Brief history

The following is a piece of correspondence concerning the history of the Shakespeare site which provides a brief timeline of its history.

Since this email was written I have located earlier logs from the site showing the search engine was indeed up and the site serving pages in October. The MIT site came online in December 1993, see the December 10 entry at the NCSA'a What's New Archive

I never received a final copy of the article mentioned. I find it disappointing that the MIT site has gone on to claim it is the web's first Shakespeare site.

From matty Thu Nov 19 16:00:16 1998
Subject: Re: history of shakespeare site
In-Reply-To:  from "Peter S. Donaldson" at "Nov 17, 1998 10: 9:29 pm"
To: (Peter S. Donaldson)
Date: Thu, 19 Nov 1998 16:00:16 +1100 (EST)
MIME-Version: 1.0
Content-Type: text/plain; charset=US-ASCII
Content-Transfer-Encoding: 7bit
Content-Length:  6456

> Dear Matty Farrow:
> I'm interested in when your Shakespeare site first went up, and any bits of
> history concerning it that you may wish to share.  The immediate context is
> an article to be published in Shakespeare Survey where I credit my former
> student Jeremy Hylton with the first Shakespeare Web site -- but he's not
> sure yours wasn't first! -- The broader context is a study of
> media/Shakespeare relationships -- from the first use of Shakespeare as a
> demo for the telephone in 1876 until the "virtual reality" Midsummer
> Night's Dream in April 98.  I'd love to know more about your early
> involvement with Shakespeare on the Web -- but if you're busy, even the
> date when it went up would be very helpful.

Professor Donaldson,

Thanks for your interest.  I've just been searching back through my
history and web log files to try and put together some kind of chronology
for my Shakespeare materials.

I must have downloaded the Moby Shakespeare sometime on or before August
29, 1992.  That's the date on the oldest files in my raw text directory.
More significantly it's the first appearance of my `bard' programme.
Bard was a perl script (and I've just looked at the comments in that
file and, wonders will never cease, I actually _did_ comment it: Late
August 1992) which enabled one to search for words and supplied some
surrounding context for the hits.  Sound familiar?  It eventually became
the basis for the web engine but that in good time...

I started bard very soon after fetching the texts since I found I had
this corpus but no easy way of finding anything in it.  I advertised it
to another postgrad that evening (the 29th).  It ran something like this
(to look up estimate in the sonnets):

~matty/bin/bard -context 20 estimate sonnets

Around this time we didn't have the web but we did have a modified finger
daemon I'd written which allowed remote users to use the finger programme
to provide arguments to a .finger file in the user's home directory which
could process these.  My .finger file was incorporated to include bard.
On September 7 I must have advertised this service to some remote users
because I've found this exemplar in some outgoing mail from that date:

% finger bard:0:estimate:sonnets%matty

This would have looked up the word `estimate' in the sonnets and returned
the number of the sonnet.  Here was one finding "Which works contain
`betime', and where are they?"

% finger bard:1:betime:.*%matty

I'd already made available some of my sonnets and the jargon file using
this mechanism so it seemed the next step in making bard useful to people
not immediately connected to the system, i.e. on remote systems.

Bard then spent a useful year as a command-line tool and remote searching
mechanism for others (via finger) until in August 1993 I set up a web
server on our research group's machine.  I remember telling the other
pgrads in my room about this `web thing'.  Certainly our supervisor
couldn't see the point in it claiming `it would never catch on!'. ;-) In
playing around with CGI scripts, bard seemed the most likely candidate
for conversion.  It was a useful tool which I wanted to make available
to a wider audience--first via finger then via the web.

The first hit I have on my Shakespeare pages recorded was - - [15/Nov/1993:20:17:37 +1000] "GET /Users/matty/Shakespeare/index.html HTTP/1.0" 200 3592 

and the first on the search engine (which was then really a very simple
wrapper around bard) - - [15/Nov/1993:20:19:49 +1000] "GET /Virtual/fsearch?context=14&text=estimate&play=sonnet HTTP/1.0" 200 1000 

I'm sure our server was up before this date but I think we changed servers
and log file formats and the earlier logs were discarded.  These are the
first entries in the log file and I know we had a web server before I
put the Shakespeare materials up and I would have expected other entries
before them.  Certainly I was testing a web-ready version of bard on
October 21,

bard -html -urlquery 14%3aestimate%3asonnets

having added the mechanisms to parse a web query and provide results
in HTML form.  I'd linked something into my Shakespeare directory the
evening before:

ln -s /usr/pgrad/matty/lib/html/Shakespeare/

(I'm sure I was supposed to be working on something else at the time).
October 20, 1993 is the earliest date for which I can find documentation
(from looking at various sources: history files, dump tapes, &c.) for
the presentation of the Shakespeare texts I had on the WWW.  (The dump
tapes seem to back up my supposition that I fetched the original archive
on August 29, 1992.)  From the `ln' command I already had a Shakespeare
directory at that time amongst my web pages.  I had made my own sonnets
available on the 16th which I imagine was part of me setting up my
web pages.

The search form and texts were tweaked over the next while.  I added
HTML versions of the plays and cleaned up bard somewhat to make searches

In June 1996 we were approach by the NSW Department of Education to
help them prepare a CD to help students and staff practice searching
the Internet.  Since a connection to the 'net couldn't be guaranteed
they wanted a reasonably well-known corpus to search and an engine which
behaved similarly to search engines such as Alta Vista.  During the next
few months the new version of the search engine was written in Python
and an index of the texts built.  This went up sometime on or before
September 19 of that year.

Since then it's been a process of tweaking things and optimising the
search engine and watching the To Do list grow and grow as I think of
improvements (and improvements are suggested).  I didn't have the time
while I finally decided to finish my thesis to implement them nor have
I since.  There have been various awards and appearances in print media
(such as books and newspapers) of which I have some record but I've
never really kept track of such things assiduously.

If you'd like anything further information feel free to contact me.

I'd be very interested in reading the final article.

Dr James Matthew Farrow                 | "For in that moment I beheld the ruin
                                        | of my existence.  My world fell dark
Farrow Norris      Tel: +61 4 1724 3183 | and my life became a shallow dream.
                   Fax: +61 2 9546 4468 | `Odi et amo. Excrucior.'" - Tlindah