- From: Melvin Carvalho <melvincarvalho@gmail.com>
- Date: Wed, 24 Jun 2015 13:16:01 +0200
- To: "public-socialweb@w3.org" <public-socialweb@w3.org>
- Message-ID: <CAKaEYhJM50UE4==6EW-nKF68+R2WTy7CgrEHB4WERJubQOcCdA@mail.gmail.com>
It's been 5 years since SWAT0 was proposed as an "acid test" for microblogging platforms. In that time it's been a struggle to implement in an interoperable way, although Aaron seems to have come very close. In looking at implementing this with SoLiD we realized that there were some pre requisites. Firstly, it requires a common understanding of "following". And it seems to me that a common understanding of profiles is a pre requisite for following. Ann mentioned that Boeing was hoping that some standardization of profiles would come out of this group, and Amy also supported this, and part of her PhD will partly deal with this topic. So, I was wondering if it could be possible to try to come to a common understanding of profiles. [[ Some background reading [can be skipped] Bradfitz original post on OpenID/Yadis http://lj-dev.livejournal.com/683939.html Terminology from the social web XG http://www.w3.org/2005/Incubator/socialweb/XGR-socialweb-20101206/#Terminology Web Identity and Discovery. A spec authored by timbl, henry, andrei (and others) on web identity http://www.w3.org/2005/Incubator/webid/spec/identity/ ]] Looking at what we have. SoLiD uses linked data (webid) to denote either people or robots. The mandatory serialization is turtle, JSON LD is rarely supported, and hash URIs are common. Indieweb overloads your homepage as an indirect identifier for a person. The mandatory (only?) serialization is HTML. The first h-card tag in the document is person. ActivityPump: ?? It seems to me that the u-url attribute is very valuable here, because it allows an indieweb h-card to have a URI. At this time the indieweb parsers dont notice this, but there's no reason why that might not happen in future, if there was a need. So I see a possible path to convergence on that front. I think the biggest mistake made in the social web XG, and OStatus was to leave profiles out of scope. We ended up 5 years later with every system making their own profiles and no two different systems having interoperable profiles (except perhaps for FOAF). I'd like to understand the Activity Pump view on this. I think it would be good if some of the participants in this group could work together to find some common ground, which would make implementing the deliverables that much easier. One good thing about this topic that it's easily testable because interoperable profiles will yield a browsable social graph. [[ Below are some examples with using SoLiD / facebook: - http://foaf-visualizer.gnu.org.ua/?uri=http://danbri.org/foaf.rdf - http://graphite.ecs.soton.ac.uk/browser/?uri=http%3A%2F%2Fmelvincarvalho.com%2F - http://deiu.rww.io/profile/card#me - http://linkeddata.github.io/profile-editor/#/friends/view?webid=https:%2F%2Fakuckartz.databox.me%2Fprofile%2Fcard%23me - http://graphite.ecs.soton.ac.uk/browser/?uri=https%3A%2F%2Fgraph.facebook.com%2F721776474%23#https://graph.facebook.com/721776474# ]] Would be happy to hear thoughts on this topic, particularly to see an ActivityPump profile. * Disclaimer * this tends to be a topic that people have strong views on, philosophically. If you'd like to reply to this thread, it would be helpful to try and be constructive, preferably from implementation experience.
Received on Wednesday, 24 June 2015 11:16:30 UTC