PeerJ / jats-conversion

Conversion and validation for JATS XML
MIT License
51 stars 29 forks source link

Improve semantic #99

Open ppKrauss opened 9 years ago

ppKrauss commented 9 years ago

Perhaps the (simplest and) best way to improve JATS with formal semantic is to use http://www.w3.org/TR/sawsdl/ at the JATS XSD, extending it... But I never used sawsdl, and not see initiatives.

By other hand, this jats-conversion project have the jats-to-html.xsl with a Microdata semantic iniciative(!). In the context of https://github.com/JATS4R/elements/issues/97 it is possible to discuss (looking for community consensus) and improve semantic description of HTML content.

... How to, discuss and collaborate in "Improve semantic" initiative? It is not so easy neither obvious, here a methodological suggestion:

  1. (continue to) Add to jats-to-html the consensual semantic.
  2. discuss and suggest (see status column) more semantic in a simple sheet (ex. columns "JATS-XPath", "status", "WikidataID", "scope", "URL1", "URL2")... See this draft.
hubgit commented 8 years ago

@ppKrauss Can you provide concrete examples of semantic markup that needs to be added to the output HTML?

ppKrauss commented 8 years ago

@hubgit , aff and contrib semantic can be translated/mapped to Schema.org markup, using for example, Organization and Person classes.

PS: there are many standard HTML5 usual semantic that can be also mapped (ex. JATS <abbrev> is the HTML5's <abbr>).

hubgit commented 8 years ago

aff and contrib semantic can be translated/mapped to Schema.org markup, using for example, Organization and Person classes.

Those are already being used here - are there any that are missing?

ppKrauss commented 8 years ago

(ops, sorry sending a correct comment now) I cited jats-to-html.xsl as example, but I think we can improve your jats-to-html with a lot more little semantic enhancements... It will be time consuming, but if you approve I can plane a first try, a draft to submit here.

hubgit commented 8 years ago

@ppKrauss Maybe you could suggest one enhancement that's missing, as a start?

ppKrauss commented 8 years ago

@hubgit See a list of XPaths here, as showed before... Each XPath can be implemented as a itemscope/itemtype as you do with schema.org 3 examples at jats-to-html.

I can implement some XPath samples into your jats-to-html XSLT to a "complete example"... At the next weekend.

hubgit commented 8 years ago

@ppKrauss Which of those is not currently implemented here?

ppKrauss commented 8 years ago

@hubgit I am checking... Well done (!), I not see before all the included XSLTs. You do a lot of semantic, and we can "look for more with a magnifying glass" later... We can improve the spreadsheet, to catalog all implemented semantics, facilitating reviews and new issues.

Examples that can be added

//collab
    use itemprop="name", but can be "legalName" or use https://schema.org/Organization

//issn
//issn-l
     need itemprop="issn", see https://schema.org/issn
//isbn
    need https://schema.org/isbn
... I can check some more, later... Let's standardize a spreadsheet.

Examples where XSLT is fine

//ref-list/ref
//ref-list/ref[element-citation/@publication-type = 'journal']
//ref-list/ref[element-citation/@publication-type = 'book']
    itemprop of http://schema.org/ScholarlyArticle and http://schema.org/Book

//front//article-title
    itemprop of http://schema.org/ScholarlyArticle

//given-names     itemprop="givenName" 

//surname         itemprop="familyName"

/article/front/article-meta/article-id[@pub-id-type='doi']
    itemprop="sameAs"

//aff    itemscope="itemscope" itemtype="http://schema.org/Organization"

// ...   ....

We not need to restrict the itemtype attribute to SchemaOrg ones, the rule of thumb for ref semantics may be:

  1. https://schema.org properties and classes
  2. https://www.wikidata.org Q-properties
  3. http://jats.nlm.nih.gov/publishing/tag-library/1.0/ or 1.1 (direct links to tags).