About this website

Notes about this website. See also: Colophon

TODO

• Should probably make this into a nix thing … trying

• Currently, I’m using imgur to host the images embedded in many pages. I’d like to have a backup plan if the “free” service goes belly up.

  • extract all the imgur urls.
  • download all the images
  • push to new location maintaining same file names
  • rewrite the content with new URLs.
  • use cloudflare r2 storage, it has generous free tier.

sitebuild

A python script replaced the nanoc setup on Aug 31, 2019

nanoc – Aug 31, 2018 to Aug 31, 2019

• Ruby 2.5

sudo apt install libsass-dev sassc ruby-dev
sudo gem install bundler

bundler install --path bundle/vendor

Old Hakyll build steps

This site ~~uses~~ used Hakyll to compile markdown files to HTML using pandoc.

Steps (TBD:document):

• Installation of Haskell
• Building the site binary
• Building the website
• Compressing the HTML files using htmlcompressor.jar
• Syncing the changes to the website

Interesting/Inspiration

Consider Paul Graham’s index page. While his front page is minimal, the site contains many, self-contained, indpendent pages. Eg: 1974 Porsche 911 Models. Currently, I’m writing all pages (only under / directory) into allpages.xml. However, this can be presented better. Perhaps multi column or even multi page pagination?

–

this and this are a nice look for long form articles.

–

“The landing page of http://deconstructconf.com is now 99 kb gzipped in a single HTTP request. It contains sixteen images and one font. Greetings. The titles are all SVGs automatically traced from generated bitmaps.” – via.

That is a nifty way to generate title images that are scalable.