Having given up on Wordpress, a static
web site seemed like a better, more modern idea for my wife’s site
content consisting of mostly images of her fabric art.
The old site was PHP-based - very basic templates that were adequate for the
turn of the century. But now that was an embarrassment for me, the original
integrator who claimed to have some web development capability.
So last August I began the journey of learning how to use
Hugo and dabbling with
Golang. Hugo’s development timeline, as well as
site generation capability, goes at internet speed and I found
the community very helpful. I
experienced a few false starts wrapping my head around Hugo’s concepts
but by December was hitting my stride on converting the old content
into Hugo’s markdown and filesystem organization in what I thought
would be relatively easy for even a non-geek to use.
The bulk of the conversion process was fairly easy once we had worked
out that each art work would become a ‘project’ in the new scheme,
with categorization into ‘galleries’ and tagging accomplished by
leveraging Hugo’s taxonomy scheme. Some of the conversion I sped up
using Pandoc to help with the markdown-ification
of her original content.
As of 2018 the fabric artist is editing
Markdown pages using
the Atom editor with a few plugins to provide some
syntax highlighting and Markdown help; a little Mac app (created with
Platypus and a bash script to
run hugo to locally create and preview her web pages automatically on each file save;
Sourcetree to commit her
changes. All without her having to be really aware of the geeky details.
Once the artist is satified, she uses Sourcetree to push the local git repo
updates to a remote git repo hosted on an Amazon EC2 t2.nano
instance. There a git post-receive hook causes the Hugo installation
to automatically rebuild her web site contents in approximately seven
seconds. That then gets served via a Caddy
Easy peasy - once all the pieces have fallen into place.