07:52
<Ms2ger>
OH: "web browsers are such broken little puppies"
09:47
<Ms2ger>
odinho, jgraham, want to review the remaining .8% of /r/368?
12:02
<odinho>
Ms2ger: awesome work. was inspired to take a new look at the manual test. still, dome of my fixups are pendig afaik. I can't well review them. tho i'm fairly certain about them being correct fixes.
12:03
<odinho>
I'll look at remaining manual in a while if noone has beaten me to it.
13:22
<Ms2ger>
odinho, the remaining manual one is similar to ones you fixed up, I think
13:25
<odinho>
Ms2ger: It's just that it's manual :P The horror!
13:25
<odinho>
Lemme take a quick lookie.
13:26
<Ms2ger>
Heh
13:28
<odinho>
Ou, I need to go out xmas shopping. I can tag along with my sister who is leaving now, so I'll do that. l8r still :P
13:42
<Ms2ger>
Damn sister! :)
19:21
<Ms2ger>
jgraham, looks like I found some more typos :)
19:27
<matjas>
for those reading the logs, the guy hosting them has gone full ppk and is now doing a donation drive: http://krijnhoetmer.nl/irc-logs/
19:38
<zewt>
ffff, amazon is probably the single worst site i use regularly when it comes to browser back
19:38
<zewt>
it's like every single link on their page goes through a redirect that gets dumped in history
19:43
<jgraham>
Ms2ger: Damn you and your careful review ;)
19:44
<Ms2ger>
Worse, that was careless review :)
19:44
<odinho>
:]
19:48
<Ms2ger>
jgraham, and while you're around, https://critic.hoppipolla.co.uk/1e2ac7b7?review=368:)
19:48
<Ms2ger>
https://critic.hoppipolla.co.uk/1e2ac7b7?review=368 that is
19:51
<Ms2ger>
Hi again, irccloud
19:51
<krijnh>
lol
19:51
<krijnh>
fu matjas
19:52
<krijnh>
:)
21:05
<odinho>
annevk: Requesting review of changes because of your issues :-) https://critic.hoppipolla.co.uk/showcomment?chain=921
21:43
<Fernandos>
I have the feel that http://ogp.me/ is very inefficient and breaks existing standards. Is there an official answer to that from w3? I thought rdf would be it. But I'm not sure
21:44
<Fernandos>
This looks like pressing machine relevant information to the client's browser.
21:57
<zewt>
wondering why system with 16 gigs of memory is swapping and dying, and find that chrome was using 7.5 gigs of memory
21:58
<zewt>
: |
21:58
<Fernandos>
woot zewt.. that's crazy.. what did you do? run an infinite loop with webworkers, indexdb and websocket connection to a db that bombs your client with data?
21:58
<zewt>
"browsed the web"
22:01
<zewt>
now it's time to play "which restored tab is playing sound"
22:02
<gsnedders>
Ah, web browsers…
22:03
<gsnedders>
Fernandos: RDFa is the cloest you get to anything from the W3C trying to address arbitrary samantics in web content; there's also microdata in the WHATWG HTML spec.
22:10
<Fernandos>
Thanks gsnedders!
22:10
<Fernandos>
gsnedders: is RDFa different from RDF?
22:11
<gsnedders>
Fernandos: RDF is an abstract model of semantics. RDFa is a syntax for serializing RDF into.
22:11
<Fernandos>
ahh
22:11
<gsnedders>
Fernandos: IIRC, RDFa cannot represent all possible RDF triples.
22:12
<Fernandos>
And is RDF what you would use to transfer information between man and machine using a audio/video interface?
22:12
<gsnedders>
Fernandos: People sometimes use "RDF" when they mean "RDF/XML" (an XML based serialization of RDF) — but strictly RDF is just the abstract model.
22:13
<Fernandos>
Because to me it looks like it's only useful to debug, when the machine does something wrong. But not a very efficient way. RDF conversion consumes precious milliseconds of computing.
22:13
<gsnedders>
Fernandos: RDF is essentially just designed as a model to represent triples of (object, property, value), and represent semantics like that. I don't really know what you mean by audio/video interface. :)
22:14
<Fernandos>
I think RDF is only suited for multi-pass or slow indexation through search-engines, but not one-pass online algorithms like realtime man - machine interfaces, right?
22:14
<Fernandos>
audio/video interface? simple, I mean NLU (natural language understanding, gestures, mimics adds more precision to the input)
22:15
<gsnedders>
Fernandos: Most realtime interfaces you /probably/ just want to use something non-standard that matches your processing of the input.
22:15
<gsnedders>
Fernandos: Okay, that's what I was guessing
22:15
<Fernandos>
There is no standard for that I guess.
22:16
<Fernandos>
I mean it's possible to store the input after it was processed, with just more ram
22:17
<gsnedders>
Fernandos: RDF is designed to be extensible, so you can make up your own objects and properties, so insofar as that RDF isn't impractical. Just depends on the exact semantics of the message you're trying to pass.
22:18
<gsnedders>
Fernandos: If you want to pass something like "person 1 is pointing to the left", you can certainly do that, but RDF arguably isn't designed for such transient semantics. #swig is a better place to talk about RDF though, with far more knowledgable people than I.
22:18
<Fernandos>
there is no standard storage format facial expressions and gestures. I know about EmotionML only
22:18
<gsnedders>
(And I'm sure there are people in there who will argue RDF is fine for transcient semantics. :))
22:18
<Fernandos>
hahaha
22:20
<gsnedders>
Fernandos: Right, there may not be any pre-existing ontology for facial expressions and gestures in RDF, but that doesn't prevent you from using it. It is designed so that anyone can introduce new ontologies without them clashing.
22:20
<gsnedders>
(Essentially by using URIs for everything)
22:21
<gsnedders>
Fernandos: But without knowing a lot more about what you're doing, I've got no idea what's a good fit. :)
22:21
<Fernandos>
I cannot think about a semantic way of storing the data that is usable for any human.
22:21
<Fernandos>
Except images. hahaha
22:22
<Fernandos>
I do what I described :) The example was actually fit
22:22
<gsnedders>
If you're just doing that, you probably don't need something so generic.
22:23
<gsnedders>
Different people would advise radically different solutions. :)
22:23
<Fernandos>
The data can be stored already in a format that makes sense for the machine, but it's not semantic.
22:24
<Fernandos>
And I'm not sure how semantic storage would make sense, other than to talk to other machines, which could be done much better in binary anyway. As said, I think it can only serve for debugging, right?
22:54
<Fernandos>
np. think I'll find a way
22:54
<Fernandos>
see you
23:55
<Jasper>
Who thought of classList, and how can I give them the biggest high-five ever?
23:56
<Jasper>
I'm surprised it's not more well-documented or used.