Website

Dependencies

You'll need to install a few dependencies to build the website; the following command will install all dependencies if you're using a Debian-based system:

sudo aptitude install python python-libxslt1 python-libxml2 perl libxml-dom-perl git-core tidy

Checking out the website

The website is managed in a git repository.

Any member can checkout the website using git-clone:

git-clone caffeine:/users/www/www

This creates a directory called "www" that contains the entire website and its history. If you want to make changes to the website, you'll need to ask someone on the Systems Committee to add you to the "www" group. You should also subscribe to our git mailing list.

Testing your changes

You should test any changes you make before pushing them out for all the world to see. The easiest way to see your changes without committing them is to create a symlink into your public "www/build" directory. For example, say your local copy of the website lives in "/users/$USER/csc/www". You could symlink "/users/$USER/csc/www/build" to "/users/$USER/www/csc". Do this by typing:

ln -s /users/$USER/csc/www/build /users/$USER/www/csc

Then you could see your copy of the website at this URL:

http://csclub.uwaterloo.ca/~USERNAME/csc/

Committing changes to the website

If you add a new file to the website, you can add it to git:

git-add file

To delete a file type:

git-rm file

And to rename or move a file type:

git-mv old-file new-file

Once you're done modifying the website you can commit your changes:

git-commit -a

The above command only commits the change to your local repository; you need to push the changes out to the master repository:

git-push

The above command will also automatically rebuild the website and send out an email to the git mailing list with the details of your change.

Building the website

Once you've checked out the website, you can build the website by running:

make

This will build the website into the build directory. When git-pushing, the website will be built into /var/www-csc.

You can remove the generated files by running:

make clean

Modifying the website

XML

The website's data is contained in a series of XML files. A series of XSLT files are used to transform the XML data into HTML files. If you want to add new information to the website you'll want to modify the XML files.

Some notable XML files:

• events.xml - past and upcoming events
• news.xml - past and upcoming items
• media/index.xml - media items (e.g. talks)

XML should be fairly easy to work with. Some important things to remember:

• All opening tags must be closed.
• All tags should generally be in lowercase.
• Tag and attribute names and case-sensitive.

If you want more information on XML, the following are good resources:

• The XML specification: http://www.w3.org/TR/2006/REC-xml-20060816/
• XML tutorials and general reference: http://www.w3schools.com/xml/default.asp

XSLT

If you want to change the HTML layout you'll want to modify the XSLT files. Almost all XSLT files are located in the 'xsl' directory. We are using XSLT 1.0 and a python front-end (xmlproc.py) to libxslt.

CSS

If you want to change the style of the HTML you'll want to modify default.css. We primarily target Firefox 2.0 but every effort should be made to use CSS that renders correctly in Internet Explorer 6.0 and other browsers. You can test various Internet Explorer versions from Linux by installing ies4linux.

Website vs Wiki

In order to keep things organized, the following guidelines should be followed.

Website:

• Information not likely to change frequently (e.g. services, general information)
• Information that should only be modifiable by the exec
• Official documents (e.g. constitution, policies)
• Current and past executive
• User and club lists
• News, events, and media items
• Anything we want non-members to see when they first encounter our site

Wiki:

• Information that changes frequently
• Information that any club member should be able to add to or modify
• Exec and office staff guides/manuals
• System and hardware documentation
• Past funding opportunities and advice (assuming this can be publicly posted)
• Contacts (e.g. CSCF, SCS) (assuming this can be publicly posted)

Also note that we can protect various pages if they shouldn't be edited in general.

Conventions

Email Addresses

Email addresses should be wrapped in <email> tags. For example, if you want to include the email example@csclub.uwaterloo.ca you should include the following xml snippet:

<email>example@csclub.uwaterloo.ca</email>

Leading and trailing whitespace will be stripped so it is safe to do this:

<email>
  example@csclub.uwaterloo.ca
</email>

Currently this just generates an standard mailto link. In order to prevent spammers from harvesting address, I want to generate an image that contains the address.

My current plan is to generate the email images while running make. This would be accomplished by creating a csc;email function that generates and saves the image to http://csclub.uwaterloo.ca/images/. The image name would be generated from HMAC(email, key), where key is some secret key/phrase.

Feature Requests and Ideas

• Possibly move club list onto wiki.
• Auto-generate exec page; this requires putting exec positions in LDAP.
• Convert cellspacing and cellpadding to CSS equivalents.
• Make RSS and ICS feeds more prominent.
• Replace or compliment /events with a web-based calendar?
• Replace all email addresses with generated graphics or javascript obfuscation?
• On media pages, get rid of trailing comma.
• Replace <br>'s with <p>'s wherever it makes sense.
• Replace <strong>, <b>, <i>, etc... tags with CSS.
• Highlight (e.g. make left bar green) current page in directory header
• Rename exec page to 'Positions' and make tables of positions.