Re: Metainformation about arbitrary Objects..

Nick Arnett (narnett@verity.com)
Sun, 1 Jan 1995 03:03:24 +0100


At 4:49 PM 12/31/94, WWW Mailing list common A/c. wrote:
>Subject: Metainformation about arbitrary Objects..
>
>Hello!
>I am fairly a novice in the world of Web. I understand the HTTP is very
>general purpose and one could have Newsgroups, or mailing lists configured
>and accessed transparently through HTTP servers. I would like to know how
>and what kind of meta information that one could produce on arbitrary
>objects such as Newsgroups, Mailing lists, etc. A newsgroup can have
>information
>like name of the group, a number of articles posted, a newsgroup editor
>(if any) etc. Can one provide all such information in HTTP headers?
>What would a browser do with such information?

The HTML 2.0 draft includes a "META" element for encoding such information,
which is probably of limited use to most browsers, but of tremendous
importance to indexing applications like the ones that we build.

We are adding embedded support in our search engine for capturing Internet
news and mail headers automatically to fields, to simplify the most common
kinds of field capture that people want.

>Another related question is: if an application like mail is permitted then
>posting a message is probably fine; however, how is one supposed to read
>his/her messages? All operations on a mail box are to be suitably converted
>into hypertext source files, etc. Isn't this really bad? Are there any
>HTTP servers on the Internet with which such services are being offered?

I suppose it would be quite bad if it were one's private mailbox, but for
mailing lists like this one, it can be an effective means of searching and
browsing discussions. This list is part of such an archive that uses some
software I hacked together last summer. It's available as part of the
Internet Virtual Library that we're building --
<URL:http://www.verity.com/library.html>.

Nick