Re: Finding Metadata for CSV Files

Oh yes of course! I did not realize this... This is certainly another way that should be documented. Although it may suffer from the same problem as the reference through the header: many data publishers may have difficulties setting this up at the server side...

Ivan
 
On 13 Mar 2014, at 22:00 , Yakov Shafranovich <yakov-ietf@shaftek.org> wrote:

> When using MIME it may be feasible to use multipart, include two
> separate files and reference one against the other via a custom
> header. Obviously this would not address the case of two files sitting
> on disk.
> 
> Yakov
> 
> On Thu, Mar 13, 2014 at 2:56 PM, David Booth <david@dbooth.org> wrote:
>>> On 09/03/14 15:07, Jeni Tennison wrote:
>>>> From: Andy Seaborne andy@apache.org Date: 9 March 2014
>>>> at 10:33:57:
>>>>> 5. (no advocacy) Naming convention : if there is a "data.csv"
>>>>> then the metadata is adjacent under "data.csv.json"
>>>>> or somesuch.
>> [ . . . ]
>> 
>>>> Yes, that would be an alternative, but I don't think we
>>>> should include it as an approach in a Recommendation. It
>>>> runs counter to the arguments in
>>>> 
>>>> http://tools.ietf.org/html/draft-ietf-appsawg-uri-get-off-my-lawn-01
>>>> 
>>>> that caution against a protocol that specifies a particular
>>>> URL path.
>>> 
>>> The cat is already out of the bag on that one!
>>> 
>>> http://central.maven.org/maven2/org/json/json/20140107/
>>> 
>>> json-20140107.jar
>>> 
>>> implies
>>> 
>>> json-20140107.jar.md5 etc
>> 
>> Agreed.  And even from a WebArch perspective, I don't think a convention
>> like data.csv.json would be bad if it were treated as a *heuristic*. The
>> WebArch issue is that you don't want to impinge on the server owner's right
>> to associate any URI (that he/she owns) with any resource.   But if
>> data.csv.json had to contain an *explicit* statement indicating that it held
>> metadata for data.csv, in order for it to be authoritatively considered
>> metadata for data.csv, then this WebArch principle would not be violated,
>> because the server owner could still use the data.csv.json URI for a
>> completely different purpose without harm.
>> 
>> In other words, the rule could be something like: "If you find data.csv, see
>> if there's a data.csv.json.   If there is, *and* it explicitly says that it
>> is metadata about data.csv, then treat it as such."
>> 
>> Bottom line: I think this is option should be seriously considered (though
>> not to the exclusion of others as well!) as a simple, practical way by which
>> anyone could associate metadata with their CSV files, without requiring
>> changes to the software that generates those CSV files, and without
>> requiring any special server configuration.
>> 
>> David
>> 
> 


----
Ivan Herman, W3C 
Digital Publishing Activity Lead
Home: http://www.w3.org/People/Ivan/
mobile: +31-641044153
GPG: 0x343F1A3D
FOAF: http://www.ivan-herman.net/foaf