CSV on the Web Working Group Teleconference -- 18 Jun 2014

<trackbot> Date: 18 June 2014

UC maturity: can we publish?

AndyS: Yes

Jeni: Yes

<JeniT> +1

resolved: ok to publish at editor's + staff contact's discretion

mappings

re http://w3c.github.io/csvw/csv2rdf/ etc

andys: [observations on jeremy's work]

<JeniT> https://github.com/w3c/csvw/blob/gh-pages/examples/csv-metadata-and-template-for-simple-weather-obs-example.md

<AndyS> The message -- http://lists.w3.org/Archives/Public/public-csv-wg/2014Jun/0092.html

jeni: I didn't get further in beyond the metadata parts

dan: nor me beyond regex

jeni: it illustrates andys' point about the complexities of such mappings

http://lists.w3.org/Archives/Public/public-csv-wg/2014Jun/0092.html

re mappings

andys: one more thing re jeremy's examples — good to be example driven

… dan could take a shot at doing same for his use case

… jeremy's could get some of the info it needs from the metadata

<AndyS> "{datetime}"^^xsd:dateTime ==> {datetime}

… e.g. for datatime, instead of [ see example ^^^ ] it could use the metadata

… which also illustrates the fact that then the template would stop being legal turtle

andys: why parse it at all, if we're treating it as an seq of chars?

jeni: treating it as text, allows us to target yaml, markdown, html, …

… but you need to be thinking about escapes etc, as diff chars used significantly in diff text formats

andys: URI templating is interesting here, as for its substitution you have to say which kind of escaping you're doing

jeni: more thinking of the escaping of the curly braces

danbri: classic security hole territory

andys: could use weird unicode 7 chars

jenit: security issue is more about inserting things that aren't properly escaped into the content

andys: sql's worse as executable afterwards, but could get data to say different things

danbri: … which could become actionable in some setting

jenit: as I mentioned before, there are lots of templating langs already out there

andys: re security/substitution, those issues also occur with other systems

… eg.. mustache does html escaping by default

… there are some tradeoffs that aren't particularly technical

jeni: wanted to note I caught up with Ross Jones yesterday, discussing json mappings, will try to do similar to jeremy's

andys: i won't have much time for next forseeable few weeks

Tabular Model spec

jeni: I think in roughly a state where it could be republished. It's hardly been touched. It now says 'cell' rather than field.

danbri: I propose we republish it

jeni: there have been a few other changes, e.g. embedded metadata

<JeniT> http://w3c.github.io/csvw/syntax/

<scribe> ACTION: jeni propose republishing Model spec on the list [recorded in http://www.w3.org/2014/06/18-csvw-minutes.html#action01]

<trackbot> Created ACTION-24 - Propose republishing model spec on the list [on Jeni Tennison - due 2014-06-25].

Metadata spec

jenit: managed to do a little work re adding some examples

need to work through that further, adding more examples.

… i made the change around terminology, having separations between schema and notes

… i.e. schemas can be reused across multiple csvs, notes are specific to specific csv file, annotations on a row or cell

… also just assuming trying to adopt json-ld throughout

…that's just an assumption

danbri: what's needed before it can go out?

jenit: depends on how polished we want it to be. there are issues throughout, some editorial tasks, some real issues needing input

danbri: in my impl I guessed, need to revise against the real doc

CSV on the Web Working Group Teleconference

18 Jun 2014

Attendees

Contents

UC maturity: can we publish?

mappings

Tabular Model spec

Metadata spec

Summary of Action Items

Scribe.perl diagnostic output