Dan, Jeremy, Pat, Henry, Michael, Kinglsey, Ivan, ack.. everyone,
Part of me feels like I should apologise for bringing this to the
mailing list (even though it was inevitable) - this is all getting out
of scope and the last thing we need is one of the most critical
communities in what's a mini revolution to be split over such matters.
Valid arguments from all sides, technical and not - but things are
really getting conflated here, at least from what I originally intended
to put forward (probably past that and insignificant now).
I respect that everybody has made large investments, time, money, data,
deployment, training and so forth; but really, non of that need be
wasted and nobody need change anything that has any impact on any
investments thus far.
My (personal) concern is really on the 10 year timeline (a bit shorter
to be honest ;), there are limitations and things in RDF that do, 100%,
prevent the web of data as a whole from moving forwards - however,
nobody has to scrap anything.
Simply, define a non serialization specific model that caters for N3 and
RDF - then let each standard or serialization specify what it
implements/supports - the point here, and I stress, isn't to break
anything, but to open it up to innovation and allow the next decades
worth of hacking to get going
So RDF/XML is perhaps broken technically and doesn't support all these
things, who cares? it obviously works just fine for a deployment of
several billion triples, why change it? why not define it as a subset of
some core model? - I can only see one reason not to, and I hate to say
it, but some kind of pride that the work done thus far and commonly
adopted *must* be seen to be 'perfect' - please, don't take that as any
insult, as none is intended.
There are clearly very strong opinions on both sides, and very valid
reasons too - there's an easy solution that would keep everybody happy
and allow all to get on being productive and innovative - why not enable
this?
In all honesty, if this doesn't happen, I personally will have no choice
but to move to N3 for the bulk of things, and hope for other
serializations of N3 to come along - I'd do that today, but you see I'm
a huge linked data proponent and see almost unquantifiable gains from
adopting linked data - but if what I do to get a full working model of
the web of data doesn't qualify as valid RDF at some level and you all
can't utilize it, then it's a wasted effort and a road to no where -
this, is the real issue, and many others have hit it, and will hit it
again and again as time moves on.
Please, do consider, nobody need loose anything here
Best,
Nathan
- :(
Jeremy Carroll wrote:
I am still not hearing any argument to justify the costs of literals as
subjects
I have loads and loads of code, both open source and commercial that
assumes throughout that a node in a subject position is not a literal,
and a node in a predicate position is a URI node.
Of course, the "correct" thing to do is to allow all three node types in
all three positions. (Well four if we take the graph name as well!)
But if we make a change, all of my code base will need to be checked
for this issue.
This costs my company maybe $100K (very roughly)
No one has even showed me $1K of advantage for this change.
It is a no brainer not to do the fix even if it is technically correct
Jeremy