Dear Wikitech (cc'd Wikisource)
A recent discussion in English Wikisource's Scriptorium was querying why commercial book companies, etc. were getting higher search hits, especially where they may just have summary information, rather than full text. In that discussion someone pointed to some of the webmaster information at Google, eg. [1], which (ultimately) talks about their microformat (preferred) or JSON-LD as a means to put in more particular metadata as explained at schema.org (for creative works [2])
I went to play, and ultimately failed, and was pointed to the inability to <script> for security reasons, and the inability to add micodata (<cite user="bawolff">microdata attributes are implemented in MediaWiki, but currently disabled via $wgAllowMicrodataAttributes</cite> thx).
So my naive questions to those that know these things are 1) How do we look to improve external search engine hits for the sister sites where they are particularly pertinent to a search [wikipedia already gets Google special treatment] 2) if the schema.org metadata is a preferred means to progress, what is the recommended means to progress such an issue 3) presumably some of this fits into the discussion about Structured Data discussion, and what means is there to include this into that discussion?
Thanks for the guidance.
Regards, Billinghurst
[1] https://support.google.com/webmasters/answer/3227642?hl=en&ref_topic=322... [2] http://www.schema.org/CreativeWork
On 6 December 2014 at 22:36, Wiki Billinghurst billinghurstwiki@gmail.com wrote:
- if the schema.org metadata is a preferred means to progress, what
is the recommended means to progress such an issue
I hesitate to mention it, as I'm aware that this is likely to start a debate with lots of heat and little light, but you should be aware that schema.org is (was?) a mostly-failed attempt by the big search engines to get a new standard for meta-data widely used faster than had it gone through the existing processes, and declaring it "a preferred means to progress" is a very good way to start a fight with certain kinds of techies. :-)
- presumably some of this fits into the discussion about Structured
Data discussion, and what means is there to include this into that discussion?
My personal preference would be for our efforts to focus on using Wikibase (either on wikidata.org or in "local" installs) for structured data and meta-data alike, rather than forking the workload. Emitting RDFa sourced from Wikibase on related pages sounds like a reasonable way to achieve more richly-laded pages which is helpful for users (and with the side-effect, rather than primary intent, of SEO).
Thoughts from others?
J.
wikitech-l@lists.wikimedia.org