Ben Kovitz
To get around the dbm problem, where the script crashes if any page is longer than 1K, I modified the script to read and write pages directly to text files. There is no more dbmopen() or db% or wiki.dir or wiki.pag files. Instead of checking to see if a hash contains a key, the script now just does a -f on the name of page. If the file exists, then the page exists.
Report on the above-mentioned experiment: after an initial phase of a bit of editing, people mostly just let it sit there. I didn't publicize the wiki heavily, and it was a team of about 8 people.
Return to WelcomeVisitors