Due 5pm Wednesday July 6
Your assignment is to write your own WikiWikiWeb engine. I would like to see it up and working at some URL that you send to me. I would also like you to turn in all your code.
A WikiWikiWeb is a collaborative hypertext database that allows users to add content using a very simple markup (no html) and to link from page to page solely by using words that are in CamelCase. The original WikiWikiWeb was dedicated to Design Patterns and Software Methodology, but later wikis have been bold enough to try and encompass the sum total of human knowledge. Anyone can edit most wikis, although Wikipedia also supports logins. The content contained in their collective databases generally grows quite quickly because it is so simple to add stuff. To learn more about WikiWikiWebs in general and the first one in specific, check out NewUserPages on the granddaddy of wikis. I have used wikis a lot for various projects, we used one in the department to collect information before submitting a grant proposal, and there is one on my home computer that I made for random thoughts and projects.
Implementing a basic WikiWikiWeb should not be too hard. Remember to guard everything against malicious input. Setting maxlength in the HTML code is not enough - those limits are merely hints to the browser, and are not actually enforced by the protocol. The requirements for the WikiWikiWeb that you will submit to me are:
\b[A-Z][a-z]+([A-Z][a-z]+)+\b<, >, and &, respectively (this requirement can be relaxed if you process the incoming text to avoid the dangers of cross-site scripting and malformed user input)Extra credit can be garnered by implementing features beyond this required set, including:
Implementing a wiki becomes a lot easier when you note that WikiNames are also valid (and
non-threatening) filenames. So if you store the text of each page in a
a file named for its WikiName, then you can make a directory containing a bunch
of files with wierd WikiNames as their filenames, and then you will know from
the filenames what pages exist and what pages do not. To find BackLinks, you
simply verify that the given string is a WikiName and then execute the command
grep '\bNameBeingSearched\b' * (possibly ggrep if you are on
Solaris) in the repository directory. and print out a list of all the files
that were found.
The are a few ways of doing all this. You could write 3 cgi scripts:
index.cgi for viewing content in edit mode or in view mode,
edit.cgi for saving pages, and links.cgi for
searching backlinks. These files will necessarily have a lot of code in
common because they are accomplishing similar tasks, so creating a shared
library that each file imports will probably save you a lot of typing.
Another approach is to put everything in one file, and have switches based on whether you are recieving a POST or a GET, and whether there is a particular type of operation you would like to engage in on the given page. Neither approach is innately superior, so do whichever one seems most intuitive to you. Comment your code well, and guard against malicious input.
I implemented my wiki using the "one big file approach", and it was a little under 200 lines - making this a pretty major project. You can see the results at http://www.cs.uoregon.edu/~peter/399/wiki.
Please ask me if there are any questions at all. peter@cs.uoregon.edu, or simply commenting below will all reach me immediately. I am also available on AIM as "jongleur peter". The most recommended course of action is coming in to office hours if you have any questions.
Turn your project in using the following form:
Questions? Answers!