Distractable (#1269)

Topics/tags: Overcommitment

Today I planned to work on the image library for CSC-151. Here’s how that went.

First, I started a musing about that library. After all, it’s better to get my thoughts in order before writing.

When I was halfway done with musing, I started exploring my various implementation options. I’ve been using Racket structs, but I decided to consider objects, especially since I was designing what looked like an inheritance hierarchy. I ended up sticking with structs.

Rather than working with images, I decided I needed to work on the color model. While the 2htdp/image library has a perfectly serviceable color model, I don’t like the name of that model because it uses color to mean rgb color. Of course, I’m going to use rgb to mean rgba color, so perhaps I shouldn’t complain.

While working on the revised color library, I decided that it was important to improve my documentation. That meant starting to work on a type predicate library [1]. That meant considering whether I should put those new procedures in the forthcoming CSC-151 reference.

Working on the CSC-151 reference reminded me that I should be writing tests [2]. And I tend to be a bit excessive on tests. For example, here are the tests for the new nonnegative? predicate.

(test-true "zero is not negative (exact)" (nonnegative? 0))
(test-true "zero is not negative (inexact)" (nonnegative? 0.0))
(test-true "negative zero is not negative (exact)" (nonnegative? -0))
(test-true "a small exact integer is nonnegative" (nonnegative? 3))
(test-true "a small rational is nonnegative" (nonnegative? 1/312231))
(test-true "a small real is nonnegative" (nonnegative? 0.00000001))
(test-true "a large exact integer is nonnegative" (nonnegative? (expt 3 100)))
(test-true "a large rational is nonnegative" (nonnegative? (expt 11/3 100)))
(test-true "a large real is nonnegative" (nonnegative? (expt 1.23 101)))

(test-false "-1 is not nonnegative" (nonnegative? -1))
(test-false "a small negative exact integer is not nonnegative" (nonnegative? -3))
(test-false "a small negative rational is not nonnegative" (nonnegative? -1/11))
(test-false "a small negative real is not nonnegative" (nonnegative? -0.0000002))

Writing these tests meant that I should be setting up a test infrastructure for the CSC-151 libraries. I took the easy way out: I set up a directory with separate test files for each procedure I wanted to test.

I also started writing reference pages for these procedures. Or perhaps I started writing the reference pages and decided those pages should also have tests. I forget.

Working on the reference pages reminded me that I wanted to find a way to identify deprecated procedures [3]. Why am I documenting deprecated procedures? Um, I’m not sure.

As I updated the scripts to build the reference pages, I realized that I had some other issues I wanted to explore, such as how I would handle multiple versions of a procedure [4]. So I started writing reference pages for the zero-parameter, one-parameter, and two-parameter versions of random. And I spent some time working on the underlying summarize documentation code.

Then I realized that I was being distractable. The only solution to being distractable is, of course, to write a musing about being distractable. So I started this musing. While writing it, I did some more of the tasks above.

At that point, I had to make dinner and clean up from dinner. By the time I was done, it was 7pm and I had not yet started writing the primary library. Oh well, back to colors. I finished colors at 8pm. Then I was stuck with the question: Should I keep coding or should I post my musings?

I haven’t covered all the other ways in which I distracted myself today: listening to shows on StageIt, making food for my family, looking at stuff on ebay [5], booking hotel rooms, answering SIGCSE TS [6] emails, and such. It also doesn’t help that I slept until 10:00 a.m. [7].

Fingers crossed that I’ll get the library done tomorrow. Unfortunately, there’s some shoveling to do and I have at least three meetings scheduled [8].

[1] I’m not sure why I don’t find any/c and or/c satisfactory. I want any? and any-of.

[2] I’m including tests in the reference materials. Well, I’m trying to include tests in the reference materials. Perhaps I’m trying to include tests in selected reference materials.

[3] A deprecated procedure is one that still exists but that programmers are discouraged from using. I’m still deciding wht to do about documentation for such procedures.

[4] What procedures have multiple versions? One that comes to mind is random. (random) procedures a real number between 0 (inclusive) and 1 (exclusive). (random n) produces an exact integer between 0 (inclusive) and n (exclusive). (random lb ub) produces an exact integer between lb (inclusive) and ub (exclusive).

map also has multiple versions. Or at least I think of it as having multiple versions. (map fun lst) builds a new list by applying fun to each element of lst. (map fun lst1 lst2 ... lstn) builds a new list by applying fun to the corresponding elements of each of its parameter lists. While the first map is encompassed by the second one, I prefer to document them separately.

square has multiple versions, as I observed in that musing about the image library library. Of course, those multiple versions are one reason I marked square as deprecated.

[5] Sorry Micki.

[6] SIGCSE TS is the annual Technical Symposium of the Association of Computing Machinery’s Special-Interest Group on Computer Science Education. The conference used to be called SIGCSE, but we now refer to it as SIGCSE TS to distinguish it from SIGCSE the special-interest grouop

[7] At some point I’ll muse about my sleep issues.

[8] SIGCSE TS 2025 [9], my therapist, and our money manager.

[9] Yeah, I’m simultaneously working on SIGCSE TS 2024 and SIGCSE TS 2025. At some point soon, SIGCSE TS 2024 will become less time consuming [10].

[10] At least until we hit conference time.

Version 1.0 of 2024-01-09.