Re: Content negotiation for Turtle files

On Wed, 2013-02-06 at 19:08 +0100, Colin wrote:
[ . . . ]
> 
>               You are..
> 
> 
> Human
> 
> 
> Machine
> 
> 
> You want to…
> 
> 
> Write data
> 
> 
> Turtle
> 
> 
> RDF/XML, JSon, Ntriples
> 
> 
> Read data
> 
> 
> HTML , Turtle
> 
> 
> RDF/XML, JSon, Turtle,
> Ntriples
> 
> 
> 
In my view, Turtle should appear in BOTH columns, and RDF/XML should be
completely eliminated from both columns.  (However, it is fine to leave
old data that is already published in RDF/XML in that form -- in
addition to providing an updated version in Turtle.)
> 
> Turtle: like Kingsley pointed many times, it's easy write at hand.

Yes, and easy to machine generate from a database.  Maybe even easier
than RDF/XML.  Ntriples is even easier, but causes bloated files and
makes the data more difficult for humans to read.   Bear in mind that
human readability is important for debugging even if the file is
intended primarily for machine use.

[ . . . ]
> I can't think of a situation where a machine would write Turtle.

I and many others generate Turtle all the time.
> 
> RDF/XML: Not readable by humans. Since XML is quite common to store
> data, the easiest way to product RDF from XML is to serialize RDF/XML.

No, that does not follow at all.  An XSLT script that interprets the XML
could just as easily generate Turtle as generating RDF/XML.  Actually,
it would probably be easier, because one would not have to deal with
escaping all the XML tags that must be generated for RDF/XML.

RDF/XML exists only for historical reasons.  At this point, there are
virtually no technical reasons for using it -- only historical reasons.
Turtle is simply better.


-- 
David Booth, Ph.D.
http://dbooth.org/

Aaron's Law, in memory of Web prodigy and open information 
advocate Aaron Swartz: http://bit.ly/USR4rx 

Opinions expressed herein are my own and do not necessarily 
reflect those of my employer.

Received on Wednesday, 6 February 2013 18:58:19 UTC