- From: David Booth <david@dbooth.org>
- Date: Fri, 19 Jun 2015 18:44:34 -0400
- To: Gregg Kellogg <gregg@greggkellogg.net>
- CC: public-csv-wg@w3.org
On 06/19/2015 04:19 PM, Gregg Kellogg wrote: >> On Jun 19, 2015, at 12:17 PM, David Booth <david@dbooth.org> >> wrote: >> >> A technical question: does a CSVW metadata file supposed to contain >> a particular string -- a 'magic number' perhaps -- that identifies >> it as being a CSVW file? If the metadata file is discovered first >> -- such as through a web search -- then this would help prevent a >> non-CSVW JSON file that just happens to look like a CSVW metadata >> file from being erroneously interpreted as a CSVW metadata file. > > The IANA registration shows no Magic Number, but valid metadata > documents MUST have the context "http://www.w3.org/ns/csvw”, which > will typically (although not necessarily) appear at the beginning of > the document. In fact, this is what my JSON-LD processor looks to to > not treat it as JSON-LD, but as Tabular Metadata. I think this is important to note in the IANA registration. I would suggest adding something like the following: http://w3c.github.io/csvw/metadata/#iana-considerations [[ Magic number(s): An application/csvm+json document MUST contain a JSON property named "@context" whose value is either the string "http://www.w3.org/ns/csvw" or a JSON array whose first element is the string "http://www.w3.org/ns/csvw". For parsers that are not JSON-aware, this means that an application/csvm+json document MUST contain the following two strings (including quotes), in order, separated by at least one character (not necessarily whitespace), somewhere in the document though usually near the beginning: "@context" "http://www.w3.org/ns/csvw" Thanks, David Booth > > Gregg > >> I did not find the answer to this question when I looked in the >> spec, but maybe I looked in the wrong place. >> >> Thanks, David Booth >> >> > > > >
Received on Friday, 19 June 2015 22:45:08 UTC