Re: Turtle version naming?

On 04/19/2013 10:12 AM, Jan Wielemaker wrote:
> Hi,
>
> While planning to sync SWI-Prolog's Turtle support with the draft, I was
> wondering about version naming.  If I understand the current state 
> correctly, the draft is fully upward compatible with the `traditional'
> Turtle, so there is not really an issue for reading turtle documents.
>
> When writing however, it may be wise to be able to save in the old 
> version.  I'm planning to have either some global application setting
> or an extra argument to specify the version, but I have no clue how to
> name the version.
>
> Also, documents may want to claim they are traditional or `new'. Is
> there something that takes care of that?

The group probably needs to talk about this before giving an official 
answer.     So this is just my thoughts.

How about "standard turtle" for stuff conforming to the upcoming 
Recommendation, and "non-standard turtle" for everything else?

If you need to distinguish among the various non-standard versions then 
I guess you'd do so by naming the spec or implementation that's being 
used as a guide.   Such as "non-standard turtle (as per 2011 W3C turtle 
submission)" or "non-standard turtle (as per Dave Becket's 2006 spec)".

I think your question kind of assumes all those specs are the same, and 
I don't think they are; plus they are kind of underspecified, so really 
you're talking about.

Hmm.  Maybe what you really want could be called the "conservative 
subset" -- a subset of Turtle that you're confident every existing 
turtle parser can handle.    A subset of all the non-standard turtles.   
Not quite sure what to call it.   "Turtle (for outdated parsers)"?

I trust we'll heard 20 other opinions, since discussions of naming are 
so fun.

       -- Sandro

Received on Friday, 19 April 2013 15:27:51 UTC