#microformats 2015-07-27

2015-07-27 UTC
fuzzyhorns, KartikPrabhu, tantek and KevinMarks_ joined the channel
#
@AllTheTwits has been sorting through Microformats-discuss emails. (twitter.com/_/status/625532799047630848)
KartikPrabhu joined the channel
#
tantek those were the days
KartikPrabhu, tantek and KevinMarks__ joined the channel
#
@shopl RT @ProjectPeachUK: We've #played with #microformats. Love the #idea of #marking up our #business data to #machines as well as #humans! #sb… (twitter.com/_/status/625562361257197568)
csarven and eschnou joined the channel
#
KevinMarks twas brian: https://twitter.com/briansuda/status/881493
#
@briansuda has been sorting through Microformats-discuss emails. (twitter.com/_/status/881493)
KartikPrabhu, eschnou, adactio, JonathanNeal_, twisted`, Left_Turn and TallTed joined the channel
#
@nodenpm microformat-shiv (1.0.0): https://www.npmjs.com/package/microformat-shiv A cross browser JavaScript Microformats 2 parser (twitter.com/_/status/625698814142361600)
KartikPrabhu joined the channel
#
@JasonWeathersby RT @dietrich: Prototyped a Microformats panel for @FirefoxDevTools today. https://twitter.com/dietrich/status/623627610619277312/photo/1 (twitter.com/_/status/625703618495295488)
KevinMarks_ joined the channel
#
rhiaro aaronpk: I wondered the same
#
aaronpk so the question i had was why are the prefixes listed under the vocabularies?
#
aaronpk e.g. http://microformats.org/wiki/microformats2#h-adr h-adr lists "p-locality" as one of the properties, but in reality, the parsed result shows just "locality" and can have a value there by one of many different prefixes
#
rhiaro I think that having prefixes listed under the vocabularies is confusing for people who haven't realised they're for parsers, and it looks like they're constraints on how terms should be used
#
rhiaro I also think listing properties under h-*s - whilst I understand they're a guide to which make properties in which contexts - is also confusing as it implies constraints that aren't there
#
aaronpk well for the h-adr vocabulary, there is a fixed list of properties that are acceptable
#
rhiaro There are some cases where only certain properties actually make sense, sure
#
rhiaro But there's no technical reason that if someone wanted to mix it up they couldn't do that
#
aaronpk it's the parsing vs vocab issue. the parser doesn't care about the vocab
#
rhiaro Indeed
#
aaronpk but the point of h-adr or h-entry is to devine vocab
#
aaronpk s/devine/define
#
Loqi aaronpk meant to say: but the point of h-adr or h-entry is to define vocab
#
aaronpk in which case you want to see a fixed set of properties
#
rhiaro One of the things I noted about mf2 was that it doesn't over constrain like many RDF vocabs do - I wouldn't want to presume how people are going to want to use terms in the future
#
rhiaro But if those lists *are* constraints, that definitely needs to be made clearer anyway
#
aaronpk i guess the lists are meant to be constraints as much as you can constrain what people do on their websites
#
rhiaro And, what's the point of of the constraints anyway?
#
aaronpk the point is so consumers have a known set of properties to look at
#
rhiaro With RDF, such constraints allow you to infer further information that's not explicit. But nobody is expecting that with micformats.
#
aaronpk sure you can put a "p-fiddlydits" property on an h-entry, but that's not going to mean anything to a consumer so they'll just ignore it
#
rhiaro Well, exactly. Consumers are going to look for terms they are interested in anyway
#
aaronpk right, so that's the point of the list of properties
#
rhiaro So it's like guidence - this is what consumers probably want, so this is what you should probably provide
#
aaronpk i think that's a fair statement
#
rhiaro When I was running through it a while ago, there were defintely h-*s that didn't have properties listed that I thought would be valid with them
#
rhiaro I forget what now. I'll have to go through again.
KevinMarks__ joined the channel
#
KevinMarks hm, this bumps into the type versus parsing thing again
#
KevinMarks some properties are dates, so dt- makes sense
#
KevinMarks some are urls, so u- makes sense
#
aaronpk but dt- is just to tell the parser where to get the value from
#
aaronpk my point is the parsed result doesn't contain "dt-", it ends up just being "published"
#
KevinMarks but p- and e- are trickier
#
aaronpk e- is the only one that actually changes the parsed result
#
aaronpk but that's a very simple rule
#
KevinMarks right, which is the conversation we were having the other day
#
aaronpk and doesn't depend on vocab
#
KevinMarks the point it to converge vocabulary based on what is being published and consumed
#
rhiaro and if you wanted a url as text but not a link on the page, you could use p-url and get the same result
#
KevinMarks hcard and hcalendar we cheats to an extent as they were bootsrapped from existing structrues
#
aaronpk KevinMarks: my point is the parser prefixes don't appear in the parsed result, so they aren't part of the vocabulary, they are just publishing guidelines
#
KevinMarks so the point of http://microformats.org/wiki/process#Standard is to show convergence of property publication and consumption
#
KevinMarks this gets tricky if things aren't consistently URLs
#
aaronpk that's a vocab issue
#
KevinMarks and u- is a shorthand for that
#
KevinMarks so, yes, thinsg are overlaoded
#
aaronpk as in, the vocab should say the "published" property is expected to be an ISO8601 timestamp, and if it's not, then the data is bad
#
KevinMarks well, strictly this profile: http://www.w3.org/TR/NOTE-datetime
#
aaronpk similarly, the "in-reply-to" property is expected to be a URL, and if it's actually an h-cite object, the value can be found in in-reply-to.value
KartikPrabhu joined the channel
#
rhiaro starting to sound like domains and ranges being specified
#
aaronpk ?
#
rhiaro if the value of a property should be a certain datatype or object type, and you want to enforce that, it's very RDF-y
#
aaronpk orly?
#
KevinMarks well, it's lots of things - json-schema tries that too
#
aaronpk i guess i don't see what's so bad about saying "published" must be a datetime or "url" must be a URL
#
KevinMarks no, quite
#
KevinMarks it's when you get into 'must be an integer between 1 and 12' that it gets annoying and people start writing validation parsers
#
rhiaro Sure. You're probably going to get people saying you're reinventing rdf whilst pretending not to be rdf, that's all.
#
KevinMarks but RDF is such a flexible representation format, I'm sure they can parse microformats into it usefully
#
KevinMarks must practice saying that with a straight face
tantek joined the channel
#
kylewm !tell glennjones when you have a chance I'd love if you could review this PR https://github.com/glennjones/microformats-testrunner/pull/2 ... There's an Apache project that's apparenntly interested in using mf2j, I'd love them to be fully informed about how ready/no ready it is for production :)
#
Loqi Ok, I'll tell them that when I see them next
#
kylewm said love twice...dumb
#
aaronpk you can never have too much love
#
tantek aaronpk++
#
Loqi aaronpk has 911 karma
#
KevinMarks kylewm++
#
Loqi kylewm has 201 karma
#
KevinMarks karma accepted?
#
kylewm hehe
#
kylewm is 911 the number for the karma police?
chiui, tantek and gRegorLove joined the channel
#
aaronpk edited /web-sign-in (-12) (view diff)
#
tantek nice
#
tantek oh dear where did I get "RESTful" from originally when I added that?!?
KartikPrabhu and KevinMarks joined the channel
#
contractorquotes edited /hcard-examples-in-wild (-7) (view diff)
fuzzyhorns and KartikPrabhu joined the channel