1. Setting up a web site
    1. Purpose: Publish information
    2. Set up a deamon ... offer http protcol.
    3. Does NOT offer ftp or gopher or telnet or news protocol.
    4. Allow users to publish their own information.
  2. Measure amount and type of accesses.
    1. Are people interested?
    2. In what are they interested?
    3. Are you overloading the system?
  3. Archicture
    1. Clients have web browsers.
    2. Clients have cache
      1. Not all requests get to server
      2. Some requests just ask to verify.
      3. Server cannot know all accesses.
    3. Connected by inet to web servers.
    4. Sometimes cache deamons come in between.
      1. Further screw up accounting.
      2. Seem not to work well (draw graph).
      3. Can get around firewall.
    5. Servers offer data.
  4. Requirements
    1. Permenant and fast internet connection.
    2. Your own IP number and hostname.
    3. Computer that’s up 24 hours 7 days per week.
    4. Fast CPU normally NOT required.
    5. Reasonable amount of disk space IS required.
      1. Should be on the machine with the http deamon.
      2. Sometimes (with user accounts??) that’s not possible.
  5. Setting up typical server.
    1. Choose Server (we choose Apache)
      1. Support CGI
      2. Support standard log format (they all do)
      3. Choose common one for support (or commercial one)
      4. Access controls
      5. Or chose the one already installed.
    2. Config files
      1. Everything in /etc/http/httpd.conf and magic.
    3. Create a /robots.txt file if you want.
  6. Test server
  7. Install useful CGI scripts (i.e. counter)
  • Accounting
    1. Access can be
      1. hit
      2. set of hits close together
      3. a site
      4. a kilobyte of data.
    2. Raw data stored in logs
      1. very volumous
      2. often compressed/summarized.
    3. Accounting done by freeware packages
      1. Pick on features/ease of installation
      2. Speed not escencial.
      3. output in HTML very nice.
    4. Notes about web browsers
      1. Some Offer users an HTML editor
        1. Best editor is ....
      1. Manage user web cache space.
        1. Netscape stores in ~/.netscape/cache
        2. Users cannot share cache space.
        3. Almost always root can clean this space.
  • How to handle offensive material:
    1. Ignore problem
    2. Police problem.
    3. Wait for problem to occur
    4. There are lots of what about racism/sexism?