RE: Request for comments about requirements

Andy;

I've updated the R-TableNormalization <http://w3c.github.io/csvw/use-cases-and-requirements/#R-TableNormalization> requirement to reflect your comments ... definitely _not_ intended in the specific RDBMS context!

I've also moved this req to the "deferred" section, as transforming a "normalized" data file to a regular tabular data file would require some generalised parsing. Easy for a programmer to knock something up, difficult for us to anticipate within a specification! So I think this requirement is out of our scope.

Jeremy 

> -----Original Message-----
> From: Andy Seaborne [mailto:andy@apache.org]
> Sent: 21 May 2014 10:53
> To: public-csv-wg@w3.org
> Subject: Re: Request for comments about requirements
> 
> On 21/05/14 10:21, Ivan Herman wrote:
> >> R-TableNormalization
> >>> Ability to normalize data that is not in normal form and possibly
> >>> vice-versa.
> >
> > This is a common comment to both requirements.
> >
> > At the moment, the parsing section of the document is non normative.
> > What we define, and base our work on, is the abstract tabular data
> > model. What this means is that, strictly speaking, we do_not_  define
> > a normative parsing of CSV, ie, we do not define normatively anything
> > that would address these two. Ie, these are not 'Accepted', nor are
> > they, I believe, 'Deferred' but, rather, 'Out of Scope'
> >
> >
> Also:
> 
> "normal form" is a technical database term that carries a very specific
> and important meaning.
> 
> http://en.wikipedia.org/wiki/Database_normalization#Normal_forms
> 
> We're not requiring extract of 3NF data (I hope)
> 
> Is this requirement "producing canonical forms for values"?
> 
> 	Andy
>

Received on Monday, 2 June 2014 13:33:36 UTC