home *** CD-ROM | disk | FTP | other *** search
- Path: sparky!uunet!think.com!zaphod.mps.ohio-state.edu!magnus.acs.ohio-state.edu!usenet.ins.cwru.edu!agate!ames!bionet!raven.alaska.edu!news.u.washington.edu!nntp.uoregon.edu!jqj
- From: jqj@nntp.uoregon.edu (JQ Johnson)
- Newsgroups: alt.gopher
- Subject: Re: Whois/finger patch for Unix client
- Message-ID: <1992Jul21.151330.27447@nntp.uoregon.edu>
- Date: 21 Jul 92 15:13:30 GMT
- References: <BrHzyC.8ID@usenet.ucs.indiana.edu> <19920717180944SEB1525@MVS.draper.com>
- Organization: University of Oregon Network Services
- Lines: 37
-
- Any new extension mechanism for gopher should do more than the present
- "add a new type" mechanism. I see the major current problem with the
- protocol being an overloading of the semantics of the type character.
- Currently, it is used to distinguish:
- - different kinds of object on a server (e.g. different file
- content types)
- - different network interaction protocols (e.g. menu based file
- retrieval, telnet, etc.)
- - [if the Schemers approach is widely adopted] different
- argument parsing protocols i.e. different interpretations of
- the selector line
- - etc.
-
- I see these as largely orthogonal. For example, I can imagine wanting
- string RPC arguments [the RS protocol] for any 0-like or 1-like gopher
- type, but I note that they are not very relevant to the telnet type.
- I can imagine a different style of argument processing also being
- desired, again applying to any type of returned object. And so on.
-
- Bottom line: a protocol extension mechanism should satisfy new needs.
- Unfortunately, I'm not convinced we know yet what those new needs are.
-
- P.S.: we do have SOME ideas of the new needs. It's clear that many
- people want more attributes/properties available for file-like
- objects, e.g. creator, who to contact, etc (the Brown proposal).
- It's clear that a more powerful alternation mechanism than the present
- "+" type is desirable to allow clients to select their preferred
- representation of equivalent data. It's clear that the present
- single-character filetype space is too small. It's clear we need to
- continue to think about scripting and such. It may be that some type
- of authentication mechanism is needed. And so on. The real challenge
- may be to build extensions that preserve protocol purity and simplicity.
- --
- JQ Johnson
- Director of Network Services Office: 250E Computing Center
- Computing Center Internet: jqj@ns.uoregon.edu
- University of Oregon voice: (503) 346-1746
-