- From: Robin Berjon <robin@berjon.com>
- Date: Tue, 22 Sep 2015 15:01:17 -0400
- To: Olaf Drümmer <olaf@druemmer.com>
- CC: "public-digipub-ig@w3.org" <public-digipub-ig@w3.org>
On 22/09/2015 13:03 , Olaf Drümmer wrote: > There are no mechanisms for complex > document substructures in other cases [no, SVG doesn't count because > it might be complex but is not well prepared to express semantic > structure], and unfortunately HTML5 cannot be extended by 'custom > tags' or tags from some other tag language (whereas that is possible > for SVG…). FWIW SVG *does* express semantic structure: the semantic structure of 2D geometric objects. But that's another discussion altogeter. > Now, if we take citations as a worthwhile example of a complex > document substructure, and also look at other examples of complex > document substructures, like diagram, bar charts, maps, musical > notation, etc. - can we think of an approach that not only serves > citations but also (some of) the other types/classes of complex > document substructure? My main interest here is not to run the risk > of fragmenting the landscape even further such that each individual > problem gets its own specialist solution… (despite the fact that one > could argue they all belong to the same class of problems) The problem we have is that we already have several means of encoding such structure, using HTML alone. Take (with some spacing added for legibility): <li typeof="schema:ScholarlyArticle" resource="sa:0C8B0478"> <span property="schema:author" typeof="schema:Person"> <span property="schema:familyName">Author</span> <span property="schema:givenName">A</span> </span>, <span property="schema:author" typeof="schema:Person"> <span property="schema:familyName">Author</span> <span property="schema:givenName">B</span> </span>. <span property="schema:name">Manuscript title one.</span> <span property="schema:isPartOf" typeof="schema:PublicationVolume"> <span property="schema:isPartOf" typeof="schema:Periodical"> <span property="schema:name">Journal Abbrev.</span> </span> <span about="sa:0C8B0478" property="schema:datePublished" content="2015-07" datatype="xsd:gYearMonth">2015 Jul</span>; <span property="schema:volumeNumber">10</span> </span>: <span property="schema:pageStart">10</span>-<span property="schema:pageEnd">100</span> </li> No doubt it's a little verbose (not counting the prefix declarations). But it's a highly structured encoding, using RDFa, of a schema.org translation of the CrossRef model. This can also be done with Microdata, with embedded JSON-LD (leaving the HTML to just be for humans), and to a lesser extent with Microformats. But hey, it works :) I can run my tooling on a document containing that and get a machine-readable understanding of the citation. -- Robin Berjon - http://berjon.com/ - @robinberjon
Received on Tuesday, 22 September 2015 19:01:45 UTC