Re: CSV parser specification?

Alf, Gregg, all,

You may have seen that I’ve marked the sections in 

  http://w3c.github.io/csvw/syntax/

about parsing and syntax as non-normative. I’ve also changed the title to:

  Model for Tabular Data and Metadata on the Web

This is because:

  (a) Ivan has pointed out that this Working Group is not chartered to 
      standardise the syntax of CSV, so we could face battles with the W3C 
      Advisory Committee if appear to be doing that

  (b) as Yakov has said, he’s happy taking standardisation forward within IETF

  (c) doing so would be a big extra drain on our time because we’d have to have
      test cases and implementations and so on to get to Recommendation

I’m aware that a lot of people in the Working Group expressed that defining the parsing of existing CSV was something they wanted to be in scope. Also I think it’s useful for us to document the variety of ways in which people use CSV, the variety of options etc that tools support, and so on (to help future implementers and to help Yakov with the IETF work).

With non-normative sections in this spec we can have that helpful text, but we don’t have to worry overly if it’s not at the level of specificity that we’d expect from normative specification text, or about the extra things you normally need to do when standardising syntax.

Anyway, do add the EBNF, Gregg (and add yourself as editor).

Alf, I definitely think it’s worth describing different levels of parsing. I would view the ‘Liberal’ parsing that you describe as transformation rather than parsing, I think: it would be interesting to try to define where that line is; perhaps something to do with parsing not removing or changing the information in the file whereas transformation might.

Cheers,

Jeni
--  
Jeni Tennison
http://www.jenitennison.com/

Received on Thursday, 6 March 2014 18:12:38 UTC