Colby Russell @colby@kosmos.social

hypothesis #2: new popular stances against lurking and lurking behavior (e.g. "not approving follow requests from accounts without a bio"), is partly responsible for (or at least has a common cause with) what we perceive to be the tendency to toxicity associated with social media; people are selecting for strong ego and (over)confidence—and selecting against meekness and thoughtfulness

Liking and upvoting someone's article is cool and all, but something that *really* shows a level of appreciation is making sure the Internet Archive grabs a copy to put into the Wayback Machine.

I'm thinking, too, of how this might overlap (or obviate) literate programming. Is it possible that by giving human names (maybe even harmless personality quirks) to a set of agents that their processes could be more readily grasped by actual humans (i.e. other programmers trying to get up to speed)? In particular, compare this to the supposed benefits of a "memory palace".

More recently, I had some shower thoughts about "proper" OO (and its proper place). What if we entertained the idea of forgetting OO, FP, etc and instead model all processes as a group attacking a problem, as in an organization (and not just the lofty ones; consider a fast food joint, for example). Your role as the programmer would be to split up the workload among Tom, Dick, and Mary. There'd still be a place for objects, but as appliances for your virtual agents' use.

Dijkstra thought we should kill the athropomorphic metaphor—personification of program components.

I've thought for a long time that there's probably something to be had by doing the opposite and embracing it in some cases. HCI is the obvious place for it. In that case, could the interaction design even end up unlocking the right abstractions for the solution to a given problem? (Suppose you didn't have a virtual assistant, for example, but instead a few, each with their own strengths.)

I think the problem is that at least half the people repeating this stuff have no idea what they're talking about. They're like Jesse Eisenberg's character in the The Squid and the Whale, parroting some BS they heard someone (like Jeff Daniel's character) say once.

Expanding on this. Russ Cox has written a series of (now widely referenced) articles on Thompson NFAs in regular expressions. There are occasionally mentions of Thompson's technique that couch it terms of it being some kind of lost art. And yet, if anyone has actually read the Dragon Book or has a copy handy, they can verify that Thompson's approach is described right there, near the beginning of the book, in Chapter 3.

Two common punching bags for which I've never seen the criticism ever substantiated:

- K&R
- The Dragon Book

The best I've seen is a bait-and-switch (a motte-and-bailey), where when you push back, people change the specifics of what they're criticizing. And that's the *best*. More often, just naked assertions.

An underexplored area: recycling abandoned trademarks. (Begin by trawling the USPTO archives.)

Here's an example of what that failure can lead to:

https://corytheboyd.com/posts/2020-03-09.html

@flancian #agora

Something I've realized within the last year or so:

If at some point you're working on a project (esp. an open source one) that involves setting up, say, a static site—or any kind of site, really—then that site should be its own knowledge repository[1] containing info about how the site is put together.

Relying on out-of-band documentation like a README that lives in a source repo somewhere is a failure.

1. To borrow a phrase from Doug Engelbart

Having spent a lot of time and energy trying to synthesize a useful description and associated shorthand for a common pattern that I see a lot—roughly, the inverse of Chesterton's fence—I recently realized the similarity of this pattern to something we already have a term for: the Just-World Fallacy.

@dredmorbius

As a practical matter, The Facility provides a pod where you are housed and fed in situ, Matrix-style.

A setting for a sci-fi plot where people derive their income from leasing out their brain cycles (and limited motor resources?) for a 4 month period of stasis. When it's time to "work", you "plug out" and your consciousness basically goes to sleep.

Like bunching up all 8-hour workdays into one block, then taking one big 8-month weekend—except from your perspective those working hours are indistinguishable from a hibernation that you don't really experience.

Maybe Finder et al should guide users towards delightful discovery of this feature when a user copies a file. Right now, they do the "Copy of File X" or "File X (copy)" or something, which tends to be less than welcome, generally. The mark of it being a quick hack (to get around the fact that a directory can't contain two files of the same name) is evident.

Instead, activate the lightweight snapshots feature in that directory and check in the original as a commit.

It's odd that this isn't already baked in, considering the official Apple support video for Time Machine <https://www.youtube.com/watch?v=jZmJFbjvKUw> uses exactly this "Rough"/"Draft"/"Final" anti-pattern.

Drastically simplified revision control baked into a file explorer in a way that doesn't suck (i.e. mostly invisible), so this can stop:

http://web.archive.org/web/20190703150800im_/http://www.pgbovine.net/two-cultures/file-copying.png

... among people who are never going to learn Git. (Although could still use Git's data model.) Like a lightweight (intentional and scoped to a directory, just like Git), less opaque form of Time Machine.

A date input in the form of a skeuomorphic dial chronometer that shows the position of the earth in a path mostly concentric with the edge. Possibly with that path segmented into the 12 regions—cells sliced out of a curvilinear grid—each representing a month.

I think this would be a really great supplement to the Computer History Museum's Oral History collections.

An idea: #PatentsExplained.

There are many expired software patents, and their inventors (who are presumably proud of their work) are likely getting to an age where, having observed the ebb and flow of the industry, would like to capture a simplified technical explanation of their work, the background that led to each patent, its place in time and what it meant, and an optional retrospective and whatever else they might have to say. I'm thinking 3–10 minutes per patent would be ideal.

Don't fool yourself; you have to create a compelling product before you can ever realistically even start thinking about selling people on a general platform—especially when the context is trying to sell that platform as _the_ way to start approaching development.