See also: IRC log
<trackbot> Date: 29 May 2014
<slightlyoff> that's me
https://twitter.com/mikewest/status/470053687050899456
https://twitter.com/frgx/status/469976128795332608
<JeniT> https://twitter.com/w3ctag/status/469830945496113152
(and the replies to that)
action Jeni to respond to feedback on Capability URLs on www-tag list…
<trackbot> Created ACTION-867 - Respond to feedback on capability urls on www-tag list… [on Jeni Tennison - due 2014-06-05].
Jeni: I would like to hear opinions - John’s question on whether it makes sense to have different URLs for e.g. read-only calendar, read-write calendar, calendar shared with specific person, etc… as being seperate resources…
Sergey: it seems reasonable since the capabiltity URL encodes the resource and the access rights to it...
<JeniT> http://lists.w3.org/Archives/Public/www-tag/2014May/0042.html
Jeni: his argument is that it
doesn't...
... the framing in the document is that a capability URL is a
resource + access privelages for that resource.
... John’s argument is that they are actually different
resources.
<slightlyoff> this seems like hair-splitting
Dan: Is that a meaningful difference?
Sergey: it seems philosophical.
<slightlyoff> once you've vended a capability URL, is that revokable? what does it mean to vend a new one for the "same thing"?
<slightlyoff> ISTM that it's reasonable to say "this is a way to achieve X" and leave it at that
<slightlyoff> I know people want us to be the church of URL True Meaning (TM)
<slightlyoff> +1
Jeni: we might want to take the relevant paragraph out then…
<JeniT> http://www.w3.org/TR/capability-urls/#web-architecture
Jeni: the impact is - I can’t share with somebody else the URL for that particular calendar without also giving them the same permissions - that decreases the value of sharing URLs.
Dan: ?
Jeni: the fact that we can give different people links to information and they can get that info too is part of what makes the web a useful platform.
<slightlyoff> I focus on "what are developers trying to achieve by vending these?"
<slightlyoff> and it seems to be what dka is saying
<slightlyoff> at some level we're blessing a scheme of "views", but those views are all the web knows about since we don't have multi-level access in our access protocol (HTTP)
<slightlyoff> twirl: right, and I think everyone goes "wait, what?" when they encounter it
Dan: I think for example if you have a calendar on a server somewhere which is actually database records and presents itself through a URL to some end user then the question of whether a Read-only vs a read-write version of that calendar are different resources is academic - and employs too etherial a concept of what a resource is.
<slightlyoff> twirl: much easier to fold it into the location but make the location itself less meaningful since ",access" isn't personal/specific
<slightlyoff> great!
<slightlyoff> enjoy the debate this has sparked
Alex: I’ve gone another round witht he EME folks - we will have a representative from EME at our f2f in boston.
… he can share some light on the state of the spec and the goals.
… I’m hopeful that we’ll have a requirements doc.
… I will share it with this group as soon as I have it.
Sergey: in your opinion does it make sense to write the “good nice clean” spec?
Alex: I think we should talk to David about the goals that the EME group has - there is tension between compatability and preseving current systems in the world. Several of the wg participants have large and widely deployed systems.
… I think there is some interest in creating a more web platform friendly [eme] by some others at the group.
… for discussion at the f2f.
Dan: I will be at the httpbis f2f next week.
Alex: I’ve been having conversation with folks involved in http2 - one thing that has changed is priotization of resources - moved from integer to dependency based model.
… may be difficult to model ...
… it would be good from web platform perspective if we had a thought on which of those two styles we like better ...
Dan: do you have a pointer to the most relevant part of that discussion?
<slightlyoff> here's the current dependency-based model: http://http2.github.io/http2-spec/#pri-depend
<slightlyoff> the previous version used integers: http://tools.ietf.org/html/draft-ietf-httpbis-http2-04#section-5.3
http://lists.w3.org/Archives/Public/ietf-http-wg/2014AprJun/0833.html
<slightlyoff> there's no data in this post
<slightlyoff> I'm happy to have an argument about what is "enough" good for web and server developers
<slightlyoff> but I'd rather we take the stance that if there are substantive technical issues, they be raised and discussed. If there aren't, then we should move on.
Sergey: I’ve read both the capability URLs draft and discussion and the web crypto and I’m quite satisfied with both…
+1 to alex
<twirl> https://github.com/w3ctag/spec-reviews/issues/3
If people can take a look at this https://www.w3.org/wiki/TAG/Reviews and update as approproate before end of next week that would be great.
Autumn f2f plans solidified september 29 - oct 1 in London.
Sergey: plan for Yandex conference - we could organize for anyone able to make a speech at the conference and if there are enough TAG members who can make it we can organize a panel.
Dan: I note that promises are on the roadmap for IE - good news.
adjourned