Experimental IRC log swig-2008-01-22

Available formats: content-negotiated html turtle (see SIOC for the vocabulary)

Back to channel and daily index: content-negotiated html turtle

These logs are provided as an experiment in indexing discussions using IRCHub.py, Irc2RDF.hs, and SIOC.

00:04:01<DustWolf>so what is wrong with foaf + wot? :P
00:04:22<danbri>nothing intrinsically, except nobody can ever remember to use pgp
00:09:47<danbri>bengee-log-reader, i think http://www.somacon.com/p109.php might be my issue
00:10:03<danbri>oops, meant for #semsol
00:10:08<danbri>ah well, he reads here too
01:53:14<DanC>.t JST
01:53:16<phenny>Tue, 22 Jan 2008 10:53:14 JST
09:06:16<uws>edd: Yuck.
09:07:50<uws>edd: http://mail.gnome.org/archives/desktop-devel-list/2008-January/msg00189.html
09:07:53<uws>edd: Last paragraph :(
09:18:42<edd>Gotta love free software politics
09:19:04<drewp>the rest of the thread was quite positive, until people wanted their AUTHORS/MAINTAINER/etc files back
09:19:30<edd>One simple makefile rule would generate them
09:19:43<edd>Years ago I made doap.rdf automatically generated on a 'make dist'
09:20:13<edd>I suspect there's all sorts of personalities and grievances at work there.
09:20:21<drewp>you mean (human-edited) doap.rdf generated the other files?
09:20:51<drewp>this gnome attempt seemed to fail due to the difficulty in parsing MAINTAINERS files
09:21:18<edd>No, I didn't do the other files then. MAINTAINERS didn't exist back then
09:21:38<edd>I don't think there's anything hard in what they missed. Seemed more political to me.
09:22:21<edd>I'm not close to the project these days, but seems like Olav's quite fed up with the reaction he gets from maintainers, who he's trying to help
09:24:53<edd>uws: you wrote 'DOAP is not XML', which is true
09:25:05<edd>uws: however, I did intend for people to be able to write it as XML, if they didn't care about RDF
09:25:14<edd>hence the relaxng schema ,etc
09:25:25<EtnaRosso>morning all
09:26:18<danieljohnlewis>morning EtnaRosso
09:31:25<uws>edd: But that poses severe restrictions on the syntax used
09:32:36<uws>edd: furthermore, the complaints were that the XML looks like "crap" and that it's hard to write.
09:32:49<uws>I proposed a decent Turtle syntax, but people kept on bitching
09:59:02<edd>uws: well, maybe i should write a few AC macros to do it for them :0
10:00:30<uws>edd: heh :)
10:00:43<uws>edd: Do you know of any existing i18n approach for RDF-XML/Turtle?
10:01:06<uws>edd: intltool preferably :)
10:01:28<uws>it would be cool to translate DOAP Turtle files
10:01:41<uws>from somefile.doap.in to somefile.doap, by getting the names from .po files
10:04:19<edd>yeah. that would be nice
10:04:54<edd>gnome does do xml intl of some sort, hence the xml intltool, but doap needs the translations collating
10:21:47<uws>edd: what do you mean by "translatiosn collating" ?
10:22:18<uws>edd: The syntax in Turtle is really nice, "Some string"@nl
10:23:03<uws>so perhaps extending the syntax a tiny bit, e.g. "Some string"@ would mark it for translation
10:24:08<ldodds>it'd be possible to find literals that don't have a translation using a SPARQL query
10:28:53<edd>uws: i mean a doap file has all possible translations in it, collated.
10:29:04<edd>rather than individual files, localized for the environment
11:35:03<uws>edd: Yeah, sure. Using xml:lang in RDF-XML or @lang in Turtle. That's how the .desktop file translations work as well (and those are handled by intltool)
11:35:26<uws>edd: Same for intltool's XML-mode btw---it produces repeating elements with differing xml:lang values
11:35:46<uws>(for the record: .desktop files are INI-style)
11:39:11<edd>i proposed RDF .desktop files at GUADEC some years ago :)
11:39:22<edd>didn't fly, as you might imagine
11:42:42<uws>edd: What would you suggest as a RDF library for usage withing the Gnome/Gtk/Glib world?
11:42:46<uws>librdf/libraptor?
11:42:52<uws>(hi dajobe)
11:43:57<edd>yes
12:54:41<danbri>hmm rss 1.0 from feedburner isn't rdf/xml :(
12:54:54<danbri>eg. http://feeds.feedburner.com/ontogame
12:55:18<danbri>the options for getting a feed into planetrdf.com if it doesn't natively have rss 1 are shrinking further...
12:56:01<jaa>RDF, dying format
12:57:27<jaa>Raptor found 22 riples for me
12:57:53<jaa>some proably hella-old version since nobody in gentoo uses it or updates the ebuild
12:59:02<jaa>e/planet/posts/%/blank/2/r1200999455r2878r125/%/http:/www.w3.org/1999/02/22-rdf-syntax-ns#_8/%/tag:blogger.com,1999:blog-5424232062276031101.post-6026369820026947582
12:59:07<jaa>TAG URIs.. people use those things?
13:05:52<jaa>curl -H "Accept: application/atom+xml" http://feeds.feedburner.com/ontogame
13:05:58<jaa>still gives broken pseudo-rdfxml-rss1.0
13:05:59<jaa>hrmm
13:06:07<jaa>how do you coax atom out of feedburner
13:06:33<jaa>raptor does weird #_24 shit
13:06:46<jaa>on the openlink dataspace feeds
13:06:50<jaa>ACTION should upgrade to SVN
13:10:57<jaa>does that mean all of blogger/feedburner is generating bad rss1.0 btw?
13:11:04<jaa>thats a pretty sizable chunk of all blogs.
13:11:37<jaa>im thinking of just breaking the feed normalizer out of planetmars/venus into its own pipeable shellscript
13:11:53<jaa>then make a one time atom -> rdf transform and let them worry abotu all the broekn crap out there
13:13:53<jaa>librdf warning URI http://feeds.engadget.com/weblogsinc/engadget:2 column 213 - Using node element 'rss' without a namespace is forbidden.
13:13:56<jaa>heh. totla fail!
13:14:11<jaa>i thought raptor was supposed to swill rss2 and atom
13:16:54<edd>don't you need to pass some command line option to do that/
13:19:11<jaa>ok yeah, turns out 'raptor' means rdf/xml in the ruby bindings, i thoguht raptor just was generic/guess
13:19:14<jaa>ACTION retries
13:21:29<jaa>nice
13:21:34<jaa>375 triples for engadget feed
13:21:51<jaa>yeah why cant planetrdf use guess then?
13:22:35<jaa> Redland::Parser.new(({'.nt'=>'ntriples','.n3'=>'n3','.ttl'=>'turtle'}.i(File.extname(uri))||'guess'),'').parse_as_stream(Redland::Uri.new(uri.match(/^(file|http)/) ? uri : 'file:'+d )) do |s|
13:22:41<jaa>seems to do the trikc
13:23:36<jaa>87 for ontogame
13:24:35<jaa>roundtripes perfectly back out w/ the same queries used for RSS1 schema stuff
13:24:37<jaa>nice work dajobe ;)
14:39:46<jah>graphpipes is a visual tool for the web to create complex SPARQL-queries and reuse them later in Networked Graphs.
14:39:54<jah>2 public empty git repos and a .de and no docs or code anywhere
14:40:07<jah> http://graphpipes.de
15:04:29<TipTop>how to manage a rdf:Seq in Swoop?
15:10:11<yvesr>i just updated the linked data last.fm service - it now provides you with foaf:knows towards your last.fm friends (see http://blog.dbtune.org/post/2008/01/22/Pushing-your-LastFM-friends-in-the-FOAF-O-Sphere)
15:12:35<TipTop>can anyone give me an example of the usage of an rdf:List
15:12:38<TipTop>i am a bit lost
15:13:48<TipTop>can't you just list the elements of the list? or is it mandatory to used rdf:first and rdf:rest ?
15:13:55<TipTop>s/used/use/
15:17:44<swh>TipTop: you mean the syntax? in N3/turtle you can write (a, b, c) etc.
15:18:08<TipTop>in rdf/xml it looks pretty difficult
15:19:46<swh>yeah, I never write anything that complicated in rdf/xml myself
15:20:13<swh>you could write it in something else and translate it with rapper or similar
15:20:17<TipTop>sure
15:20:27<TipTop>in fact, i will use a rdf:Seq
15:20:39<TipTop>but this is the first time i meet rdf:List in rdf/xml syntax
15:20:40<swh>those are easier - easier to query in SPARQL too
15:24:08<TipTop>damn, i have problems with rdf:seq syntax too (in rdf/xml again :)
15:31:36<swh>rdf:Seq should be easy
15:34:45<TipTop>swh: the problem is with elmo
15:35:01<TipTop>it should map rdf:Seq to java.util.List
15:35:04<TipTop>but it does not
15:36:09<TipTop>ACTION is unsure of the uppercase/lowercase of rdf:seq
15:36:32<yvesr>TipTop: there are some examples of rdf:Seq here: http://musicbrainz.org/MM/mm_examples.html
15:36:39<swh>Seq
15:37:14<TipTop>i think my syntax is correct
15:37:21<TipTop>but elmo still conplains
15:37:29<swh>hi yvesr, we met at a rdf london pub meet I think
15:37:46<dajobe>joh, jaa has gone but 'raptor' parser is a legacy name for 'rdfxml'
15:38:18<TipTop>and my greater concern is the lack of support of rdf:Seq by swoop
15:40:04<yvesr>hi swh! yes, exactly, last summer at these linking-open-data drinks :)
15:40:35<yvesr>we should do that again, btw (although i think the semantic barcamp will be a good opportunity for beers :) )
15:42:58<uws>edd: http://svn.usefulinc.com/svn/repos/trunk/doap/examples linked from http://usefulinc.com/doap/ gives a 404
15:43:38<TipTop>noone uses swoop here?
15:44:19<swh>yvesr: yes, I was planning to go to that, but may only go to one day, or none
15:44:45<TipTop>what is that semantic barcamp?
15:45:03<swh>dajobe: did you see a bug report from one of our guys about some roundtripping problem in raptors turtle parser?
15:45:19<swh>dajobe: I didn't actually understand the bug myself
15:45:26<dajobe>swh: I don't recall that one.
15:45:38<swh>it was in email... I'll find a reference
15:45:46<edd>uws: my SVN server is broken :(
15:46:05<edd>ACTION just needs a spare half day
15:47:13<swh>dajobe: 14 January 2008 22:51:42 GMT, "Turtle definition vs implementation"
15:47:47<dajobe>oh that one. I ignored it due to the rather rude attitude
15:48:08<swh>yeah, that guy is a bit abrupt, sorry
15:48:37<uws>dajobe: Are the python bindings to librdf/libraptor cover the C API completely?
15:48:38<dajobe>timbl and I did agree \> was a spec mistake
15:48:46<uws>s/Are/Do/
15:49:01<dajobe>uws: no but they cover the parts I thought made sense
15:49:04<yvesr>TipTop: http://barcamp.org/SemanticCampLondon
15:49:07<danbri>who is coming to sem barcam here?
15:49:10<danbri>i am :)
15:49:15<yvesr>i am too :)
15:49:16<danbri>the london one i mean
15:49:27<uws>dajobe: Things like parsing multiple DOAP file to an in-memory store and getting some useful data out of it is possible?
15:49:27<danbri>frenchies are doing a parallel event i believe, we shoudl have eurovision-style linkup
15:49:36<yvesr>heh:)
15:49:38<uws>dajobe: (I'd like to know before eventually digging into it)
15:49:40<swh>dajobe: the thing that was biting us is the whitespace problem
15:49:40<dajobe>uws: sure, that's no problem
15:49:46<TipTop>ACTION is going to the Paris event
15:49:54<duck1123>does anyone here use the Flickr FOAF exporter?
15:50:05<uws>dajobe: Great, thanks.
15:50:05<yvesr>yes, i did hesitate to go to both, but i am only going to the london one
15:50:28<TipTop>i am afraid the date is not the same in london and paris :'(
15:50:33<dajobe>swh: I'll look at that part, just for you :)
15:50:34<duck1123>I noticed it's using the same uri for the document, and the foaf:Person
15:50:43<swh>dajobe: thanks dave :)
16:00:39<bijan>tiptop...swoop issues?
16:01:48<bijan>also, parseType=Collection makes rdf:Lists pretty easy to write in rdf/xml
16:01:53<bijan>so long as you don't sue literals
16:09:37<duck1123>yvesr: when I the last.fm link into my FOAF file, tabulator is showing it as "Is user of" -> "Listened to..."
16:09:45<duck1123>Is that how it should be?
16:10:12<duck1123>http://kronkltd.net/data/foaf.rdf
16:13:08<danbri>not sure how it should be!
16:13:20<danbri>sorry can't chat much now, am having mail probs i need to debug urgently :(
16:21:58<TipTop>bijan: hello bijan
16:22:27<TipTop>bijan: i try to define a property to point at a sequence of resources
16:23:23<bijan>Ok
16:23:25<TipTop>bijan: at the moment, it is serialized in rdf/xml as separate XML tags, i need then to become a rdf:Seq with several rdf:li
16:23:33<bijan>ACTION 's usual line on that is "don't do that" :)
16:24:01<TipTop>what's the bad idea in my plan?
16:24:54<bijan>I generally think containers and collections are bad ideas in rdf (and moreso in OWL)
16:24:56<bijan>But that's me
16:25:19<TipTop>i can understand that, and honestly, i already see the flow in using that solution
16:25:26<TipTop>flaw
16:26:38<TipTop>but elmo has a very good support of rdf:Seq->java.util.List mapping
16:26:48<TipTop>and that is convenient
16:27:25<TipTop>is your disliking of collections related to OWL reasoning theories, or not at all?
16:27:51<bijan>A bit
16:28:10<bijan>Containers are a mess on many levels
16:28:17<bijan>No semantics, really
16:28:22<bijan>numbered properties?!?
16:28:25<bijan>yeek
16:28:42<bijan>Lists were invented to help encode daml+oil syntax
16:29:08<bijan>And if youwant to do "real" list reasoning, they are just wrong (on a number of levels)
16:29:13<LeeF>yeek++
16:29:23<bijan>if you are using them to encode a data structure, that works ok
16:29:38<bijan>But hten I prefer to just *use* a data structury thing like xml
16:30:58<TipTop>XML has good support for lists
16:31:05<TipTop>but it has an awful semantics
16:31:14<bijan>Uhm...not really
16:31:16<bijan>esp. for lists
16:31:20<TipTop>yep
16:31:21<bijan>esp.c ompared to the rdf semantics
16:31:25<TipTop>but for anything else, i mean
16:31:36<TipTop>superclass/subclass for example
16:31:37<bijan>That's not so clear to me
16:31:49<bijan>Oh, xml schema
16:31:57<bijan>That's usually not what i want for that, i agree
16:32:14<TipTop>may be data as XML and schema a OWL
16:32:34<bijan>Well, that's not quite how I look at it
16:32:41<bijan>SOme data is fine in rdf/owl aboxes
16:32:43<bijan>No worries
16:32:48<TipTop>:)
16:32:51<TipTop>i think so too
16:33:14<bijan>Things like lists which require well founded structures and transitive closure don't fair as well
16:33:29<bijan>But you don't always need all that
16:33:49<bijan>But then you might as well construct your own aggregation pattern
16:33:59<bijan>E.g., "contains"
16:34:05<TipTop>i see
16:34:06<bijan>as a property
16:34:12<bijan>foo contains bar, baz, bat
16:34:16<TipTop>and then use some kind of custom inferencing?
16:34:23<bijan>You can use owl
16:34:33<bijan>Owl works well for those sorts of object
16:34:35<bijan>poorly for lists
16:34:35<TipTop>for transitivity you mean?
16:34:46<bijan>For transitivity, for typing
16:34:51<TipTop>yep
16:35:03<bijan>I can easily say foo contains at least one element that has a child that has a tree
16:35:24<bijan>OWL is used for anatomy quite a bit
16:35:31<bijan>and configuration
16:35:37<TipTop>wow, isn't it called second order logic?
16:35:48<bijan>No
16:35:51<bijan>that's all first order
16:35:54<TipTop>k
16:36:18<TipTop>oh yep, i see how to express that in owl
16:36:27<bijan>foo rdf:type [onProperty contains; someValuesFrom [onProperty...
16:36:28<bijan>etc.
16:36:45<bijan>The main problem is that existentials occur in tree like patterns only
16:36:52<bijan>(with some trickery)
16:37:08<bijan>Boris motik et al just developed a trick to allow certain sorts of graph patterns in owl tboxes
16:37:18<TipTop>that was my next question
16:37:23<TipTop>what's the status of hermit?
16:37:40<bijan>Which will be very nice for all sorts of patterns
16:37:49<bijan>Quite a ways along I'm given to understand
16:37:55<bijan>Will cover shroiq
16:38:09<bijan>Plus loads of stuff
16:38:14<bijan>Boris is a machine :)
16:38:15<TipTop>which is?
16:38:17<TipTop>good?
16:38:27<bijan>SROIQ is OWL 1.1
16:38:34<TipTop>wow, cool
16:38:34<bijan>So owl dl + a buncha stuff
16:39:49<TipTop>unfortunately, hermit is currently running on top of kaon2
16:40:05<TipTop>you think you will integrate those algorithms in pellet?
16:40:30<bijan>It's not running on top of kaon2
16:40:34<bijan>it used the kaon2 api
16:40:39<bijan>but that's goign to be merged with the owl api
16:40:58<bijan>As for pellet, we definitely are going to grab as much as we can :)
16:41:03<TipTop>cool
16:41:09<bijan>subject to resource constraints per usual
16:41:27<TipTop>owl api is probably cooler but jena is still the main api used today
16:41:43<bijan>I've no real sense
16:41:49<TipTop>?
16:41:57<bijan>I mean, jena has lots of RDF users
16:42:06<bijan>But I don't know how many people are using it for OWL
16:42:08<TipTop>owl in jena is a pain
16:42:13<TipTop>but it workd
16:42:13<bijan>Exactly :)
16:42:15<TipTop>works
16:42:21<bijan>TopBraid composer uses it
16:42:27<bijan>but prot4 uses owl api..
16:42:28<bijan>so
16:42:31<TipTop>yep
16:42:41<TipTop>but look at the status of both IDE
16:42:52<TipTop>and tell me which you would prefer for industrial stuff
16:42:58<bijan>Protege4
16:43:05<TipTop>tss tss
16:43:09<TipTop>the correct answer was swoop
16:43:14<TipTop>(of course)
16:43:26<bijan>I would take protege4 and skin it with swoop
16:43:33<bijan>The swoop internals were crappy
16:43:37<TipTop>nice idea
16:43:48<bijan>matthew had a prototype
16:43:51<TipTop>but ergonomy is wayyyy better that protege3
16:43:54<bijan>we'll get around to it at somee point
16:44:23<TipTop>were you the one who blogged the 5 semantic web tools ?
16:44:39<bijan>that was kendall
16:44:40<TipTop>a blog entry listing the tools to use for semantic web (a few years ago) ?
16:45:08<bijan>yep that was kendall at the c&p blog
16:45:14<TipTop>when i first read about swoop over protege, i dismissed the whole blog entry
16:45:25<TipTop>but then i had a look at swoop
16:45:27<bijan>See: http://clarkparsia.com/weblog/2005/12/28/5-best-semweb-tools/
16:45:29<TipTop>and never looked back
16:45:39<bijan>Why thanks
16:45:52<bijan>Protege 4 is much better than prot3
16:45:56<bijan>completely redone
16:46:05<bijan>has some swoop features like hyperlinks
16:46:11<TipTop>cool
16:46:20<bijan>yeah, doesn't make me cry :)
16:46:21<TipTop>i am afraid to try topBraid
16:46:35<TipTop>bcause of that protege3 GUI legacy
16:46:42<TipTop>may be they improved it
16:46:56<bijan>My understanding is that they did better but I've not worked with it
16:47:14<bijan>anyway,/me heading home.
16:47:55<duck1123>I like TopBraid, but I don't know what I'm going to do when my 30 days run out
16:50:05<TipTop>duck1123: does it still look like protege-owl 3 ?
16:50:43<duck1123>not really
16:51:01<TipTop>fine, then
16:51:08<duck1123>imo, topbraid looks better
16:51:47<duck1123>also, I use Eclipse anyway, so the eclipse integration is really good for me
16:52:21<TipTop>k
16:52:33<TipTop>that is one good point
16:52:54<duck1123>do any of the other editors allow you to edit the triples in a jena store?
16:53:58<duck1123>I must have downloaded every single editor I could find last week, and Topbraid was the only one I could find that did what I wanted
16:57:04<TipTop>direct connection to a database you mean?
16:57:15<TipTop>and live update?
16:57:19<duck1123>yeah
16:57:24<TipTop>i dunno
16:57:28<TipTop>i coded that myself
16:57:41<TipTop>but i have no idea about standard editors
16:58:08<TipTop>the jena store is accessed through what kind of protocol?
16:58:23<duck1123>I'm just getting started in my RDF career, so a lot of this is new to me
16:58:34<TipTop>this is a perfectly valid requiremen
16:58:36<TipTop>t
16:59:19<duck1123>what do you mean? My store is in a MySQL database, my java code and topbraid access it directly, I think
16:59:45<TipTop>k
16:59:48<jive>http://blog.whats-your.name/post/2008/01/22/element%3A%3A-planet
17:00:01<jive>B: mucking about with 'planets' in ruby
17:00:27<TipTop>was mysql on another computer?
17:00:54<duck1123>no, it's all hosted on my computer atm
17:02:05<TipTop>k
17:08:56<TipTop>as soon as the update features of SPARQL are implemented, we will have generic editor for remote triple store, i think
17:09:23<TipTop>anyone knows the status of implementations yet?
17:17:49<BenO>bengee: Can I ask a small q. about the Image Region vocab (docs at http://www.bnowack.de/w3photo/pages/image_doc_rdfsowl) - why is the #boundingBox coords defined differently to #Rectangle?
19:18:39<bengee>phenny, tell BenO I don't remember 100%, but boundingBox is a convenience method, and I think there were requests to use width,height instead of x2,y2
19:18:41<phenny>bengee: I'll pass that on when beno is around.
19:33:59<phurl>hi all
19:34:12<Shepard>hi
19:34:17<phurl>hi Shepard
19:35:51<lordi>does anyone have a hint on how to get the "most specific" property of a property assignment in pellet?
20:32:06<duck1123>does anyone here know if there is a way to get firefox to display text/turtle as text/plain (in the browser)
20:43:46<CaptSolo>anyone here who has got Redland python bindings to compile under OS X?
20:43:54<CaptSolo>ACTION trying to do that but is getting errors
20:44:30<kasei>what sort of errors?
20:44:50<CaptSolo>gcc -g -O2 -I/usr/local/include -DSWIG_PYTHON_SILENT_MEMLEAK -g -O2 -Wl,-F. -Wl,-F. -bundle -framework Python Redland_wrap.so `redland-config --libs` -o Redland.so
20:44:53<CaptSolo>/usr/libexec/gcc/i686-apple-darwin8/4.0.1/ld: Undefined symbols:
20:44:59<CaptSolo>_raptor_locator_byte
20:45:01<CaptSolo>...
20:45:19<CaptSolo>redland libraries are in /usr/local/lib
20:47:07<mortenf>do you have a preinstalled (old) raptor somewhere?
20:47:21<mortenf>could be a raptor version mismatch
20:47:48<CaptSolo>no
20:49:29<mahound>hello
20:50:00<kasei>not sure, but might be a 32/64 bit confusion between the libraries... i know this has bitten some perl code on the new os x.
20:50:08<CaptSolo>hm, redland-config --libs returns
20:50:30<CaptSolo>-L/usr/local/lib -lrdf
20:51:00<CaptSolo>since there is no -lraptor -lrasqal the linker does not know it needs those libs
20:51:15<CaptSolo>now, how to feed these libs to make?
20:52:21<mahound>is there any d2rq-like approach to non-relational databases, like OODBMS ?
20:53:38<kasei>there's been some work on putting sparql front-end on ldap. not sure about oodbms.
20:54:00<LeeF>duck1123, I use the Open in Browser FF extension to do precisely that
20:59:18<duck1123>leef, thanks I'll check it out. I'm getting sick of reading Triplr's output in Notepad++
21:00:21<mahound>kasei: weird... it though someone would have already worked on that
21:00:36<mahound>i thought
21:01:36<mahound>but it seems people assume that !rdf == relational
21:02:38<kasei>there might be work like what you want... i just thought i'd mention the ldap thing (since it matched your "non-relational" question)
21:06:47<CaptSolo>ok, works now
21:06:55<CaptSolo>(or at least builds and installs)
23:29:49<mahound>if any of the rdflib users can help me with this one... http://groups.google.com/group/rdflib-dev/browse_thread/thread/cb8cb29a890d34a3

Back to channel and daily index: content-negotiated html turtle