#microformats 2015-04-25

2015-04-25 UTC
KartikPrabhu and gordonb joined the channel
#
kevin marks
edited /microformats2-parsing-brainstorming (+111) "/* rel-tag */ add "urls" example"
(view diff)
#
kevin marks
edited /microformats2-parsing-brainstorming (+200) "/* xfn */ wiht urls"
(view diff)
#
kevin marks
edited /microformats2-parsing-brainstorming (+199) "/* rel-enclosure */ add urls example"
(view diff)
#
kevin marks
edited /microformats2-parsing-brainstorming (-180) "/* rel-tag */ just the urls mode"
(view diff)
#
kevin marks
edited /microformats2-parsing-brainstorming (+570) "/* xfn */ multiple values"
(view diff)
#
kevin marks
edited /microformats2-parsing-brainstorming (+88) "/* more information for rel-based formats */ link to my fork of mf2py"
(view diff)
#
tantek
KevinMarks - the top level "urls" key is misleading, because they're actually <link> / <a> / <area> elements that you're representing a collection of
KevinMarks__, KevinMarks___ and KevinMarks joined the channel
#
KevinMarks_
what name would you suggest?
#
tantek
I was going to suggest "links" but that's no better because they could be <a href>s instead of just <link href>s
#
tantek
"urls" at the top level makes it sound like *all* the URLs in the document are listed there
#
tantek
which is not true
#
tantek
hence it's a misleading key
#
KevinMarks_
I could make it get all urls
#
tantek
no that's the wrong methodology - the "could do" methodology, rather than the - "what's the use case?" and minimal design / implementation requirement methodology
#
KevinMarks_
rel_urls ?
#
KevinMarks_
related_urls ?
#
tantek
which is why I used "rels" to being with
#
tantek
s/being/begin
#
Loqi
tantek meant to say: which is why I used "rels" to begin with
#
tantek
as the top level key
#
KevinMarks_
so are you saying urls should be under rels?
#
KevinMarks_
I tried adding more values to each rel entry, but that was a breaking change, and made it complex
#
tantek
yeah. well, "rel-urls" would be better than "urls" at least.
#
KevinMarks_
it does get the css etc links as well
#
tantek
well so does the existing microformats2 parsing spec
#
KevinMarks_
where's the stuff about alternates?
#
KevinMarks_
currently I'm looking for "media","hreflang","type","title" attrs
#
KevinMarks_
shouls I check fro more?
#
kylewm
I may be interested in it eventually, for following/follower stuff
#
kylewm
i like organizing by url, i didn't quite understand that that was the suggestion earlier
#
KevinMarks_
that wasn't the original one; I found it better when I tried
#
KevinMarks_
as it works well for xfn
#
tantek
KevinMarks - again "should I check for more" is the wrong question. The right question is, "are there any use-cases that need any more?"
#
KevinMarks_
that si waht I meant
#
tantek
rarely how people interpret it
#
KevinMarks_
enclosures use type and title, feed etc use title and type
#
KevinMarks_
your point about them being a fixed set was well taken
#
tantek
kevinmarks - the attributes in your design have array values for no reason
#
tantek
e.g. name: [ Technology ]
#
tantek
HTML elements can only have a single attribute by a given name
#
KevinMarks_
ah, I followed the name pattern from the rest of mf2
#
tantek
KevinMarks - the use of "name" there is still wrong
#
tantek
it's not the same thing as the use of "name" elsewhere in microformats2
#
tantek
and using the same key to mean two different things is only going to confuse devs
#
KevinMarks_
what do you suggest instead?
#
tantek
go back to the use-case, what precisely do you need, and nothing more?
#
KevinMarks_
Name fits well for xfn and fro alternate in the feed case
#
tantek
you're conflating parsing and semantics incorrectly
#
tantek
don't do "magic" with the "title" attribute
#
KevinMarks_
you're saying it is confusing, I'm saying to plays the same semantic role as name elsewhere in mf2
#
tantek
just have it be there, along with "type" etc.
#
tantek
it doesn't because elsewhere in mf2 there's always the opportunity to have an explicit p-name
#
tantek
and you're conflating innertext and the title attribute and that's no good
#
tantek
no I retract that - I see the explicit "title" attribute now
#
tantek
"title": "real title"
#
tantek
also - in the rest of mf2, a p-name can sometimes come from other attributes, e.g. the <img alt>
#
KevinMarks_
that's why name worked - feedparser uses "term" and "label for the 2 parts fo tag
#
KevinMarks_
not sure where that si from
#
tantek
we really should just use the DOM terminology here
#
tantek
since that's essentially what we're representing
#
KevinMarks_
which would be what, innertext?
#
tantek
looking it up now...
#
KevinMarks_
sizes was mentioned by aaronpk as useful for rel-icon
#
tantek
yeah - need to write-up the use-case
#
KevinMarks_
do we want coords and shape from area ones?
#
KevinMarks_
for the person tag case?
#
tantek
no that doesn't help because rels are global/top level only
#
tantek
we need to solve that in a microformat h-* object scoped manner
#
tantek
ok found it
#
tantek
textContent
#
tantek
e.g. document.body.textContent
#
tantek
there - re-use at least
#
tantek
instead of reinvention
#
KevinMarks_
with camelcase?
#
tantek
sorry
#
tantek
yeah - might as well go with that literally, as camelCase is a JSON convention
#
kevin marks
edited /microformats2-parsing-brainstorming (+984) "/* more information for rel-based formats */ list attributes"
(view diff)
#
KevinMarks_
hm, did I mis-nest that
#
tantek
fixing
#
tantek
edited /microformats2-parsing-brainstorming (+755) "/* more information for rel-based formats */ add feedback about name, no need for array, urls"
(view diff)
#
kevin marks
edited /microformats2-parsing (+97) "/* parse a hyperlink element for rel microformats */ add title"
(view diff)
#
tantek
edited /microformats2-parsing-brainstorming (+40) "/* feedback on more rel info */ since only one at most"
(view diff)
#
KevinMarks_
see the edit i just made to parsing
#
tantek
edited /microformats2-parsing-brainstorming (+135) "/* attributes parsed */ don't need coords and shape on rels"
(view diff)
#
tantek
KevinMarks - yes that looks good
#
tantek
also wrote up the feedback on your brainstorm a bit - pretty sure I covered the different points http://microformats.org/wiki/microformats2-parsing-brainstorming#feedback_on_more_rel_info
#
kevin marks
edited /microformats2-parsing-brainstorming (+232) "/* rel-alternate */"
(view diff)
#
tantek
is there a use-case for textContent on alternates?
#
tantek
e.g. <link> has no textContent
#
KevinMarks_
your example doea
#
KevinMarks_
"French mobile homepage" is useful if offering alternatives
#
tantek
as does "Mobile homepage"
#
KevinMarks_
right - in the same way as it's useful for rel-me
#
tantek
to show a label
#
KevinMarks_
and probably for syndication too
#
tantek
alright I'm going make that a separate brainstorming item
#
KevinMarks_
did you see the mf2 I found in the embbede tweet iframes?
#
KevinMarks_
benward is marking them up well
#
tantek
edited /microformats2-parsing-brainstorming (+69) "move more information for alternates to its own suggestion since it's orthogonal"
(view diff)
#
tantek
edited /microformats2-parsing (+270) "/* parse a hyperlink element for rel microformats */"
(view diff)
#
tantek
oops meant to say: move title and add textContent on alternates to proposed
#
KevinMarks_
also should link to the brainstorming
#
tantek
edited /microformats2-parsing (+104) "link to brainstorming for more information for alternates"
(view diff)
#
tantek
KevinMarks: now that we have interop among parsers, we need to add proposed changes like that explicitly to a proposed section for review by other parser developers
#
tantek
rather than just edit the spec inline
#
tantek
also relates to keeping stable things stable - helps / necessary for being able to normatively reference e.g. from W3C drafts
#
tantek
to keep stable / proposed things clearly delineated
#
tantek
edited /microformats2-parsing-brainstorming (+101) "/* more information for alternates */ discussion, +1 makes sense"
(view diff)
#
tantek
edited /microformats2-parsing-brainstorming (+33) "/* feedback on more rel info */ discussion"
(view diff)
Soopaman, KevinMarks, KevinMarks__, iwaim__, joskar_, benward, Erkan_Yilmaz, voxpelli, tantek and tommorris joined the channel
#
tantek
also since "textContent" is both ugly camelCase, I could live with shortening it to "text".
KevinMarks, termie and AcidNerd joined the channel
#
kevin marks
edited /microformats2-parsing-brainstorming (+236) "/* feedback on more rel info */ 'text' not 'textContent'"
(view diff)
Atamido, eschnou, chiui, KevinMarks__, KevinMarks, KartikPrabhu, Soopaman, Davidino, netweb, danielfilho_, tantek, Hixie, twisted`, elux, ben_thatmustbeme, JonathanNeal and tommorris joined the channel
#
KevinMarks_
tommorris: did you see the parsing discussion?
Garbee joined the channel
#
KevinMarks_
if tantek's name changes make sense to you too, I'll make them and send you a pull
#
tantek
KevinMarks - want to update your proposal on the wiki first accordingly?
#
tantek
I'd rather we have a precise spec update proposal on the wiki we can point to for any pull requests
#
tantek
so that anyone can verify that the pull request does something according to a proposed spec update
KevinMarks, KevinMarks__, eschnou, tantek and KevinMarks___ joined the channel