  http://validator.w3.org/
15:04:40  New validator version released
15:04:41  (AaronSw) Nice work to Gerald and Terje -- looks really great.
15:14:41  ambient has joined #rdfig
15:16:58  ambient has quit
15:19:25  ambient has joined #rdfig
16:11:23  ambient has quit
16:36:56  sbp has joined #rdfig
16:41:26  sbp has quit
16:53:12  * danbri admits defeat...
16:53:38  Just spent a day and a half trying to fix up my linux box, replace a patched up redhat 5.2 with a shiny new SuSE installation...
16:54:29  Oh well.
16:54:47  Total screw up. Partly due to some kludgy disk-partition juggling on my part, partly confusion due to SuSE installer (it didn't install a kernel), whatever, wasted whole bunch of time.
16:54:54  * danbri shrugs
16:55:51  So you never got SuSE running?
16:57:27  Not properly, no. I don't have much disk space left in the < 1024 part of the disk I'm targetting, and everything else is the 80G reiserfs driver which isn't good for booting off; windows is hogging most of the 3rd drive, so I'm shuffling stuff about trying not to trash the old setup before new works...
16:58:05  I just did most of that with debian replacing redhat
16:59:17  How was it? I've heard Debian's good but a bit down-to-the-metal installerwise, whereas SuSE supposed to be pretty slick. 
16:59:38  the isntaller really really really sucks
16:59:53  * danbri didn't think "didn't find kernel on installation media; sorry no kernel" too slick (SuSE FTP install...)
17:00:03  but with reiserfs in the mix, you are on shakey ground for most everything
17:00:33  * AaronSw remembers installing Redhat via FTP... over, and over, and over again
17:00:46  pah, redhat's dead easy to install
17:00:56  Yeah, if you shell out the cash.
17:00:59  can do it over a short lunch
17:01:03  nope; 1 floppy
17:01:18  I didn't touch the reiserfs disk this weekend, mostly. Trying to squeeze a modest SuSE into 2Gb alongside the older living redhat.
17:01:48  modest? 2G is rather huge for linux
17:02:10  Yeah, but most of it is clagged up with the old one. 
17:02:33  First time I succumbed to reinstallation greeed, admittedly ;-)
17:02:41  (all those intriguing looking new packages...)
17:03:23  bijan has joined #rdfig
17:03:30  hi bijan
17:03:33  Hi dan.
17:06:03  oierw` has joined #rdfig
17:06:26  * bijan got a buncha common lisp books for b-day. Having fun.
17:06:45  http://www.aw.com/product/0,2627,0201719622,00.html
17:06:46  C: http://www.aw.com/product/0,2627,0201719622,00.html from bijan
17:07:26  C:|Building Parsers with Java(TM) (non-TM javaish implementations need not apply) by Steven Metsker
17:07:27  titled item C
17:07:42  C: Spends a *lot* of time about implementing simple logic programming facilities in Java.
17:07:43  commented item C
17:08:43  C:I only skimmed it in thet bookstore, but it's definitely worth further exploration.
17:08:43  commented item C
17:09:09  C:If it's as competant as it appears, it will be a valuable text for semantic webbers of the Java taint.
17:09:09  commented item C
17:17:55  Oh, is it your birthday today?
17:22:00  No.
17:22:28  Well, happy unbirthday then!
17:23:03  Er...thanks.
18:35:20  dmiles has joined #rdfig
18:40:31  tav|offline has quit
18:40:41  lasDesk has quit
18:41:44  lasDesk has joined #rdfig
18:48:48  tav|offline has joined #rdfig
19:13:30  [GlobalNotice] Hi all.  Services will be down for 15 minutes, perhaps longer.  We'll take it down in about 5 minutes.  Apologies for the inconvenience.
19:18:04  sbp has joined #rdfig
19:20:39  dmiles has left #rdfig
19:29:11  sbp has quit
19:41:46  Ooh, version 0.8.10alpha1, the first mostly feature-complete version of E is out
19:41:52  Time to learn E, I guess.
19:42:20  Heh. I'd wait for the mostly working mostly feature-complete version is out :)
19:42:40  Heh.
19:44:17  http://www.mathematik.uni-osnabrueck.de/projects/dcqual/qual21.3.1/qual29-8-1.html
19:44:19  D: http://www.mathematik.uni-osnabrueck.de/projects/dcqual/qual21.3.1/qual29-8-1.html from danbri
19:44:46  D:|Expressing Qualified Dublin Core in RDF, 2001-08-29
19:44:47  titled item D
19:44:56  Ooo really?
19:45:35  D:authors: Stefan Kokkelink, Roland Schw�nzl 
19:45:35  commented item D
19:45:58  D:a product of the dublin core architecture WG
19:45:58  commented item D
19:47:40  * danbri and Aaron are trying to help wordsmith this doc, improve the english; suggestions welcomed
19:49:52  So the first paragraph ends with a couple of sentences about URIs/URLs/URNs that sort of fizzles out...
19:50:49  * bijan is getting nothing from it, sorry.
19:50:55  Adding something like "DC in RDF adopts URI syntax as a common convention for naming resources, and for naming the elements of metadata vocabularies."
19:51:00  from what? whole doc?
19:51:06  It's not skimmable, alas.
19:51:09  ...might justify the mention of URIs
19:51:12  sandro has quit
19:51:12  mids has quit
19:51:12  kifbot has quit
19:51:12  bijan has quit
19:51:30  bijan has joined #rdfig
19:52:03  The notion of resource in RDF is formal. Any resource can be given a URI, but one is not forced to expose a choice visibly to the outside world. There is at most one choice one can provide to the outside world. RDF treats different URI as different resourc
19:52:08  mids has joined #rdfig
19:52:11  I really feel that all the intro parts need to be cut out.
19:52:12  is this coherent? Or useful?
19:52:22  sandro has joined #rdfig
19:52:33  A literal is an object one cannot make any assertions about. 
19:52:37  That seems false.
19:52:52  Unless it means "literals cannot occur in the subject position"
19:53:19  The latter seems to be the sense aimed at.
19:53:20  But that's a *covoluted* way to say it.
19:53:30  [GlobalNotice] Please bear with us all, as we reconnect services in a new configuration
19:53:31  Plus, doesn't make clear that they can't be predicates either.
19:53:46  It's getting into 'intro to rdf' territory.
19:54:11  Taken together the notion of resource and the notion of literal, it appears obvious to require resources as subjects of assertions, but not to make limitations on the object.
19:54:13  The RDF Core WG are just beginning work on an RDF primer. It might be that some bits of this doc are best moved to general RDF intro... 
19:54:43  Yes, pleeease�
19:54:44  Why one earth are they trying to derive the restrictions on literals from some weird first principles.
19:55:08  Plus, I consider rdf statements to be as much assertions about the object as the subject.
19:55:16  bijan: is the 'it seems obvious' bit a quote? 
19:55:18  * bijan doesn't like the whole SPO stuff.
19:55:23  Yes.
19:55:36  found it, sorry.
19:56:57  Kill all the "metametadata" stuff.
19:56:59  bijan: that's a fair point; the direction one names the label for shouldn't really matter (worksFor / employee etc...)
19:57:01  Hmm, I wrote a big long rant about corrections it needed, but I can't seem to find it.
19:57:13  The whole 0.2 can go.
19:57:47  danbri: predicating something of a subject always involves an assertion "about" the object as well as about the subject, I'd think.
19:58:03  Yes, I'd drop 0.2.
19:58:03  * AaronSw is itching to just rewrite it...
19:58:12  To think otherwise is to embed the object in the predicate.
19:58:15  bijan: I thought I was agreeing
19:58:23  yes, but not enough!
19:58:24  :)
19:58:38  I.e., it's not because of directionality.
19:58:44  It's because of two placeness.
19:59:00  [GlobalNotice] Okay, one more shot folks....in a few minutes....this configuration should be a bit more stable
19:59:08  Aaron, go for it! I'm in pretty useless shape today (toothache, with Linux installer fatigue) but I'll happily chime in if you want to go for a substantial reorganisation/rewrite...
19:59:15  Ouch.
19:59:20  * bijan stops picking on danbri.
19:59:29  Ow. :-(
19:59:32  I meant, the directionality is pretty arbitrary
19:59:41  Here's my question: what's the useful content of this doc?
19:59:46  I.e., where is it.
19:59:52  Section 2?
19:59:55  Useful content: it's better than the _last_ RDF DC spec. 
20:00:02  Section one basically sucks.
20:00:16  kifbot has joined #rdfig
20:00:27  Section 1 or section 0?
20:00:38  I mean, I presume the issue is how to take the informal dc set of elements and qualifiers and represent them in rdf.
20:00:39  The technical content, the way it suggests DC folk use RDF, it a vast improvement. Previously, DC didn't try using schema, so instance-level data bore the burden of doing the qualification.
20:00:46  Aaron: section 1 too.
20:01:01  "The hedgehog model" Huh? What on earth.
20:01:20  The issue is how to get a spec published by DCMI that says "here's how DC looks in RDF"
20:01:44  danbri: that woudl be wonderful. I'd love to have a canonical dc/rdf.
20:01:56  But does this document contain that?
20:02:14  I'm thinking for section one, I just say: these are the elements, you use them like this.
20:02:15  It's the closest we've got, to date...
20:02:38  The word 'elements' suggests XML; some other DC folk seem to be using 'terms' lately.
20:02:48  yes, terms is good.
20:02:54  Aaron: section 1 is worse than that, it tries to explicate "past practice".
20:03:09  If it's a spec, it shouldn't try to be a tutorial.
20:03:14  If it's a tutorial, point me to the spec.
20:03:28  Yep, that's why I'm rewriting
20:03:30  RDF syndrome; RDFCore's going through the same thing.
20:03:37  danbri: oh yeah.
20:03:51  I mean, *all* I want to know is how to express the range of DC in rdf.
20:04:26  I'd like to have the machine bits (the schemas) and then nice direct explanatory text telling me how to go at it.
20:04:40  [GlobalNotice] Okay folks, here goes.  Expect some splitting as we attempt to connect services in a new configuration.  Apologies for the inconvenience.
20:05:01  Hmm, should I list the terms or simply tell folks how to convert DC specs into RDF?
20:05:04  dmiles has joined #rdfig
20:05:13  There have been more schema (drafts?) floating around lately, tangled up with DC registry stuff?
20:05:16  Who is this targeted to?
20:05:25  "metadata implementors"...
20:05:29  "dc implementors".
20:05:29  I mean, which ones?
20:05:32  DC experts?
20:05:34  RDF experts?
20:05:36  A technical audience, really. 
20:05:40  DC and RDF experts?
20:05:45  Neither?
20:05:49  I'm aiming for folks that know what RDF is, what DC is, what XML is, etc. and simply want to know how to put things together.
20:05:51  USer guides etc would be more for content creators.
20:05:59  Bijan "He Who Must Be Catered To" Parsia?
20:06:05  ChanServ has changed the topic to: Semantic Web Chat - http://www.w3.org/RDF/Interest/#irc
20:06:05  This channel is logged and blogged: http://logicerror.com/rdfIRCWelcome
20:06:09  Aaron, should we sit back and wait to see what you come up with?
20:06:22  Well, give me some advice! ;-)
20:06:30  Well, first go for value add.
20:06:42  [GlobalNotice] That looks like a more reliable connection.  Thanks to everyone for your patience, it looks as if we're done.
20:06:49  If there are a slew of other docs that *already* do stuff (like explain RDF, RDF/XML, DC, etc.) just point to them and try to fit in.
20:06:51  I'm wondering whether I should list the elements definitively or describe how to convert... sticky either way...
20:07:40  danbri? advice?
20:07:46  If that turns out to be inadequate, or one has extra resources, then one can write "How to get into RDF using DC" and "How to get into DC using RDF" And "How to get into XML using DC and RDF" (ick) and "Why Java Sucks for DC and RDF" (oops) and...
20:07:48  Am i making sense?
20:07:55  OK: Up front, it should say: Dublin Core metadata is written in XML, modelled in RDF. It follows the RDF convention of using URIs (and XML namespaces) to identify properties/attributes/relations, as well as types of resource (classes). 
20:07:56  Label OK not found.
20:08:22  Actually, there are documents about expressing rdf in HTML and XML.
20:08:26  Model yourself on them.
20:08:30  I'm not convinced, danbri.
20:08:31  ...and that we use the RDF notion of property and class hierarchies to express qualification, and specialisation
20:08:38  you're not?
20:08:52  bijan, which do you have in mind?
20:08:56  I think we should just say: this is how to do it in RDF, and leave that stuff for other guides
20:09:11  * AaronSw is looking at http://dublincore.org/documents/2001/04/11/dcmes-xml/
20:09:16  * dajobe groans
20:09:29  Should I copy you, dajobe?
20:09:31  That groaning is clearly my fault :)
20:09:45  hey, I didn't say anything about you
20:09:46  hey, qualify your attributes, dajobe!
20:09:58  Yes, but context setting: people need to know that the conventions we're using here (URIs for properties, classes) come from the W3C RDF (model/syntax/schema) and XML (namespace) specs.
20:10:08  dcmes-xml in that state is under review, don't look at it now
20:10:32  ie. we punt off to W3C for further articulation of those specs / technologies (rather than leave as if further DC specs might provide detail).
20:10:43  dajobe: not even as a model? I.e., for a draft extention?
20:10:51  It doesn't seem so terrible.
20:11:10  OK, danbri -- you can do that section (context setting) ;-)
20:11:14  I mean, I'm not dancing in the streets singing "I've got that groovy feeling", but it's not like I'm reading Javadoc.
20:11:26  Heh!
20:12:08  http://www.ietf.org/rfc/rfc2731.txt
20:12:08  E: http://www.ietf.org/rfc/rfc2731.txt from bijan
20:12:22  E:|Encodign Dublin Core Metadata in HTML
20:12:23  titled item E
20:12:38  E:Another reference for Aaron, the document-serf.
20:12:38  commented item E
20:13:04  context: ok, will do...
20:14:43  I'm still missing whether the orginal doc set up an RDFS for the qualifiers.
20:15:40  No, this is the first doc that mentions qualfiers.
20:15:48  I'm thinking of punting to http://dublincore.org/documents/dcmi-namespace/, however
20:15:57  'orginial doc' == 'doc under revision'
20:16:13  Oh, yes, it did.
20:16:20  I think it's an appendix.
20:16:39  So there's two aspects, yes?
20:16:47  What's the other one?
20:16:49  The syntax (uri choice/namespace, etc. etc.)
20:17:00  The inferential constraints (i.e., the schema)
20:17:24  The first I think we can punt to http://dublincore.org/documents/dcmi-namespace/
20:17:27  The doc seems to muddle these two issues together.
20:17:34  For the second we can punt to the actual schemas
20:18:00  I want to focus on (the third) which is about how to use these terms in RDF
20:18:14  i.e. in what contexts they make sense, and what are their semantics
20:18:33  Hmmm.
20:18:48  There semantics are given by the DC specs, yes?
20:18:54  They're
20:19:04  Err, yes... for the terms.
20:19:16  But it doesn't define, what, say, repeated elements mean, vs. bags, vs. subclasses, etc.
20:19:43  now *that* bit is interesting.
20:20:02  Seems to me that there's a simple case and then there's all the complications.
20:20:20  Very much analogous to the core terms + literal case vs. using qualifiers, etc.
20:20:29  Yep.
20:21:08  So, it might be worth enunciating as clearly as possible the core case and then deal with the complications as elaborations or departures from the that case.
20:22:11  OK, let me try and write that up and then you can comment...
20:22:28  muhWAHAHAHAHAHAHA!
20:22:36  'scuse the drool.
20:24:44  http://logicerror.com/dcrdfDraft
20:24:44  F: http://logicerror.com/dcrdfDraft from AaronSw
20:24:51  F:|Dublin Core in RDF Draft
20:24:51  titled item F
20:25:03  F:My rewrite of the Dublin Core draft, updated live
20:25:04  commented item F
20:26:35  We encode metadata elements by making the first letter lower case, and appending them to their specified "namespace". (See [DCNS] for namespaces of Dublin Core elements.)
20:26:49  I don't see why you can't list the pertinent namespaces there.
20:27:09  Er...
20:27:11  Is that it?
20:27:16  because they're going to be changed/added to and DC process for updating drafts is torturous
20:27:30  so i figure we leave one doc (dcns) to be fixed
20:27:34  You have to do that for your examples anyway.
20:27:49  You mean you're not generating this using a function which snags the right namespace and inserts it everyone!!!
20:27:51  :)
20:28:02  heh
20:28:14  * bijan singsongs "Aaron's using static HTML..."
20:28:41  I'm not!
20:28:47  DCMI is, tho
20:28:55  If only  supported HTML...
20:28:58  
20:29:13  * bijan singsongs "Aaron's using something that might as well be static HTML but isn't which makes it worse..." :)
20:29:31  pfft.
20:29:39  :-)
20:29:45  Er..and is using n3/triples as the example really the best way to explicate? Especially without mentioning that you're doing it?
20:30:35  I might have an example block, i.e., : In rdf/xml:...in rdf/n3:...
20:30:43  Actually, a switch woudl be better.
20:30:58  Have all the examples in whatever, and I can flip a switch and have the form I prefer show up...
20:31:02  Heh -- perhaps I should just write the spec in Smalltalk?
20:31:07  YES!
20:33:09  Heh, don't think that'll pass DCMI process. ;-)
20:38:21  Hmm. PhilG seemed to go from bashing Tcl for not having CL level numerics (correctly) to bashing Common Lisp for not having Tcl level string munging (falsely).
20:38:37  Hmm, pointer?
20:38:47  (to the second one)
20:38:56  Chapter 10 or 11 of ESBW. Second edition.
20:39:07  Er...DBBW :)
20:39:30  DBBW?
20:40:07  Database backed websites with random letters thrown in :)
20:40:19  # we have to trim the leading zeros because Tcl has such a
20:40:19  # brain damaged model of numbers and thinks "039" isn't a number
20:40:19  # this is when you kick yourself for not using Common Lisp
20:40:40  Yes, I know about that one -- I meant the second one (about Common Lisp not having string munging).
20:41:01  The new edition: http://www.arsdigita.com/books/panda/server-programming.html
20:41:27  Good old Gates-Clock.
20:41:27  Tcl can be better than Lisp because string manipulation is simpler. For example, in Tcl 
20:41:27  "posted by $email on $posting_date."
20:41:27  will generate a string from the fragments of static ASCII above plus the contents of the variables $email and $posting_date. These were presumably recently pulled from a relational database. The result might look something like 
20:41:27  "posted by philg@mit.edu on February 15, 1998."
20:41:27  In Common Lisp, you'd have 
20:41:29  (concatenate 'string "posted by " email " on " posting-date ".")
20:41:36  But that's just baloney.
20:41:41  Likely you'd have used format.
20:41:55  Which is only slighltly trickier for this case.
20:42:16  And for anything more complex likely to smack Tcl hard :)
20:42:35  Speaking as a former Tcl programmer, it's still no fun... I'm waiting until Python implements Ka-Pings $strings addition. ;-)
20:42:54  I say former so that Bijan won't hit me _too_ hard.
20:43:04  ;-)
20:43:18  It's like saying that in python you have to do 'posted by' + email + ' on ' +  posting-date '.'
20:43:33  Which many people *do* do, but c'mon.
20:43:49  Well, it's prettier than using % gobbledy gook
20:44:03  "posted by %s on %s" % (email, posting_date)
20:44:08  or the dict version.
20:44:15  But aaron, prettiness isn't the argument.
20:44:31  True
20:44:47  And it's only prettier, much of the time, for this really really trivial case.
20:44:51  Or interestingly prettier.
20:45:08  Format has all sorts of interesting stuff, including conditionals, etc.
20:45:24  field width stuff.
20:45:30  Etc. etc.
20:45:35  Hmm, I don't think it's that trivial, since people who write website scripts do it a lot. ;-)
20:46:01  But that's because you stuff a lot of formatting stuff elsewhere.
20:46:04  So this is just output.
20:46:32  And for the trivial case, which this is, there's only a marginal gain in Tcl.
20:46:50  Worthwhile? Sure, that's why python and CL have % and format :)
20:48:01  Not to mention that it's trivial in CL to write a simple macro that translates '("posted by" email "on" posting_date ".") into the right output.
20:48:23  I still want Ping's Itpl to be included in the standard Python distro.
20:48:44  Sure. Not that that's at all to the point, but...
20:49:11  I invoke my vast influence in the python relams to fullfill your wish.
20:49:31  Aww, thanks. :-)
20:52:42  It's still a bit of a turn around by philg co-inciding nicely with his greater interest^H^H^H^H^H^Hexperience with Tcl :)
20:53:54  Well, he's back to LISP now.
20:53:59  Heh.
20:54:19  If you're interested: http://www.xanalys.com/software_tools/reference/HyperSpec/Body/sec_22-3.html
20:54:51  Layout Control and Control-Flow Operations are particularly interesting, IMHO.
20:55:46  Neat.
20:55:59  * bijan freely confesses that format control strings look like gobbeydegook to him, but he's never written any.
20:56:00  Yes.
20:56:21  It's all very interesting. The example of following the english conventions for lists is especially cool.
20:56:34  At the bottom of: http://www.xanalys.com/software_tools/reference/HyperSpec/Body/sec_22-3-7-2.html
20:56:45  (format nil foo) =>   "Items: none."
20:56:45   (format nil foo 'foo) =>   "Items: FOO."
20:56:45   (format nil foo 'foo 'bar) =>   "Items: FOO and BAR."
20:56:46   (format nil foo 'foo 'bar 'baz) =>   "Items: FOO, BAR, and BAZ."
20:59:44  Hmm. It's even worse, IMHO. Since even if you ignore format, I'd wager that you *wouldn't* use strings for html generation.
20:59:52  You'd generate lists that wrote out as html.
21:00:16  And backquote, comma, and atcomma let you do the kind of splicing he's talking about.
21:14:19  danja has joined #rdfig
21:18:19  OK, http://logicerror.com/dcrdfDraft is looking pretty decent now
21:21:02  Comments? Comments?
21:21:04  Please?
21:21:32  rreck has quit
21:25:05  Aaron - I've not got time to read it properly, but a quick query anyway - what're the advantages over just using dc:blahblah in the usual fashion?
21:25:21  What's the usual fashion?
21:25:32  This document is supposed to define the use of dc:blahblah.
21:25:42  well, the proper use, anyway
21:25:54  hmm, some proper uses, perhaps. ;-)
21:31:18  sbp has joined #rdfig
21:34:33  Hi Sean.
21:34:44  Wanna take a look over http://logicerror.com/dcrdfDraft ?
21:34:49  Please?
21:40:26  * danja eek! a tiny scorpion just ran under bookcase
21:42:35  Ah sorry, it is the usual fashion, I was thrown by N3
21:43:02  N3 sent a tiny scorpion under your bookcase!
21:43:12  * bijan restricts or eliminates his use of n3
21:43:21  This is N-Triples -- not related. ;-)
21:43:28  must have been
21:43:40  That's why a *tinY* scorpion.
21:43:47  sorry again Aaron
21:43:54  God knows, if you'd used full fledged n3 it would have been HUGE!
21:44:00  When I use RDF/XML I am attacked by hordes of angry ants, so I opted to leave it out of this draft.
21:44:06  ;-)
21:44:37  so please correct it to RDF/n3 AaronSw
21:45:03  Correct what, dajobe?
21:45:12  the DC in RDF/n3 draft
21:45:13  * danja ok now, I jarred it and chucked it out
21:45:37  I'm not using N3, dajobe. I'm using N-Triples.
21:45:46  n-triples is a proper n3 subset, isn't it?
21:46:00  I mean, that n-triples statement *is* a valid n3 statement, isn't it?
21:46:09  Err, perhaps... I think there's some clashes on character escapes.
21:46:27  DC thing: Seems fine to me
21:46:28  But not in your example :)
21:46:30  n-triples is a proper subset. n3 may not define the \u and \U
21:46:34  actually, there is an inconsistency
21:46:37  recommend including RDF/XML equivs then
21:46:54  My examples are just fine.
21:46:55  N3 lets you use _:1 as an anon node (well, CWM does)
21:47:01  I don't think NTriples lets you
21:47:08  But that's why N3 is a superset. ;-)
21:47:13  yeah
21:47:52  that's a bit sucky though, because CWM does that on -ntriples output
21:48:02  so that would be a bug 
21:48:15  with a 1 line fix probably
21:48:16  yep
21:48:21  absolutely
21:48:34  so email DanC the fix
21:48:45  but... wouldn't it be easier to let name chars start with a number? why is there that restriction anyway?
21:49:00  too close to rdf:_1 ?
21:49:06  http://www.isacat.net/2001/09/books.htm
21:49:06  G: http://www.isacat.net/2001/09/books.htm from danja
21:49:09  oh, I guess it'
21:49:19  s for synonimity to some extent with XML names
21:49:33  G:| A short list of SW-related books
21:49:33  titled item G
21:50:04  dunno: perhaps it's something that needs to be discussed. It's a very small pedantic thing, but important to get right...
21:50:53  it isn't worth this much typing in irc; email the fix to Danc, lets move on
21:51:04  agreed
21:51:53  G:Heh! "Must-have book for lovers of techie autobiography - more readable than the source code to cwm."
21:51:53  commented item G
21:52:00  * AaronSw frets over Dublin Core draft until danbri gets back
21:52:03  G: with short (personal) reviews
21:52:03  commented item G
21:52:45  G: I wasn't going to post it yet, but last week I nearly bought the book bijan recommends below 
21:52:46  commented item G
21:53:23  G: suggestions welcome, as would be a scrape of the chump logs for books
21:53:23  commented item G
21:53:54  G:Note that I only somewhat endorse it, based on topics covered. Quality of the technical content is unknown to me.
21:53:55  commented item G
21:54:33  * danbri catches up
21:54:39  hey danbri!
21:54:42  Hey Aaron, it's looking good :)
21:54:44  Hi DanBri
21:54:55  Thanks!
21:55:04  I like it a lot, right now.
21:55:23  I like the technical content of the old doc, but I find it rather hard to wordsmith cos it really needs a rethink; yours will be easier for me to nitpic over...
21:55:35  You wrote all that today?
21:55:44  Yep. just now while you were out. Yes, i like the new one -- it's much leaner than the old one and it covers just about all the necessary material.
21:56:02  Can I start nitpicking now? ;-)
21:56:03  I stole most of the flow and examples from the old draft, though.
21:56:10  Yes, please begin the nitpicking!
21:56:44  The n3/ntriples might be troublesome for a more casual audience; while I'm happy saying its intended audience is 'RDF folk, some XMLers', ntriples/n3 is even more niche.
21:56:48  Or was until recently...
21:57:07  Hmm... I really don't want to scare people off with XML, though.
21:57:15  Post both.
21:57:16  XML is what some people *want* to see
21:57:21  s/list some of the basic/lists some of the basic
21:57:23  I said that from the start! :)
21:57:51  ta, sbp
21:57:53  "[DCNS]" should be in a  element
21:58:00  * AaronSw decides to go with Bijan's plan
21:58:26  XML: that's what I was getting at with "written in XML"... I think we need to accept the need for XML in this doc. 
21:58:27  Label XML not found.
21:58:34  (Also, it's useful in that it gets people familiar with the alternatives)
21:58:54  Hmm... it's difficult being nitpicky with your documents; too damn perfect
21:59:16  I'm happy having ntriples in there, for simple examples; the more hairy nits of n3 might be appendixes or separate files.
21:59:33  No, N3 stuff is in there.
21:59:35  Just triples.
21:59:49  I'll add RDF for the complicated ones.
22:00:19  I think it's good to have the XML serialization and the raw triples.
22:00:23  "<> in N-Triples" might be a bit confusing, seeing as how the actual term "<>" used to (still does?) represent the containing document in Notation3. You mean "terms surrounded/delimited with '<' and '>'"
22:00:34  (aside: I'm actually adding statement groups, or lets call them n3 contextx to redland now; this is an experiment)
22:00:49  dajobe: great!
22:00:51  I think it's acceptible to start using n-triples for the "raw" triples though I'm not super duper happy about it.
22:00:58  Actually <> isn't allowed in ntriples
22:01:00  this might be a bug
22:01:13  since I put character+ rather than character* - I think
22:01:14  Something that occured to me as an intereesting example: we could produce a chunk of XML that _didn't_ even contain the string 'dc' or 'dublincore.org' or 'purl.org/dc/elements' anywhere in it, ie. using 3rd party subProperties etc... And describe how the scheme stuff shows this can be processed as qualified Dublin Core.
22:01:29  dajobe: Well, N3 uses "this" now to refer to the context of the document
22:01:48  yes, but 'this' ahs no meaning for a graph description language; all URIs are absolute
22:01:50  (RSS, in fact, is such an example...)
22:02:34  well, TimBL seems to be using "#_formula" for the root context, which is pretty terrible if true...
22:02:39  * danbri deletes Redhat and re-installs SuSE 
22:02:51  go debian danbri - you'll like it :)
22:03:03  * AaronSw fixes bugs that were raised, updates draft
22:03:10  * AaronSw continues to add RDF/XML examples...
22:03:34  debian: thought you said the installer stank...
22:03:42  once it's installed though....
22:08:33  Are you done nitpicking?
22:31:04  * danbri leaves suse installing, takes another look
22:38:13  danja has quit
22:38:38  What is a Grecian URN?
22:39:18  Oh man...
22:39:29  whatsup?
22:39:53  :-)
22:52:29  danbri, haven't you received enough debian propaganda from w3c sysadmins yet? suse... tsk.
22:53:17  * ger-laptop added some debian propaganda to http://validator.w3.org/about.html the other day
22:54:05  I was thinking Debian, though had also collected enough views from others about it being a bit tough to get used to, rough diamond etc; SuSE looked interesting.
22:54:24  I wouldn't go with SuSE now, except I've invested a chunk of my weekend learning way around their installer.
22:55:06  in my experience, debian might take an extra day or two to get up and running in some cases, but you won't be sorry afterwards
22:55:09  Any installer that fails to find a kernel to install (ftp site structured wrong) and doesn't properly alert the user is pre-alpha, imho :(
22:55:50  I did my first few deb installs at the 'tute with experts nearby...
22:55:52  yes, I recommend debian for what danbri wants - easy to get new cool tools, daily
22:56:05  Well I've quite goofing around trying to preserve my old partition with an ancient redhat; that's wiped now, so I could debianise it pretty easily
22:58:08  gerald, what kind of things went wrong with your debian installs, first times around? 
22:58:43  or dave for that matter, did yours go wrong, any obvious things to avoid?
22:59:12  installing the stable, then deciding to goto unstable
22:59:18  hmm... good question; can't recall. (should've recorded it on fogo.) Maybe hardware detection stuff, and lack of knowledge what config files to tweak?
22:59:24  and tricky to get installing with reiserfs, but I fixed that
23:03:38  next time, I'm keeping a full swearLog...
23:04:12  * ger-laptop might have related bits in an irc log somewhere
23:04:40  that could be useful. Were you installing on a laptop or pc?
23:08:30  btw #suse isn't very useful; #debian by contrast looks pretty useful
23:09:41  I was installing on a pc, my friend's at home
23:10:56  got redland contexts partially working; added statement to context, listed context
23:14:21  cool. Does that work across in-mem and on disk versions? or do you have to re-implement for each storage form?
23:14:35  Rael will be pleased too :)
23:14:47  just done in-mem as first step,s o I can get test cases working nicely
23:14:57  reimpl: yes
23:15:42  right. makes sense. 
23:17:43  sbp has quit
23:18:53  * danbri puts the laptop down for a bit...
23:19:16  yeah, time to quit for a bit - g'nite all
23:19:20  dajobe has quit}}