W3C home > Mailing lists > Public > public-vocabs@w3.org > June 2011

notes from Q&A session at SemTech, RichSnippets session

From: Sandro Hawke <sandro@w3.org>
Date: Wed, 08 Jun 2011 17:38:26 -0700
To: public-vocabs@w3.org
Message-ID: <1307579906.2485.13.camel@waldron>
More very rough notes.   Apologies in advance for bits I got wrong.

    -- Sandro

====================


Q:  This seems like two talks.  What's the connection between rich snippets
    and schema.org?

A:  schema.org is intended to power future rich snippets.

Q: are you talking to people making CMS's and platforms?

A: Easy to see how author-of-blog-post could be baked in, but recipes...?

    people just write the text.

Q:  (Juan)   You've created a vocab for what people can say.  How
    fast can you expand the schema...

A:   How do we scale up new vocabs, that people want.

     We're laying the barebones.   There is an extension mechanism.

     eg a particular kind of event.    extn lets you say this is a subtype
     of this....   it has to be widespread before we'll pick it up.

     [[ so they're doing subtype reasoning...???? ]]

Q:  (Roger)  you targeted a particular user group.   do you have plans for other user groups?

A:   Do you mean "people who work in a particular domain?"

     Roger: eg B-to-C, not enterprise

     Kavi:   we used what people search for on the web.

Q:  Any thought to extn for registry, where communities can develop common extension, to crowdsource the dev of stds

A:  This is a big open question.  There's a lot of work left to be done.  Interesting suggestion.   We haven't decided on anything yet.

A2:  That would be a great problem to have.

followup: Bigger is almost never better -- you need app/community-specific extension.  Eg Biomedical.

A3:  Like how do I markup dogbreeds or cars, etc.   There was a balance.

Q: It struck me, something you said.  An alternative path.   You said "there could have been a better semantic markup tool".   You could have ....    back in 1995 HTML was too hard.   You COULD have said RDF is too complex, lets make better authoring tools.

A:  The question was -- you could have produced better tools.  Implied, "pick RDF".   But there are more standards than RDF, and others have adoption as well.   I don't think it would have been possible to create the perfect tool.

A2:  As a webmaster, I should be able to just write ingredients in HTML, why do I need markup for it?    Maybe keywords?

followup: your rhetoric has leverages "complexity vs simplicy" a lot; you seem to have not purused the tool part.

Q: When you decide to make a new rich snippet, say for sport stores, do you and MS and Yahoo argue about that schema?  Who will be at the table?

A: So far, but we've all be compelled by the urgency of doing something.  For any domain, it's nice if domain experts can be part of the conversation.

followup: So you intend to reach out like that?

A: Yes, there are such conversations underway.

Q:  Other than rich snippets, where within Google will this be surfaced, eg APIs.  And will the markup include the ability to opt-out of some of the aggregations.  I might want my photos tagged, but not aggregated into a different context.

A:  Google is a large company, we can't represent everyone there.  Withint google, there are lots of ways data is collected.  Some of the ways, already: local info using hcard, it can be used by geo team for place pages.   Video search and image search teams.   Fairly limited clues so far.   Ads.   Social.   Tons of potential places.    Can we provide controls....?  that's a big open question.   Our phil. so far has been that whatever we're doing is ultimately helping the content creator.   We want to keep that up.    Also, what could the data be used for besides the search engines.

Q:  The next step.    A version of the data without the mess.   We might tell you what the web pages is about.    Something more direct -- instead of marking up content, we just tell you what's in the page, a bit like sitemaps.

A:  It sounds like you have a particular idea.

followup:   .rdf on every webpage

AF:  Seems like we're moving in the direction from inferring what web page is about, that the content creators could SAY what web page is about.   Lots of questions about how that should be done.   It could be baked in, it could be a feed, it could be a separate file.

Q: Above and beyond a particular schema...     there are rumours about, say, how you handle GoodRelations.

A: Q is what do we think about non-schema.org vocabs?

   Certainly we dont want to decrease the activity in this space, we want to build community.

followup: More direct -- what do you now do with these vocabs?

AF: If they only did goodrelations, it wouldn't make sense of any search company to not use it.

A2: Speaking specifically about GoodRelations -- we're not shutting off support for anything.  GoodRelations and RDFa and Microdata are going to continue to be supported.  It's not in our interest to stop consuming data we're already consuming.  The question is about new things.

Q: How much spam have you gotten?

A:  Not a lot so far.   We've been fairly conservative, though.   We've erred on the side of not showing good content.   Once it becomes poluted, it would be hard to recover.

A2: Since there is a UI for some of these types, .... there are some
incentives against spam.

Q: Metadata layers of video, 

A: I'm not an expert on that peice, I don't know.

A2:  If there's enough information and enough tags, then we should look into it.
Received on Thursday, 9 June 2011 00:38:34 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Tuesday, 22 May 2012 06:48:54 GMT