social-ISSUE-44 (removeMFfromAS): Remove microformats examples from ActvivityStream core and vocabulary [Activity Streams 2.0]

social-ISSUE-44 (removeMFfromAS): Remove microformats examples from ActvivityStream core and vocabulary [Activity Streams 2.0]

http://www.w3.org/Social/track/issues/44

Raised by: Sarven Capadisli
On product: Activity Streams 2.0

re ACTION-26: http://www.w3.org/wiki/Socialwg/2015-06-02-minutes

"<Arnaud> RESOLVED: revert change that removed microformats examples, and add a warning e.g. "This example needs review." to any microformats examples that have not been verified/fixed per action 26. Also others are welcome to help with action 26!"

To the best of my knowledge, this action:

http://www.w3.org/Social/track/actions/26

was neither carried in full nor with sufficient quality. See also http://www.w3.org/Social/track/actions/26#n163162 where it states "First pull request to fix the first two microformats examples in AS2." And, there was no response to https://github.com/jasnell/w3c-socialwg-activitystreams/issues/83#issuecomment-99995834 before closing the issue.

Action 26 is incomplete, but more importantly, the AS mf examples still do not properly translate in any shape or form to the *intended statements*. See also:

https://github.com/jasnell/w3c-socialwg-activitystreams/issues/194
https://github.com/w3c-social/Social-APIs-Brainstorming/issues/14

Therefore, this is a second request to remove the mf (microformats) examples from AS core and vocab.

There are two distinct issues here:

1. Possible mapping between the AS and mf terms
2. Quality of translation from AS statements to mf statements

#1 has partial coverage (some preliminary mapping here: https://www.w3.org/wiki/Activity_Streams/Microformats_Mapping ). #2 has been pointed out more than once in which the translations are inappropriate (i.e., weak or not possible). #2 is an important point which should not be prematurely dismissed. AFAICT, #2 is not possible in a reasonable way, within a reasonable amount of time, within a reasonable quality for the SocialWeb WG because:

1. mf does not have the vocabulary (since it is not using AS's vocabulary, but rather trying to find the closest "equivalent" from its own vocabulary), so, it may need to come up with equivalent terms/expressions, which is actually entirely dependent on:

2. mf's own vocabulary development process (i.e., document what's going on the whole Web, .., ..)

Is this something the mf community currently or will be working in parallel to the AS vocab, in order to showcase the examples on time? What's the progress like on the mf wiki at the moment?

3. mf "statements" using the RDF language ends up being obscure, e.g., at this time, a mf "statement" essentially looks like this when transformed to RDF Turtle:

_:bnode1 _:url "http://tantek.com/" .

Read: bnodes are problematic, and properties can't even be bnodes in RDF.

Even if we apply code/mapping magic, the best we get is something like this:

<http://tantek.com/> urn:microformats:url <http://tantek.com/> .

Which is arguably not very interesting/useful.

On the other hand, there is a clear and meaningful conversion process between JSON-LD, Microdata, RDFa, and Turtle snippets.

Keeping the mf examples in the AS spec only adds clutter, and confusion (for current and future readers) since the mf examples *do not* even use the AS vocabulary.

The removal of the mf examples from the AS spec:

* is likely to resolve these issues
* free up resources

Proposal is to move the mf examples to a wiki page (where interested parties can carry out the desired exercise), or create a separate NOTE. The latter was previously suggested by James here:

https://lists.w3.org/Archives/Public/public-socialweb/2015Jun/0009.html


All of the other representations; JSON-LD, Microdata, RDFa, and Turtle, are capable of exemplifying *the AS vocabulary* - which is the primary point of showing the examples in different serializations in the first place. If we side-step that, having the mf examples is equivalent to having arbitrary formats making arbitrary statements (counter argument to that is: if you look at it from a different angle, combine x and y when z is present, but not u, then it is equivalent). Even the following CSV is more accurate and meaningful than the mf and is in fact in line with AS (using http://jasnell.github.io/w3c-socialwg-activitystreams/activitystreams-core/index.html#example-1 ):

"Activity", "actor", "object"
"Create", "http://www.test.example/martin", "http://example.org/foo.jpg"

If Microdata, RDFa, and Turtle are moved to a separate document/NOTE, it should be different than the microformats one - where it may be accompanied with CSV etc., since there is clear distinctions between them. If somehow all of that ends up in the same document/NOTE, the differences between them and their intentions should be absolutely clear.

PS: I'm not advocating CSV examples. It was only used to compare.

-Sarven
http://csarven.ca/#i

Received on Thursday, 6 August 2015 08:58:09 UTC