>From [EMAIL PROTECTED]
>On Thu, May 10, 2001 at 03:06:44PM -0400, Benjamin Coates wrote:
>> Content-Type is how the client decides which viewer to pass the data on to.
>> Clients are unlikely to do anything at all with things like "Description"
or
>> "Publisher" or "Title". This is the sort of thing that would be useful
>> *before* you download the CHK, like in an index or something, but putting
it
>> into the CHK itself means you have to download the whole file to get the
>> subject (for example).
>I don't have the ability to decode mpegs, it would be nice to know if
something
>is an mpeg before fetching the whole bloody file.
I suppose a redundant content-type tag in whatever indexing/searching system
you are using would be useful, and harmless enough. But I don't see the
benefit in adding dublin-core syle metadata to the CHK... what would you do
with it?
>Any argument you make for Content-Type will apply to any other generalized
>metatdata.
Here, however, I have to disagree. I don't think 'generalized metadata' is a
very useful concept, there is metadata you need to use the data (content-type,
content-encoding, blah...) and metadata used to find or index the data.
>Just because -your- particular pet application is browsers and other
web->centric stuff doesn't necessarily make it special in any one else's view.
Fair enough, but as a practical matter, I don't think many (and by no means
all) applications involve being able to identify data without any context at
all.
>(for instance)
>
>I just want to get the damn data in and out of Freenet, I expect that I'll
>know what it is and what's going to be done with it *before* it's inserted
>or fetched.
>
>What the hell do I care about Content-Type?
>
>David Schutt
How do you know what it is before you request it (or afterwards, in the
absence of a content-type tag or the equivalent)?
--
Benjamin Coates
_______________________________________________
Devl mailing list
[EMAIL PROTECTED]
http://lists.freenetproject.org/mailman/listinfo/devl
