The whole website is location independent, i.e., the directory structure can be copied from one webserver to another without having to deal with broken internal links: all internal links are relative to eachother. E.g., I have three copies of this website. You can transparently visit three (almost) identical websites:
- http://www.godot.be/site/index.html (tertiary mirror),
- http://godot.studentenweb.org/site/index.html (secondary mirror),
- http://www.esat.kuleuven.ac.be/~decockd/site/index.html (primary mirror).
No other tools or bash commands/constructions than those commonly available on almost all machines have been used on pure text to produce my marvelous (ahum) website.
An exhaustive enumeration: basename, cat, cd, cp, date, dirname, echo, find, for-do-done, grep, if-then-else-fi, mkdir, mv, rm, sed, seq, sort, test, time, touch, uniq, wc, while, zip.
The scripts also launch a few (optional) tools to improve the website's overall quality, look and feel, and to make sure the html pages are consistent with the current (x)html standards:
- checklink, a great but time-consuming tool which nicely discovers dead links, but it produces rather many false positives, e.g., on https urls
- apcalc, an advanced arbitrary precision calculator
- fortune, adds some random fun to the pages
- ispell, to reduce the number of misspelled words
- perl, to determine the dimensions the various logos
- tidy, makes sure that the pages contain syntactically correct html
- wget, to eliminate false dead links reported by checklink