Damian’s notes

I rather wish every logging framework had at the top of its web site landing page a prominent link to the effect "Are you working on a project that already has SomethingLog4J logging? Maybe you want a _recipe for adjusting logging levels of a particular class_!"

Almost always I do not want the 10-page tutorial that starts with explaining the concepts of loggers, appenders, etc.,— I just want to be told where to find and how to tweak the config of an existing system.

Oddly named package for producing a diagram of your Django project's object model

Carry-propagation is expensive in modern CPUs because it reduces parallel processing, so adding 256-bit integers by switching to base-2**51 and deferring carry propagation until the end can counterintuitively be faster.

Simon Willison’s approach to creating template repositories in GitHub, with a Cookiecutter step to set the name of the components in the source code automatically when someone uses the template.

Django SQL Dashboard provides an authenticated interface for executing read-only SQL queries directly against your PostgreSQL database, allowing exploration and visualization of your data with minimal friction. Queries can be bookmarked and turned in to publish dashboards.

One of my bugbears is user databases that assume or impose first-name, surname fields, based on assumptions about how people's names are structured. This W3C note on personal names around the world outlines some of the ways names are understood around the world, and has suggestions for designing UIs accordingly.

Entertaining twitter thread on a bug (that wasn't a bug really) caused by changing keyboard-layout conventions and the accommodations made to support old keyboards on new computers.

Access-control lists are often set up to answer the obvious question "can X access Y to do Z?" but not other useful things like "What are all the Ys X can do Z to?" But this needs to be planned in from the start, or you will be sad later on.
(Twitter thread.)

We are used to treating databases as a pluggable back-end for storing inert data for processing in our code, but the PostgreSQL database has many features for slicing and dicing the data and doing analysis than one might think.