For years now, I have been thinking of programming languages like human languages. The two have a lot of features in common. My wife is a bit of a linguist and can speak conversationally in multiple languages. I have never had much luck with human languages, until after I had learned a few programming languages.… Read More »
I stumbled across this blog post (https://steveklabnik.com/writing/rusts-golden-rule) a short time ago. The author spends a bit of time on a concept that is pretty much normal for compiled languages, and almost completely missing from dynamic languages: function signatures. In particular, he points out that Rust’s requirement that we are explicit about function signatures means that… Read More »
I was listening to episode 2.8 of the Lost Terminal podcast (a science fiction story in podcast form). One of the characters questioned whether a system they had encountered was an AI or just a Markovian Spirit. This is such a wonderful expression for something that appears intelligent at a glance, but on closer examination… Read More »
A 4.5 year old tweet about serverless: https://twitter.com/kylerankin/status/1067576106352427008 It’s much easier to understand #serverless apps once you realize they are just CGI scripts from 20 years ago, but with proprietary interfaces so you are locked into one provider. #devops — Kyle Rankin
Repeatedly in my career, I’ve come back to the parallels between writers and programmers. I found this article from an old note about Managing developers as writers: https://medium.com/coaching-notes/youre-not-managing-a-team-of-software-engineers-you-re-managing-a-team-of-writers-b263d3a10cc7
It probably shouldn’t surprise me, but I am often taken aback by people’s misunderstanding of the Burden of Proof. The normal failure mode is “I am presenting this new thing and you need to prove it doesn’t work (is not great, etc.)”. The correct use of Burden of Proof is a requirement that the new… Read More »
https://arxiv.org/abs/2304.15004 suggests the some of the seemingly intelligent behavior of LLMs may be more a result of choice in how to measure it, than an actual phenomenon. Definitely need to read and understand the whole article for balance. Plenty of breathless optimism out there already.
Following up on the different types of IDs from last time, it’s often useful to have an ID that can be sent in a URL. The unique ID may not be appropriate for many reasons. Many systems use something called a slug for this purpose. A slug must be unique, but can change under some… Read More »
Several times in my career, I’ve needed to deal with data structures that lasted more than a short period of time. Sometimes they were actual objects, sometimes just collections of data that belong together without any actual behavior. Many times these structures needed to be stored somewhere. When working on projects like that, I find… Read More »
I need to write up a rebuttal of 5 line code principle. My experience has been that it is sometimes useful to extract a concept as a method when it makes sense by itself and can be named. If you can’t think of a good name that covers the concept, then you don’t have a… Read More »