- From: Andrea Crevola <andrea.crevola@3juice.com>
- Date: Tue, 09 Aug 2005 10:08:03 +0200
- To: public-bpwg@w3.org
Hi,
Probably we do not need only a controlled vocabulary for device
capabilities description. We *run the risk* of stealing freedom to our
users if we delegate all the work to formal descriptions of device and
context distribution.
We need something that could let the user choose the quality and the
quantity of information. It can be a menu putted on home page, a browser
button or a server module.
I think that one thing on it is possible to work is the rendering of the
element <link /> by user agents. It may be used to give information
about relationship between the actual page and the other information
units of the web. Something like:
<link rel="headlines" href="/headlines.php" title="Headlines only
version" />
I know that Opera Browser recognizes this kind of tag and creates a
navigation bar. Firefox (with some extension) do the same.
A mobile browser can perform in the same way: it looks for these tag and
build a menu (or a button) and the user can make use of it.
An other way to reach a similar goal is develop a sort of "Rosetta
stone" that could relate terms written in various languages. It could be
a controlled vocabulary used by servers, proxies or clients for give the
same meaning to different words. I do not know if it could be name
"onthology"...
Let's imagine: this element receive a request by a client, that is - for
example - http://www.bbc.co.uk/solotitoli/ ("only headlines", in italian
:) ). This request is processed trying to find any occurrence in a data
repository (a database, an xml file...). Imagine that - in our example -
our module finds that "solotitoli" is an *alias* of "headlinesonly".
It translate the request in a standard form and passes it to the origin
server.
The <link> strategy is a client-side strategy, the second one could be
performed by a proxy or by a server.
What are your opinions?
Andrea
Received on Tuesday, 9 August 2005 08:09:16 UTC