◊(Local Yarn Code "Update of "Tools and Methods"")

When it comes to computers, here’s how I like to work:

• I like the things I make to look good and right. If they don’t, I get frustrated and dissuaded from creative work.
• I prefer to put in a lot of thought and extra work on the front end in order to get repeatable, high-quality results with a minimum of fuss later on.
• I don’t have time for bullshit maintenance work: stupid repetitive tasks that merely keep things working vs. actual creative work.

These priorities mean that I try to select tools (computers, software, workflows) that

• are reliable
• don’t depend on any one company or person
• can be automated
• have appropriate simplicity for the job

Pollen

I use Pollen as the framework for organizing writings and publishing them as HTML and as printed books.

Pollen gives me a clear framework for defining my own markup language, and for specifying exactly how that markup gets translated into HTML, LaTeX, or any other format. This means I’m not dependent on other peoples’ Markdown processors and I don’t need to rely on flaky hacks or beg them to implement functionality I need.

Pollen’s author describes his rationale for creating it:

It oc­curred to me that what I wanted was not a sim­ple but reg­i­mented sys­tem like Word­Press—it just wouldn’t let me work with the so­phis­ti­ca­tion and de­tail I needed. In­stead, I wanted a flex­i­ble tool for de­scrib­ing com­plex HTML & CSS lay­outs with sim­pler, high-level notation.

In short: I wanted my own pro­gram­ming language.

Pollen is free software (LGPL licensed).

Fossil

Version control is inherently complicated, but its benefits make it an essential part of any programming project, even a small one with only one developer.

My reasons for picking Fossil for this job have to do with my personal tastes (see above) and specific reasons described in Why Fossil?

Web Server

Because Pollen produces static HTML files, the website is pretty agnostic about most aspects of the server stack.

• Server: I use Apache on Linux. It’s easy to install on any distribution, and Apache has more than adequate performance to serve a static site with minimal or no tuning.

• Analytics: I rely solely on Apache log files and tools like GoAccess that can make use of them; no JavaScript and no third-party services.

• RDBMS: No database client/server system is needed.

• Backups: The use of static HTML along with distributed version control (Fossil) mean that I always have a local copy of everything. My local copy (along with everything else on my personal computer) is backed up locally and to a cloud backup servce; and if the hosting service provides automated backups, I use those.

For certain mildly interactive aspects of the web site, such as contact forms, I am not yet certain whether I will use PHP or if I will attempt to implement them in Racket for purity’s sake.

The server configuration is maintained with Ansible, so it is the work of a few minutes to spin up and configure a new server if necessary.