Agenda, meeting of Wednesday, 27 May 1998

Organization

• Assignment of minutes
• Confirm that everyone purchased a notebook
• Summary of notebook format (by Becker and Rebelsky)
• I've gathered the information from project clio in rough form.

Discussion of Readings

• Key ideas
• What is applicable? What isn't?
• ...

Planned File Format

• What data will we record?
• What kinds of log files? Possibilities include
  • One file per time period
  • One file per person
  • One file per "session"
• What type of information to record?
• Note: we should probably consider the standard http log file format and consider what is good and bad about it.
• Do we want to do post-processing of these files to insert or remove information?

Architectures of recording system

• What are some of the tasks we need to do in the recording system?
  • Transform pages to include
    • recording capabilities
    • modified links
    • ...
  • Record appropriate events
  • Authentication ...
• How do we record?
  • A cgi script can record entries
  • A small Java applet can post regular messages to a server "still on this page"
  • Cookies can provide some session-level information
  • Query strings can provide some session-level information
  • JavaScript scripts can tie parts together
• How should the recording system go together? I can suggest a few simple possibilities.
  • We can run everything through a cgi script that does some logging and transformations. For example, it might translate <a href="foo.html"> to <a href="clio.cgi&url=foo.html&session=...">
    • Makes it easier to do authentication at each page load.
    • May be costly in computing time.
    • Can be used for external pages as well as internal pages.
  • We can transform ("compile") each page to another page with some modified components.
    • Less computing intensive
    • Better urls
  • We can transform ("compile") each page to a cgi script
    • Easier customization
    • Decent urls
    • Computing intensive
  • We could use frames (ugh).

Miscellaneous Goals

These are near-term goals; clearly we won't be able to do many of them by our next meeting.

• Some analysis of the student notes on Project Clio. I have a tool that can help with some of these analysis, so whoever is doing this should speak with me.
• Install some alternate hypertext systems for ideas
  • HyperWave
  • WebCT
  • Stanford annotation system
• Read more about httpd logs and additional possible information
• Figure out how to write Java-based cgi scripts
• Gather additional papers on event logging
• Investigate other competitors
• Work on small systems to test alternate architectures (may be SamR's responsibility)