To: wais-discussion@think.com
From: Brewster Kahle <brewster@Think.COM>
Subject: WAIS-discussion digest #53
Forum On Wide Area Information Servers and Electronic Publishing
Brewster Kahle
Contents:
sources in foreign languages (fwd) (Elham Chounet)
WAIS clients and the directory of servers (rhys@cs.uq.oz.au)
Re protecting publishers (Ken Kahn)
Please send contributions to wais-discussion@think.com
----------------------------------------------------------------------
From: chounet%FRULM63.BITNET@pucc.Princeton.EDU (Elham Chounet)
Subject: sources in foreign languages (fwd)
Date: Mon, 11 May 92 12:31:06 MET DST
Hello,
While using WAIS to create a source for our department library, which contains
essentially french and english documents, I was very favourably impressed by
WAIS, but was faced by the problem of our french accentuated characters,
and could not process them in a way satisfactory for me.
Here are two small extensions I have made, both related to processing sources
in languages other than english.
First, I have add an option to "waisindex", usefull while indexing sources
in languages containing accentuated (8-bit) characters: applying it allows
the subsequent search process in sources containing accentuated characters to
run correctly.
Second, I have parameterized "waisindex" by the stoplist which determines
a list of words that are ignored during indexation, this list being obviously
specific to each language (maybe even specific for some sources?).
If interested, get the patch by anonymous ftp from
snekkar.ens.fr (or 129.199.104.3)
at file
/ftp/pub/unix/appl/wais-8-b4.patch.tar.Z
(which includes a patch-README file)
All suggestions and bugs are welcome. (chounet@ens.fr)
Hoping this may help someone!
Elham Chounet
------------------------------
Date: Fri, 15 May 92 11:15:39 +1000
From: rhys@cs.uq.oz.au
Subject: WAIS clients and the directory of servers
Hello WAISers,
A friend of mine and myself were having a chat about WAIS this morning and
he brought up a particular problem with WAIS clients, and xwais in particular.
Namely that of searching the list of sources for keywords.
Hey, what about the directory of servers I hear you cry? Well, what about it?
It's quickly becoming a solution to the wrong problem, at least locally here
at the University of Queensland. I automatically fetch all sources from
quake.think.com each week, and store them locally for our users. Hence,
connecting to the directory of servers is supposedly not necessary.
Currently, if the user wishes to perform a keyword search on the list of
sources, they must connect to the directory of servers and do so, to find
which sources of the local ones are relevant to their query, and then choose
the local ones, increasing network bandwidth in the process. Alternatively,
they can manually grep the source files, which is hardly friendly enough.
So, what's the solution? I could create a local directory of servers that
indexes the sources I fetch, so my users can search that, but this still
requires a two-stage search which is clumsy. What would be better is a
keyword search built into the "Add Source" function of xwais (and similar
functions of other clients) that takes care of the search in the client and
adds automatically all those sources that match the search. That is, combine
the two steps of searching and adding the found sources into one.
The keyword search doesn't need to be as complex as the full WAIS search
engine, since the sources have a fairly simple text format, and can be indexed
"on the fly".
As the number of sources increases, the directory of servers will not be
enough - more smarts need to be built into the clients to permit searching
the sources that are locally available. WAIS already assumes that clients
are smarter than dumb terminals, and rightly so - we should take that further
and build some of the initial source-finding operations into the client to
alleviate the burden on the user.
Cheers,
Rhys.
--
Rhys Weatherley, University of Queensland, Australia.
rhys@cs.uq.oz.au "I'm a FAQ nut - what's your problem?"
------------------------------
From: Ken Kahn <kahn@parc.xerox.com>
Subject: Re protecting publishers
Date: Sun, 10 May 1992 21:43:07 PDT
Apropos the recent discussion about how publishers can protect their
goods if they are available in electronic form, Dave Patterson from
Berkeley was asked about that in a recent PARC Forum. He mentioned a
solution he had heard that seems to make sense to me.
The idea is that your terminal has a key and decryption abilities. When
one purchases an electronic document, one gets it encrypted for the
terminal being used. The bits can be saved and read multiple times on
that terminal but to read it on another terminal one must purchase a new
copy.
While he didn't mention it, I assume this could easily be generalized
for printers for those who prefer paper. The document one buys would
only be printable on that particular printer. It is true that one could
print lots of copies on that printer, but then one can use a copier
today.
Patterson indicated that the goal not would be perfect protection but
protection from illicit copying by most customers. Sophisticated
high-tech hackery might still be able to work around the scheme. The
analogy would be the way cable TV works today.
I wonder if publishers annouced that they would make lots of documents
available were such terminals (and printers) to exist, whether terminal
manufacturers would soon start making them. I assume that extra cost
isn't very high (or if it is that it won't be a couple of years).
-ken kahn
------------------------------
End of WAIS-discussion Digest 53
************************
-------