#social 2015-11-05

2015-11-05 UTC
#
azaroth
Is the desire to have something like: {"value": "<b>foo</b>", "language": "en", "mediaType'
#
azaroth
": "text/html"} ?
#
aaronpk
do we have the ability to provide both plaintext and html values for content?
#
aaronpk
(where consumers know the two values are different representations of the same content)
#
jasnell
no, one or the other
#
jasnell
the only variations we can provide currently on content are language-tagged
#
azaroth
nods
#
jasnell
which is the *simplest* solution
#
azaroth
kicks RDF
#
jasnell
in other words, "content": "foo" == no language tag unless there's a default language context established
#
aaronpk
in mf2 whenever the parser encounters HTML, it also converts it to plaintext and provides a plaintext version
#
jasnell
"contentMap": { "en": "foo" } == content == "foo"@en
#
aaronpk
e.g. an e-content tag results in "content": {"html":"<b>foo</b>","value":"foo"}
#
jasnell
there's nothing stopping implementations from doing something similar, but I'd argue that makes a nice extension :-)
#
aaronpk
the nice thing is consumers who want to display HTML can sanitize the html tag and render that, or if they don't want to deal with it can just use the plaintext value
#
jasnell
{"content": "<p>foo</p>", "plain:content": "foo"}
#
jasnell
{"content": "<p>foo</p>", "jf2:content": "foo"} :-)
#
jasnell
that would keep the most common cases simple but allow implementers do provide for the alternate case. Not everyone is going to want to put two alternate copies of the content in the object
#
azaroth
FWIW, we have skirted around the same issue in Annos, and have it in spades here: http://iiif.io/api/presentation/2.0/#property-values-in-html
#
aaronpk
yeah, that's the advantage of it happening at the microformats parsing stage. publishers are only providing one version of the content that way
#
aaronpk
but that doesn't really exist with activitystreams
#
aaronpk
azaroth: wow you ahve to check the first character to know if the content is html?
#
azaroth
Yeah :( :( :(
#
Loqi
aww, cheer up
#
azaroth
We couldn't get agreement in that community to always do HTML
#
jasnell
boo that sucks
#
azaroth
And it's not valid to have @value, @type and @language on the same literal
#
jasnell
nor can you put additional metadata in there... it gets dropped by the json-ld algorithms ... http://tinyurl.com/ohqvloa
#
azaroth
@language has more variation, and more value, so we put in a kludgey processing rule
#
jasnell
aaronok: there's nothing stopping an AS2 parser from doing the same thing with content.
#
azaroth
Yep.
#
aaronpk
yeah i suppose
#
aaronpk
okay did you get those changes for the wording queued up somewhere? I can summarize in that github thread
#
jasnell
not yet but I can. basically, the proposal is to: (1) rename displayName to name, (2) drop title, (3) clarify that summary is always html, (4) clarify the relationship between mediaType and content in an object
#
jasnell
separately, we need to look at clarifying the definition of published
#
aaronpk
i will make a new issue for 1 and 2
#
jasnell
I'm +0 on 1 but if it helps us move forward and makes implementers happy, then awesome
#
azaroth
is +0,+1,+1,+1 too
#
jasnell
ok, I gotta get back to node.js commits now :-)
#
jasnell
btw, LINK and UNLINK support has landed in node.js master. woo!
#
jasnell
but that's an entirely other matter
#
azaroth
Nice! :) LINK++
#
jasnell
aaronpk: fwiw, I may come off as a stubborn ass at times but I'm always open to be convinced on almost everything. :-)
#
aaronpk
i'm also not opposed to being convinced that I am wrong
#
aaronpk
it's just good to talk things through and improve documentation, since chances are if I am unclear on something right now, others will also be unclear on it later
#
azaroth
+1
#
jasnell
yep, and to be honest, I've been staring at these specs for years now, what's obvious and familiar to me based on past discussions and work isn't always obvious to folks coming into it fresh... I often need to be reminded of that
tilgovi and kevinmarks joined the channel
#
kevinmarks
azaroth's worry about published dates was dealt with in Atom by source
#
kevinmarks
so if you republish something and change things you put in a source with the original values
#
kevinmarks
I think this is rarely used in practice though
azaroth, bblfish, tantek, kevinmarks and jasnell joined the channel
#
jasnell
kevinmarks: I've rarely ever seen source used in the wild
#
jasnell
it had been discussed for AS1 but ruled out because of lack of consistent use
#
jasnell
I've toyed with it a few times as an extension but it's really not all that useful in practice
#
jasnell
a "via" extension might be useful to explore but not in core
#
kevinmarks
agreed - I think it's a theoretical problem
#
kevinmarks
on the indieweb side, we just defined a repost model: http://indiewebcamp.com/repost
#
kevinmarks
via isn't quite the same - in practice that is a 'I found this link on someone else's site and I'm giving credit'
Shane_, azaroth, tilgovi, bblfish, Arnaud, jasnell, JonathanC, melvster, tantek, the_frey, ShaneHudson, kevinmarks, elf-pavlik, shevski, timbl, MarkS and daviding joined the channel
#
melvster
for those interested in Solid apps
#
melvster
*work in progress*
the_frey, jasnell, azaroth, melvster, MarkS, bblfish, kevinmarks and ShaneHudson joined the channel
azaroth, azaroth_, tantek and jasnell_ joined the channel
#
azaroth
jasnell_: Am here if you have questions/comments re #249, #250
#
jasnell_
hey... ok, taking a look
bblfish joined the channel
#
azaroth
The only question that came up was whether Leave can be the opposite of both Join and Arrive?
#
azaroth
Eg can you Leave a Group, and Leave a Location?
#
azaroth
(In that you can't Travel or Move yourself without stating a destination)
jasnell joined the channel
#
azaroth
hmmm. {actor: azaroth, type: Remove, object: worms, origin: can}
timbl joined the channel
#
azaroth
timbl: Congratulations on http://www.csail.mit.edu/solid_mastercard_gift :)
jasnell_ joined the channel
#
melvster
azaroth: oh just saw that, great news! :)