For the last few weeks, my normal hobbies have been preempted by my obsessive refactoring of an old SmugMug keyword setting program. Many years ago I wrote a small application that helped me maintain meaningful SmugMug keywords. I used this application for years until inevitable software changes broke it. Replacing it with something better has been on my to-do list for ages.
I kept putting it off because I knew the tools I used in the past, mostly J and VB, were not well suited to REST API programming. Yes, I know just about any programming tool can be forced to serve, Turning completeness guarantees it, but I have grown weary of programming to prove tiresome points. If a tool makes something more difficult or tedious than it should be then change tools!
The downside of changing tools is that it usually results in learning yet another programming language. I always try to avoid this! I briefly considered languages with decent REST support that I already sort-of-know1, C#, JavaScript, VB.Net and decided that proper 64 bit versions were either too expensive (C#, VB.Net) or that I didn’t want to stare at ugly bracket infested syntax for weeks. I’m talking about you JavaScript. Even your fanboys admit that you’re an ugly duckling.
I eventually decided to go with Python. I’ve always found Python code more readable than many other programming languages. This is a widespread opinion. Python is free, open source, and comes with a large diverse set of tools. One of the best tools to come out of the Python world is Jupyter. Jupyter is the first public domain open source literate programming tool that has gained a large following.
I’ve been a fan of literate programming ever since I read Knuth’s seminal book. He used the technique to create some of the best program documentation ever created. I always wondered why literate programming never caught on. I suspected the basic problem was simply that many programmers are not particularly literate. Well, Jupyter is proving me wrong.
Jupyter is certainly helping me write and program with such clarity that all my ideas seem trivial.
Compare this notebook, (use the first link for the best layout), to my earlier blog post about setting SmugMug print size keywords.
- Nbviewer: Setting SmugMug Print Size and Geotag Keywords with Jupyter and Python
- GitHub: Setting SmugMug Print Size and Geotag Keywords with Jupyter and Python
- Be highly suspicious of people who claim to fully understand any programming language. Only delusional nitwits would make such a claim for any natural language. Does anybody, even luminaries like Shakespeare, truly understand English? We all sort-of-know our mother tongues and if we’re honest, we’re continually surprised by how others make use of it. Literature, it’s a thing. The same holds for programming languages. Every day I’m surprised, by unusual, stupid, silly, clever and freaking brilliant code fragments in programming languages that I have used for decades.↩︎
Analyze the Data not the Drivel 