This is a public chat log generated from the #semsol IRC channel.
11:14:56
bengee updates ARCs RDFa extractor
11:15:32
will try to comment on process section while I'm walking through it
11:19:14
* nitpick: why is it longer than the one from the previous spec, I would've expected a cleaned-up/simpler final process
11:19:20
* step 9 is missing
11:27:42
* CURIE is not a valid abbreviation for "compact URI", should be cURI, or CURI, no? or do I need Marie Curie capabilities to spot the E? ;)
11:29:55
kwijibo reads on in amusement with bengee's nit picking
11:35:08
* steps 1-6 seem fine to me, easy to follow
11:38:50
* step 7 seems to require a [new subject] in order to create a triple, this is not mentioned in the intro sentence (which is different from step 6, which explicitly says "none of this ... if there is no [new subject]")
11:41:36
* step 8: fine to me
11:42:38
* step 10: understood, I guess it's identical to the previous spec version
11:43:34
* hmm, "once the triple has been created". There can be multiple, so maybe s/the/a/
11:44:38
* i.e. *any* XMLLiteral object stops recursion (as I understand it)
11:46:08
* step 11: "using the rules described here". What exactly does *here* refer to? step 11, or the whole process sections
11:50:43
* hmm, the distinction between the passed-in context, the current context, and local variables is a bit confusing, I just used the passed in context and overwrote it's values while I was makinf my way through the process, that worked fine in the prev. spec, but is apparently not correct now
11:51:09
* tweak, teak, tweak
11:51:22
* ok
11:51:50
* keeping a copy of the passed-in context
11:56:26
* hmm, I fail to grok the text in the blue box in step 11
11:58:19
* the final step (step 12, below) involves returning a flag. If the flag is true, then incomplete triples are completed in the next step (step 11)
11:58:59
* I *am* in step 11, s/next/this/ in he last sentence
12:00:00
* ok, as I understand it, I have to proceed with step 12 for each child node first, and see if I can set flag
12:01:24
* err, no. step 12 expects a flag already
12:01:31
wow bengee you're working through it quickly :)
12:01:53
* something is wrong with the step numbers
12:02:06
maybe too quick
12:07:56
* ok, I'll try to read it as "text correct, numbers wrong", then 12 is the final step, and step 11 is actually step 10 and can point at step 11 as "next step"
12:21:42
bengee is still confused
12:23:30
* "after having recursed into the processing of descendants"
12:27:06
* "forwarded to step 11", not sure PHP understands GOTO 11
12:27:44
ok, more seriously now again
12:28:30
* I loop through and process the subnodes
12:28:40
bengee: where does it say about the < 1000 lines for semantic scripting challenge?
12:28:46
* a process returns a floag
12:28:53
s/floag/flag/
12:28:57
lemme see
12:29:02
kwijibo wonders if he can get to go
12:29:39
shame it's not in innsbruck again - that looks a much more fun location!
12:30:04
ah I see it bengee
12:30:18
'we define "small" ... less than 1000'
12:30:28
how do they judge what's an external library?
12:30:40
or ... you - are you a judge?
12:30:42
that's the ultimate backdoor
12:30:52
lol
12:31:08
just place everything into a separate file and call it a lib
12:31:13
you could just make it one line
12:31:15
yeah:
12:31:19
right
12:31:20
here's my app:
12:31:33
require('scripting-challenge.php');
12:31:39
lol
12:31:41
:)
12:34:12
The application must be implemented using a scripting language and Semantic Web technologies. It should access, visualize or combine information from more than one source in novel or unforeseen ways.
12:34:15
wow :)
12:34:26
I'll use the IRCBot class as an external lib and submit sparqlbot. then I'll use sparqlbot as an external lib, add a widget generator, and submit that, too. then I'll use the widget generator and create an ajax portal from it, submit it as well, etc
12:35:05
lol
12:35:13
exactly!
12:36:37
I was thinking of doing a js forms designer for RDF
12:36:45
cool
12:36:53
doesn't meet the criteria though
12:37:07
oh actually it kind of does - or could
12:37:34
accept user input, and use schema-cache to retrieve stuff for drop down menus etc
12:37:48
if user input counts as a source
12:38:30
it would be built on a js RDF library I'm writing, which could be submitted too I suppose
12:39:09
and that depends on my convert service, but that's more than 1k lines
13:13:50
* ok, I've done step 13 (12) now, that was easy
13:14:35
* now back to 12 (11)
13:15:42
* ok, this is all about incomplete triples in the current context
13:17:12
* that means step 12 is not done as a loop
13:17:36
* but after processing the sub-nodes, which might have set the flag
13:18:15
* ok, I think the penny dropped
13:21:38
* the "not the local list of incomplete triples" hint is helpful, I would've been confused again w/o it
13:28:17
* in step 12, everything from the 3rd block sentence ("Note that if [new subject] is a bnode, then ... during this step") should be moved to a separate block. The first 2 sentences tell what I should do, I had the impression I had to add a bnode check for new subject here.
13:33:31
* "if direction is not 'forward'": are there any other possible values than "reverse"?
13:34:15
* ok, I think I'm done. time for the test suite
13:37:53
* ok, minor bugs, "parent object" is part of the eval context, not the local values
13:38:28
* the spec more or less mentioned that, my fault
13:42:14
* adding an empty "incomplete_triples" var to the initial context, to get rid of an explicit check at step 12 on the root node
13:43:24
* ok, test 1: passed w/o php notices
13:43:35
bengee: if I just do extractRDF from the semhtml parser, would that invoke the rdfa parser as well?
13:43:56
(is the current rdfa parser working at all?)
13:45:27
it depends on you config settings, RDFa is not part of the default extractors
13:45:34
ah ok
13:45:46
you can do extractRDF('rdfa') though
13:45:53
(I think)
13:46:06
will rdfa be part of the default set?
13:46:22
or is that a design decision?
13:46:44
it generated unwanted triples
13:47:00
yeah?
13:47:22
can rdfa be made not to do that?
13:47:31
so, once the parser is fixed (which might be now, I'm running the tests) it'll get re-added to the default extractors
13:48:01
is the unwanted triples a bug with the parser or the spec?
13:50:32
my fault, I think
13:51:46
coincidentally bengee, edsu in #talis was just asking about rdfa support in my Convert service (which of course uses ARC for all the RDF parsing stuff )
13:52:57
* 5 tests passed, /me gets dauntless and clicks on "run all tests"
13:54:04
hmm, it fails on a lot of "onhold" ones
13:54:20
and also approved ones...
14:02:47
hmm, pass/fail: 40/32
15:38:25
* nope, the processing rules don't generate the triples they should
15:38:34
* or I fail to find the bug
15:38:53
damn
15:38:58
bengee turns on context dumping
15:49:36
* ok, the "skip" flag prevents the uir set by @resource to be propagated to the child nodes, hmm
15:49:43
s/uir/uri/
15:50:17
* where is that set?
15:51:05
* whenever there is a parent_object...
15:53:16
* aha, c&p error, skip is only set to true in step4, not in step5!
15:53:30
* for whatever reason, though ;)
15:53:47
* PASS!
15:55:07
bengee re-runs test suite
15:56:17
ok, pass/fail improved from 40/32 to 48/15
15:57:02
minus 9 skipped "onhold" tests
15:58:37
* ok, that seems to be related to "incomplete triples"
16:34:24
hey bengee: I meant to ask you about named graph support in the rdf/json structure
16:34:36
ugh
16:34:47
the least intrusive place to put it is in a key in the object
16:35:13
ah, right
16:35:23
{graph: '_:465', value:'x', 'type': literal}
16:35:33
it's also the place where the triple gets "completed"
16:35:39
yeah
16:35:47
would it be usable in that position?
16:35:57
maybe graph*s*
16:36:09
graph*s*?
16:36:19
if a triple is in more than one graph?
16:36:20
a triple can occur in multiple graphs
16:36:23
y
16:36:39
ah, didn't occur to me :)
16:37:28
what I'm wondering is what kind of manipulations do you want to do with named graphs in scripting, and is that solution sufficient for them?
16:38:42
could be handy for triple editing
16:40:00
don't see what you mean?
16:40:25
update 0 where s= &p= &o= &g=
16:40:37
update o rather
16:41:09
w/o the g info, you'd update all graphs where s/p/o matches
16:41:31
(which is what I'm doing in knowee/grawiki at the moment)
16:41:51
ah
16:42:05
well, I could add it, but haven't for efficiency reasons
16:42:11
that's just pure information though
16:42:17
so this is not necessarily related to rdf/json
16:42:27
yeah
16:43:52
can you think of anything you would use the graph id for if you were just iterating over the data in a template or something?
16:44:10
hmm, not sure
16:44:20
i can't really
16:44:46
but I'm not using named graphs, so not in a good position to say what's useful :(
16:44:49
provenance/trust info
16:45:22
at the bottom of a page: this page is based on info from a, b, c, d,
16:45:43
yeah, object position is good for that then
16:46:13
otoh, if you wanted to break it into boxes, X says: and Y says: top level would be better
16:46:32
kwijibo trying to imagine other use cases
16:47:14
* current processing rules do not seem to cover test0033
16:55:40
* ok, let's make the processing rules a bit more tolerant
17:02:31
* ok, looks like we're getting somewhere
17:03:31
* pass/fail: 52/11
17:04:42
* ok, no support for reserved xhtml-rels/revs yet
17:13:50
* another run
17:14:55
* pass/fail: 58/5
17:15:43
* ok, : seems to be a default prefix
17:16:40
* for 'xhtml + #'
17:16:51
* where does that come from?
17:18:02
* aha CURI(E) Syntax Def
17:23:02
* ok, next run..
17:26:22
* pass/fail: 59/4
17:26:31
ooh
17:26:35
getting close :)
17:34:48
* hm, the <base></base> container tag sems to confuse ARC's tagsoup parser
17:35:23
is that in XHTML 1.1?
17:35:37
dunno
17:40:13
* tweaking my tagsoup parser
17:42:23
* and yet another run
17:45:22
bengee reads Talis Platform News
17:45:37
* pass/fail 63/0
17:45:42
* yay!
17:46:26
awesome!
17:46:31
congrats bengee
17:48:17
btw, there's a new talis dev blog: http://blogs.talis.com/n2/
17:48:19
* making the processing rules in step 12 a tiny bit more restrictive (to see where exactly the spec might need tweaking)
17:48:32
ah, cool
17:48:52
ARC is plugged at least twice on it already ;)
17:49:14
heh, yeah. super cool
17:49:28
and always with "excellent" :)
17:49:48
oops yeah, a bit of unconscious plagiarism there maybe :)
17:50:09
- it seems the natural adjective ;)
17:50:52
* nope: pass/fail 60/3
17:53:08
* step 12: "If the [skip element] flag is 'false' ...": this condition should be removed
17:59:44
* final pass/fail (including "onhold" tests): 67/5, with all failed tests being "onhold"
18:08:30
bengee calls it a day
18:08:33
cya