Evolution.
The Turtle spec was February 2014 -- there wasn't a formal definition
before RDF 1.1 -- and the grammar for prefix names and the rest is the
same as SPARQL 9all the grammar rules ending "s")
This is on output, so hand written files are not a strong factor here
because they can be read in.
PREFIX and @prefix are both legal inputs to the parsers.
Teaching this stuff does get the question "why?".
Another argument has been "storing RDF text file in git". There is some
point to that but output isn't stable anyway (bNodes! hash tables! Small
changes of RDF triples can cause big changes in the text output)
What do other toolkits do?
I think it is a combination of migration and doing the expected.
Andy
On 20/09/2019 21:10, ajs6f wrote:
Ditto, but I'd be interested in hearing why you suggested it, Andy. IOW are
there some benefits that aren't obvious?
ajs6f
On Sep 20, 2019, at 7:17 AM, Claude Warren <[email protected]> wrote:
For me this falls under "if it ain't broke don't fix it"
so
-1
On Fri, Sep 20, 2019 at 11:47 AM Andy Seaborne <[email protected]> wrote:
The Turtle and TriG writers output "@prefix".
RDF 1.1 allows PREFIX.
Should we change the writers to output PREFIX? (after the next release)
(we can add options but majority of users don't set options and exisintg
code doesn't)
Andy
--
I like: Like Like - The likeliest place on the web
<http://like-like.xenei.com>
LinkedIn: http://www.linkedin.com/in/claudewarren
