home *** CD-ROM | disk | FTP | other *** search
- Path: sparky!uunet!elroy.jpl.nasa.gov!swrinde!zaphod.mps.ohio-state.edu!sol.ctr.columbia.edu!The-Star.honeywell.com!umn.edu!news2.cis.umn.edu!gopher-news-request@boombox.micro.umn.edu
- From: gopher@boombox.micro.umn.edu ("The Gopher Team" )
- Newsgroups: alt.gopher
- Subject: Gopher Phase II Wish-List Compilation
- Message-ID: <9207241825.AA01032@boombox.micro.umn.edu>
- Date: 24 Jul 92 18:25:17 GMT
- Sender: news@news2.cis.umn.edu
- Distribution: alt
- Lines: 85
- Approved: alt.gopher@news.cis.umn.edu
- Original-To: gopher-news@boombox.micro.umn.edu
-
- This is being written to put rumors, fears, injured egos, and
- alt.gopher flames to rest.
-
- Yes, internet Gopher originated from the University of Minnesota.
- Yes, enormous amounts of invaluable contributions from other folks on
- the Internet have changed its look and capabilities from a campus-wide
- information system to an Internet-wide document search and retrieval
- system. And yes, this growth and change has been evolutionary and
- somewhat disorderly (as evolution often is).
-
- While the original Gopher Team does maintain an archive site for
- convenient distribution of all the gopher software, mods, and tools,
- and while we'll continue to incorporate enhancements designed here and
- elsewhere, we're not a Standards Committee. Our small size and
- relative freedom to move allows us to get things done quickly, and we
- value that. Gopher was put together with a handful of guiding
- principles, and we try to view changes in the light of those
- principles. To recap the basic ideas:
-
- + Keep the protocol connectionless (one request/one response per transaction)
- + Keep intelligence in the server. Keep the protocol simple.
- + Keep requests and responses (directory ones at any rate!) to readable text.
- + Be able to debug clients and servers using telnet.
- + Make the client-writer's job as simple as possible
- (client writer's real job is the user interface).
- + Clients must run efficiently on Macs and PCs, (the world is not just UNIX).
- + Information producers should run their own server (Mac, PC, UNIX box) rather
- than rely on the "computer center" to maintain their data for them.
-
- And so, things we'd like to avoid:
-
- - Complex back-and-forth option or feature negotiations between
- client and server.
-
- - Proliferation of document/resource types that are platform-specific.
- Excessive proliferation of types makes it harder to write a good client.
-
- - Breaking compatibility with old clients and servers.
-
- Finally, here is the most requested and needed wish-list for the Future of
- Gopher (in no particular order):
-
- 1. Retrieval (and perhaps display) of documents other than text
- (pictures, sounds, binaries, smells)
-
- 2. Encoding an access method (eg. gopher, ftp, wais, etc.) and an
- ident in a gopher list (directory) item.
-
- 3. Generic index servers (for UNIX, Mac, PC) that allow incremental
- update.
-
- 4. A gopherspace cruiser daemon that, coupled with #3 could be used
- for autoindexing at least all doc names in some subspace... a sort of
- Archie for gopherspace.
-
- 5. More formal process for enhancements (Oh no! Not that :-))
-
- 6. Primary server organization questions and conventions.
-
- 7. Mechanism for identifying and contacting the administrator of a server.
-
- 8. ID and password to allow controlled access to a document.
-
- 9. A "put this document on the server" facility. Note: This is
- pretty contrary to the basic Gopher philosophy: "You got data? Run a
- server off your desk."
-
- 10. Generic "prompt user, get input, feed it to a server-based
- process, return output to the user" facility. Note that this will
- subsume #8 and #9 above. It could be used for password protected
- stuff, ftp, submitting documents, etc.
-
- 11. A push for easier resource discovery, server communication (called
- by various names). Perhaps simple, static-list based global indexing
- of document names on "all" gopher servers. Note: This must wait for a
- good #3 & #4; we're working on that.
-
- 12. Single efficent data transfer method. Probably a headed table
- (length count + data block)
-
- 13. Alternative representations of document types while preserving the
- lowest common denominator: TEXT.
-
-
- - The Gopher Team.
-