RE: ISSUE-66: LinkedData(tm)

> -----Original Message-----
> From: John Walker [mailto:john.walker@semaku.com]
> Sent: Wednesday, August 06, 2014 9:20 PM
> To: Kingsley Idehen
> Cc: public-hydra@w3.org
> Subject: Re: ISSUE-66: LinkedData(tm)
> 
> Hi Kingsley
> 
> >
> > [2] http://bit.ly/WAJGCp -- Linked Data (HTTP URI based denotation and
> connotation) in a single slide.
> >
> 
> I did not see this before. Personally I always thought of the hash part as
> identifying a fragment of the document, so it always felt a bit weird to use a
> hash URI to denote a non-information resource. This slide presents it as a
> "local identifier" which gives a subtle but important twist on that.
> 
> Note I have no desire to re-hash (excuse the pun) the whole range 14
> discussion and remain pretty much agnostic on the subject. Both hash and
> slash work technically, so it's just up to the implementor to choose which
> flavour they prefer.
> 

The simple guidance I found useful in the past (if you're able to ignore range-14) is to consider using hash if you expect to only ever define a relatively small set of terms under a URI (e.g. http://example.org/small-vocab#firstItem).

But consider using a slash if you think it possible that the set of terms under that URI could potentially grow very large. This is because the slashes allow you to introduce more and more organizational categories as the set of terms grows, e.g. http://example.org/big-vocab/category1/sub-category9/sub-sub-category123/millionthItem).
 
Of course you can always combine them - e.g. if you consider one of your sub-categories to only ever have a 'small' number of terms, e.g. http://example.org/big-vocab/category1/small-sub-category17#billionthItem).

> Cheers
> John

Received on Thursday, 7 August 2014 21:43:27 UTC