W3C home > Mailing lists > Public > public-wai-ert@w3.org > May 2007

Re: Antw: Re: [HTTP-in-RDF] "Simplified" approach

From: Alan Dean <alan.dean@gmail.com>
Date: Thu, 31 May 2007 08:03:54 +0000
Message-ID: <6fb173f50705310103x5e9ed843r88aa4da19562030a@mail.gmail.com>
To: public-wai-ert@w3.org

Reinhard Ruemer wrote:
> Allowing both, would mean that information is kept in a redundant
> manner (as in your example).

I don't actually accept your premise that the information is actually
redundant. But, for the sake of discussion, let's say that it is.

Why do you object to the redundancy? It would be entirely optional.
For me, the important point is not redundancy, it is flexibility
(let's face it - the RDF/XML syntax is plenty verbose anyway). I think
that it is iniquitous to force a choice between the two
representations when allowing a side-by-side representation allows
support for consumers who want the primitive *and* those who want the
formalised decomposition.

By inference, therefore, I assume that if you are against redundancy
between http:fieldValue and http:headerValues in my example then you
are also against the redundancy between http:fieldName and
http:headerName in the original proposal at

> I think it is hard to keep track of both
> representations at the same time that they are equivalent.

I would be very interested to see what the semantic equivalence rules
are and how you would carry out a comparison using, say, XPath.

> The formalised decomposition is in my opinion the better way of
> presenting the data because it is the more structured and formalised
> way.

Other than the subjective, what practical benefits does formalised
decomposition bring that merits it's preference over the literal

I can rattle off a set of benefits in the opposite direction that
would prefer the literal (ease of use being the primary one, but there
are others too).

> If someone "needs" the literal string it is not a big thing to get
> it out via a script - again because the decomposition is very a
> structured way of presentation.

I am naturally suspicious of a representation that requires you to
execute a script merely to obtain the original input value (see my
XPath comparison question above).

By the same token, a script could emit the formalised decomposition when needed.

However, the critical point is that my proposal does not favour one
over the other - it applies a "live and let live" principle that
permits both representations without requiring one to be overwritten
by the other.

Alan Dean
Received on Thursday, 31 May 2007 08:22:25 UTC

This archive was generated by hypermail 2.3.1 : Tuesday, 6 January 2015 20:55:55 UTC