The goals of a redesign are:
- Base it on Git
- This allows me to make updates to the website from any computer, without having constant SSH access (which is a corporate security violation in many places)
- Make it more attractive for others to read and navigate
- Add some basic templating, so that there's a header for each page, that includes: 1) a title, 2) breadcrumbs, 3) a "history" link for that page
- Make it easier for others to separate the wheat from the chaff
- Add YAML metadata to each page that notes that it's "high-polish", "low-polish", or "personal interest only"
- Generate my own index pages. On the index pages, when the user hovers their mouse over a link, it should pop-up a brief description of the page. (this brief-description would also be part of the YAML metadata)
- Make history easier to browse
- History should be directly tied in to the Git log: 1) The front page should be generated from the git log, and should only be updated when a git push happens. 2) Each page should have a link to display that specific page's history.
- The ability for users to view individual-page-history is important. The bliki concept means that it's not always clear to readers HOW MUCH content has changed on a recent update; it could be as trivial as tweaking punctuation, or as large as an entire page rewrite.
- Provide some minimal markup functionality
- <h3>'s and such should be automatically given an <a name=...>
- a <title> should be automatically generated from the YAML metadata, along with the same string wrapped in <h2> at the top
- (optional) being able to link to wikipedia or search.cpan.org might be nice
- Auto-link URLs within .txt files
things that may not make the cut, because of amount of work reqiured, or the technical details haven't been nailed down
- have the ability to track page moves and directory renames, and point the user to the new page
- 404 JS handler — that's the way to go
things that may not make the cut, because of bloat, but are fairly easy to implement
nice features that come automatically with a "baked" setup
- able to compile SASS/Less and CoffeeScript (to CSS and JS, respectively)
- Provide some way to generate HTML tabular information from .yaml data files. (see these examples)
- I'm currently heading towards using more and more client-side jQuery-based transformations... is there a good way to move these server-side?
- I regularly resort to the command-line to search for things. These search features should be available to ALL readers.