We ship code All the way We are coders Night and day

Tap and type Click and slide Drag and enter Show and hide

Merge complete Submit request Boot develop Run the test

Reference call Document Optimising Memory spent

Layout render Draw the frame Style the model Symbol name

Open close Abstract tree Brace the bracket Whitespace free

Worker work Run parallel Threads combining Terminal shell

Build component Catch exception Nest together

Memory buffer Manual manage Scope enclosure

Binary value Floating point

Lambda function Template string

Undefined

Not a number Not a number Not a number

Null Null Null Null Null

Overflow

...

We ship code All the way We are coders Night and day

Hope you don't mind a quick remix :)

strudel.tidalcycles.org/?zNF1PMnnvgS1

Oh my god AMAZING

are you going to post it on fedi?

wasn't intending to as it's not my poem, but feel free to !

not sure if the speech synth works in all browsers

A version with a tune by felix roos added strudel.tidalcycles.org/?KWrkRVHQboWo (link updated with bugfix)

Wow that's so fun! I need to get involved in this kind of thing. Everything I do is MIT license unless otherwise stated, feel free to do anything you want with it! :)

Yeah I had to switch to chrome to get it to work. works great in chrome though

That’s interesting. I’ve always wanted to integrate those ideas into my subsequent work but it never fit. Maybe someday…

I’d love to see how you model conditionals in a spreadsheet.

Here's a recent one. It's from a function that determines the state to transition to in a state machine.

The C code had grown organically for a while and I noticed some duplicated branches, so I converted it to a spreadsheet approximation of a schematic table to be able to see which transitions I could dispatch most easily (like column B and C, then F, because they depend on the fewest values / have the most blank cells). I moved those branches up front in the C code.

It helped me simplify the function in the way that I wanted without any existing tests failing, and (surprise) found a bug that I added a new test case to cover. I don't think I could have done that if I'd only worked with the C code (because I tried and failed 😅).

I remember now that I had to reorder the rows a few times before the structure fell out nicely. If I put "is_active" below the two classifications, for example, then things doing coalesce as well because I'm laying the table out by hand.