Minutes for Thursday, 27 May 1998

• Recorder: Kevin
• Begin: 1:15
• End: 1:55

Administration

There is a version of Perl installed in /home/rebelsky/perl/bin/perl. There is probably a library for helping retrieve remote web pages in a lib directory under the above path. (Look in /home/rebelsky/perl/lib. One promising location is /home/rebelsky/perl/lib/site_perl.

We now have a group on the MathLAN. It is called hyper.

Reports

Kevin implemented a version of Clio in Perl. It currently takes a local file and prints it out. It retrieves the filename from the query string. Kevin expressed concern over how we would replace the urls but after discussion this did not seem to be a problem.

Raphen implemented a version of Clio in Java. It was easy to have Java go out and grab a webpage elsewhere. He was working on finding the links and replacing them with the appropriate revised link. Problems occurred with links that went over one line. Took Raphen about 110 minutes.

Hilary was working on sticking something into the header and the body of an html file. She did this in Perl. Rebelsky commented that if there is not head it should just enter a head with the new line.

Sarah was working on the same project as Hilary in Java.

Division of Labor

How do we divide labor on this project?

• One person on logging system, etc.
• We divide logging into small parts.
• Have consultants to larger parts.

Proposed division of labor:

• Kevin expressed interest in the logging system.
• Raphen is making a formal specification of our log files.
• Sam orders Perl books. :-)
• Sarah expressed interest in the annotation project.
• Hilary expressed interest in the identification subsystem.
• Hilary and Sarah are going to install existing annotation and web authoring systems and evaluate them.

Places for Hilary and Sarah to start:

• Hope College, Mike Jipping might be a good person to start with. (jipping@cs.hope.edu)
• WebCT, go to Rebelsky's origin there should be a link. http://homebrew.cs.ubc.ca/webct/
• A project at Cornell named Huttenlocher.
• Terry Winnograd at Stanford does some research. There is an article on this with some URLs.
• Get and install Hyperwave which is a hypertext publishing system. You can probably find it at http://www.iicm.edu (a later look suggests that it can be found at http://www.hyperwave.de

For Tomorrow

• Think what the identification subsystem is going to look like (what functionality it will provide)
• Draw a picture of your desired architecture.
• Work on your component of the project.

Misc.

How do we store the id and session number and if they don't have one are we going to bring up a page. Should we let the session id time out if somebody tries to use a session they started 4 days ago. Should this be part of Kevin's responsibility or somebody else?

Rebelsky noted that we can combine the two architectures by creating small cgi scripts that simply call clio.cgi with the appropriate page.

A question occurred about what happens if the remote host sets a cookie? We decided to differ this question until later.

We went over a quick overview of what a server sends to a browser. It sends the following:

• Content-type: text/html
• Cookie:
• Location:
• etc...