Well, lots of us do, what with the ongoing Panasonic. But also, I recently encountered a very clever website called I Miss My Bar, which is a very simple idea, cleanly executed: several toggleable and volume-adjustable channels of distinct sorts of bar-sounds ambience (conversations, the clink of glasses, street noises outside, rain on the windows) plus an embedded Spotify playlist of the sort of music a bar might be playing. You toggle the different sounds on or off, and adjust their relative volumes, to get a mix that approximates the background noise of your own favorite local spot, and it’s really surprisingly soothing.
It got me thinking, somehow, about Bernband and about Shamus Young’s Pixel City, and about the good old days of fanciful WinAmp audio visualizers like MilkDrop. I have an HDMI cable running from my desktop to the office TV, which I usually use to watch games on WNBA League Pass, but I can also put a browser window pointed to I Miss My Bar over on the TV and have its ambient audio piped through the attached Sonos; but then the screen itself is just showing the static web page.
It would be nice to have a more appropriate, and more dynamic, visual on the TV to go with those sounds.
Sometimes things take longer, but also sometimes I get distracted.
Well, I said “by the end of the month,” but technically I didn’t say which month.
Anyway I’m continuing to work on Shtory. I’ve had my usual struggle to balance “drilling down into tiny details instead of looking at the whole picture” with “drawing high-level block diagrams and blithely assuming the implementation will be trivial” and also with “just plunging ahead writing code without much plan, in the hopes that it will somehow come together.”
I’ve realized I need to add join and quit commands, and that I need to try to get all this PlantUML/Graphviz business working, and sort out the VSCode plugin situation, so that I can write myself out some sequence diagrams for how things are supposed to communicate over the sockets.
Also, because who can just keep their attention on one project at a time, I’ve gotten interested in DIY audio electronics — specifically guitar amps, which I dabbled in about a million years ago, and associated stuff. I used to hang out on the Solid State Guitar hobbyist forums, and I got an email recently that the guy who runs them had launched a Kickstarter for a 9V-battery amp kit of his own design, based, like my previous efforts, on the venerable and ubiquitous LM386 amplifier IC. The “Honey” amp kit has beaten its funding goal, so I’m looking forward to receiving my kit once those get shipped out; and in the meantime it’s stirred up a bunch of ideas from the dusty corners of my brain.
The first one is that I should repair the old Ruby circuit I housed in that Balvenie packing tube, and the second one is that I should actually do something with the components I still have lying around — including a pretty nice 10″ Jensen speaker, a big transformer, and an LM3886 chip —that I once meant to build a ~40-watt amp from.
The third one is that I could build a cabinet to house that speaker — quarter sheet of 1/2″ plywood with a nice hardwood veneer, splined miter corner joints with internal corner braces, 14″x14″ face plus 2″ high instrument panel, 8″ deep, yes I’ve already sketched out the cut list — but wire the speaker to a jack, and use the combo cab as a modular platform for trying out different amplifier circuits with different power supplies, features, etc. Even the little ~1/2-watt Ruby can drive a proper full-size speaker, so as long as all the amp circuits are designed for an 8Ω speaker (and don’t put out more than 50W) that should pose no problem. I’d also like to try building another Ruby or similar design, but add in some extras like an effects loop or two, the recommended bass-boost circuitry (perhaps with a bypass switch), a headphone jack, etc. Ideally I could design the cabinet so that swapping in my original Ruby board, the Honey, such a modified Ruby, or even the big LM3886-based design if I ever get around to making that work, would be quick and easy.
“But Scott,” you might ask, “are you actually any good at playing guitar?”
Work on Shtory continues, but has been a bit slow. The first three weeks of January were a pretty wild year, what with the fascist coup attempt which nearly resulted in members of Congress being lynched, and in the end just barely managing to keep some semblance of a representative democracy intact long enough that now we have a chance to actually improve things. So it was a little hard to concentrate for a while there.
That said, I still expect to have Shtory up to nearly-MVP-level functionality this week, and “ready enough” to put up on github by the end of the month.
The tentative feature list for the V1 milestone is:
local operation only — no following remote users
shtory list command and lisht alias (also the default behavior of shtory with no arguments): list users with current stories, marking users with unread stories with a *
shtory post command and posht alias: read stdin until EOF, then post to current user’s story
shtory read command: read all unread posts from followed users
shtory read <user> command variant: read all current (read and unread) posts from specified user, whether or not current user follows them, if they have not blocked the current user
shtory follow <user> command: follow specified user, if user exists and has not blocked current user
shtory unfollow <user> command: unfollow specified user, if current user follows them
shtory block <user> command: block specified user from following current user or seeing their stories
shtory unblock <user>: remove specified user from block list, allowing them to see stories from and follow current user if they choose to
This doesn’t fix the underlying conditions, but it does shift the possibility space.
And not a minute too soon. In many ways — often aesthetic, but also his unabashed corruption and criminality, willingness to openly embrace the extreme right in ways even Reagan wouldn’t — Trump was anomalous, but in just as many he was an inevitable outcome of what we are accustomed in the United States to calling “conservatism”. (In the GW Bush era, we used to call it “movement conservatism”, trying to draw a distinction that might always have been more spurious than we wanted it to be, between it and a more “traditional” notion of “conservatism”; but “movement conservatism” won decisively, and there is no other kind of “conservatism” anymore.)
The transfer of power doesn’t instantly fix any of the horrors Trump inflicted, but it does prevent him and his array of accomplices and sycophants — the Millers, Bannons, Barrs, Pompeos, Kushners, et al. — from continuing to worsen them, and it gets a lot of open white nationalists and would-be genocidaires out of positions of power. It doesn’t fix the underlying conditions, too many and complicatedly interrelated to get into here, but it does shift the possibility space: it’s not a given that Biden will do everything right (indeed, it’s a given that he won’t do everything right) but it’s a more realistic possibility that he can be pushed to do most things better.
Fascism isn’t dead, and the next four years will, I fear, see a lot of homegrown right-wing terrorism; there’s going to be a lot of work to do in and out of politics to try to make real progress on repairing the harms of the past and building a better future. A lot of that work will be done pushing against the Biden administration, not necessarily in cooperation with it, but it’s still a plain fact that this government will not be so resolutely opposed to progress as the previous one was, and will be more responsive to pressure toward doing the right things.
We can’t be sure of a better future, but there’s a little more reason to hope — if enough people work hard enough at it — for one than there has been for the past several years. That’s not nothing.
Democrats in Congress are looking to pass a $15/hr federal minimum wage as part of their first big bill when the Biden administration takes office, so naturally Online is full of Discourse about the minimum wage.
First of all, $15/hr is more than twice the current federal minimum wage, which has remained unchanged at $7.25/hr since 2009 — the longest stretch without an increase since the federal minimum hourly wage was first passed in 1938.
The previous record-holder was 9 years, from the 1981-01-01 increase to $3.35, to the 1990-04-01 increase to $3.80. In 2020 dollars, those are the equivalent of $10.03 and $7.68 respectively. (I’m using the Bureau of Labor Statistics’ CPI Inflation Calculator for all relative value figures in this post.) $3.35 in 1990 was worth the equivalent of only $2.26 back in 1981, or $6.77 in 2020; in other words, over that nine year span, someone working full-time — that is, 40 hours a week, with two weeks off a year (that should also be reconsidered, we should consider a 20- or 30-hour week the “full time” standard and organize our society accordingly, but that’s a whole other argument) — at minimum wage went from making the 2020 equivalent of about $20,000 a year to about $13,500 on 1990-03-31. Then the next day, our hypothetical worker got an oh-so-generous raise, of about $1800/yr in 2020 dollars.
The careful reader will have noticed that one thing that happened in between those two minimum wage increases was the entirety of the 1980s, and all of the economic, societal, and moral damage that the Reagan era did to the United States. Indeed, if you look at the 2020 equivalent minimum wage over time, there are two very distinct eras: from $1.00 ($9.72) in 1956 — I’ll use the “nominal (2020 equivalent)” convention from here on, to save typing — to breaking $10 at $2.00 ($10.61) in 1963, a high of $1.60 ($12.19) in 1968, through to the 1981 increase to $3.35 ($10.03); then a sharp drop in the neoliberal ’90s, and hovering in about the $7.50–$8.50 (in 2020 dollars) range ever since.
If our worker starts a full-time minimum wage job on January 1st, 1979, and works 2,000 hours a year without fail for 42 years, bringing them up to today, their pay has declined from $5,800 ($22,120) a year to $14,500 in current dollars. Over approximately the same period, median rent in the US has gone from $308 ($1031) in 1980, to $600 ($1226) in 1990, to $1064 ($1287) in 2009 when the minimum wage was last increased, to $1588 ($1669) in 2018.
Year
Annual Pay
Annual Rent
Rent as % of Pay
1980
$20,760
$12,372
59.6%
1990
$15,360
$14,712
95.8%
2000
$15,900
$15,576
97.9%
2010
$17,440
$15,540
89.1%
2018
$15,240
$20,028
131.4%
Annual pay at minimum wage vs. median annual rent, per the sources linked in this post
Obviously it’s been some four decades since it was even remotely possible to live on a single full-time minimum wage job without other ways of generating income or reducing costs. Indeed, even an increase to $15/hr — $30,000 at full time — would only return minimum-wage workers to approximately the 1980 norm of spending “just” two-thirds of their pay on rent.
A quick note here that opponents of raising the minimum wage like to insist that it’s “not meant to be a living wage,” that it’s supposed to be for “starter jobs” for teens who still live with their parents, etc. But this is an absolutely ahistorical claim in the first place — at the time the original Fair Labor Standards Act was passed in 1938, it was understood to establish a wage standard that would support a “minimum standard of living necessary for health, efficiency and general well-being, without substantially curtailing employment” — and wouldn’t be morally acceptable even if it were true, because the plan fact is that millions of people who are not dependent teens supported by their parents do, in fact, work minimum-wage jobs. This is all the time I’ll spend on that garbage argument.
So ultimately my position is that passing the biggest minimum wage increase we can as soon as we can is great, and if that means the increase is only to $15/hr in order to get a bill passed in Biden’s first hundred days, well, that’s fine. It’s a lot better than not passing an increase or taking longer to pass one, and it’ll help a lot of people, but it’s also not enough. (I believe I read, though I don’t currently have a source to link, that the bill would also close the loophole that allows employers to pay disabled people less than minimum wage, which is very good if true; I don’t know whether it would eliminate or even change the tipped minimum wage, which is also something that really needs to be done, because that’s a major vehicle for wage theft.)
The thing is, the government has data on various costs of living around the country. If someone asked me what the minimum wage should be, I’d say: it should be automatically set every January 1st to the higher of 1) whatever it currently is, or 2)
let r1 … r10 be the median rent in the most recent year for which data is available, in the 10 largest cities in the country
then let r be the mean of r1 … r10 (r1 + r2 + … + r10) ÷ 10
then let y be r × 12 to give the mean annual rent
then let s be y × 3 to give an annual income high enough to pay that rent and still have a decent standard of living
then let w be s ÷ (40 × 50) for full-time employment, giving an hourly wage over a year of 40-hour weeks with two weeks off.
If we assume the median rent figures above are close enough to current to substitute for r, instead of hunting down the individual components of that mean and calculating it, then we have r = $1669, y = $20,028, s = $60,084, and w = $30.04/hr is our new minimum wage.
So $15/hr is a big improvement, but it’s literally only halfway there, and unless the law is changed to incorporate a formula (maybe like mine, though mine is very off-the-cuff) for automatically adjusting it on a regular basis (maybe every year, maybe every two or five years; I think more than five years without an adjustment is clearly too long), even raising it to the $30/hr it should currently be would only be good enough temporarily.
Nothing’s funnier than a joke about something people have long since stopped talking about. Introducing shtory: stories, for the Unix shell. You’re welcome, and I’m sorry.
First Snapchat introduced its “stories” feature as a broadcast alternative to its original model of sending self-destructing snaps directly to individual users; then Facebook was rebuffed in trying to buy Snapchat and Mark Zuckerberg directed Instagram to reproduce an identical “stories” feature in Instagram, prompting all kinds of jokes about what software would have “stories” next.
Excel, perhaps
Then in 2020, Twitter announced its — also identical — feature, this time called “Fleets” (see, like “tweets”, but they’re “fleeting”), and there was a whole new round of jokes.
Tweet from Jef Poskanzer (@jefposk), November 18, 2020, reading “/bin/sh has stories now too.”, with an attached image of a shell prompt under a row of ASCII-art faces
And I thought, “heh. that’s pretty funny.”
And then I thought, “you know, I bet I could actually write a program to do that.”
And then I thought, “that’s a terrible idea.”
So, obviously I’m doing it. Introducing my new project: shtory — Snapchat/Instagram/Twitter-style stories, for the Unix shell. You’re welcome, and I’m sorry.
Everyone knows nothing’s funnier than a joke about something people have long since stopped talking about, so in keeping with that principle I hope to have an alpha of shtory ready to put up on github by about the end of the month. Ultimately the concept here isn’t very dissimilar to the traditional Unix utility finger, which displays the .plan and .project files, if any, that a user has in their home directory; but individual poshts in a shtory will, like their social media inspirations, only live for 24 hours before being automatically deleted, and I plan on implementing more granular privacy controls of the type we’re used to in modern social media, like follow lists, blocking, locked accounts, mutual-only posts, etc.
Well, it’s 2021. 2020 was pretty bad! it’s gonna take a lot of work to make 2021 good, but maybe we’ll all manage it together.
One positive change I’m making is that I’ve quit Twitter. I don’t currently plan to delete my account — it’s useful to have posts here automatically linked over there, and there are a lot of people there I’d hate to lose touch with, so if the account stays accessible they can at least find out why I’m not tweeting anymore — but a couple of weeks ago Twitter notified me that it was my tenth anniversary on the site and asked if I wouldn’t like to make a commemorative tweet with a special “10” graphic they’d prepared, and I thought, well, ten years is definitely too long to be here.
I have more thoughts about the ways in which “social media” as it currently exists, and Twitter in particular (I quit Facebook about ten years ago, so I don’t have any first-hand knowledge of its current state), is bad for us as individual people and as a society, and why, and what might be better; and maybe at some point I’ll organize those into a post here. I want to work on, and write about, more software projects first, though, so look for more on that soon.
Anyway although time is largely fake, there’s something nice about choosing to mark the new year a few weeks after the solstice — it’s about when we start to actually notice that the days are getting longer. It’s been a few months of it getting darker and colder, and it will stay cold, and even get a little colder yet, for another couple, but we can see it’s starting to get a little lighter, and we know it’ll get warm again, we just have to get through the hard depths of winter.
A metaphor, if you like. Happy new year, wear a mask, don’t go to restaurants or weddings or bars or generally spend time indoors with or near people you don’t live with, get the COVID vaccine as soon as you can, don’t vote for Republicans, tip servers and delivery people extra, do what you can to help other people.
People (or more precisely, pundits, politicians, and reporters) keep talking about the CARES act and the current deliberations in Congress as being about “a stimulus” or “stimulus checks”, etc., though some in the left wing of the Democratic party (including Reps. Ayanna Pressley and Alexandria Ocasio-Cortez) have started insisting on calling them “survival checks”, but I see hardly anyone using the term “relief”, and I wonder why that is.
On the one hand, the last time the government spent a lot of money all at once (or in a relatively short time) to try to help as many people as possible, as quickly as possible, was the ARRA under Obama, in response to the Great Recession, and that was constantly discussed as an “economic stimulus” bill, so in that way it makes some sense that the word “stimulus” would be sticky. On the other hand, “stimulus” and “relief” have different objects. You “stimulate” an economy; you give “relief” to a person. And of course you help a person “survive”.
I wonder if some politicians consider “survival” a more “marketable” word, a way to sidestep debates about moral hazard and so forth, because it implies an emergent, temporary crisis. If someone is in the path of an oncoming car, or is about to drown or fall off a roof or has been stung by a bee and needs their Epi-Pen, why, of course anyone would help them survive. If the economy — and it’s an article of faith in modern American politics, an axiom not to be questioned, that The Economy is an independent and unruly force that must be appeased at all costs — needs stimulus, why, we can pick and choose which economists we listen to about what form that stimulus should take, and decide we like the plan where it’s all tax credits and business loans.
To offer relief implies caring about people, and not just whether they’re barely existing, but whether they’re doing well. One needs relief from hardship, from suffering, from deprivation, from worry and stress. That’s an ongoing, maybe indefinite problem, and a problem of human empathy, not the abstract, imaginary machine-god of The Economy simply demanding that the numbers go up.
Right now people don’t need “stimulus”. People don’t have jobs, their bills and rents and mortages are coming due or overdue, and it’s not safe to be around other people. We’ve known since March what the right thing was to do, all along, and the government (which is to say, mainly Republicans, who control most of the veto points; maybe Democrats wouldn’t have done a good job either, but we know that many of them have at least argued for doing better, and we know the Republicans actually didn’t do the right things) has refused to. We should have had ongoing, monthly, non-means-tested, no-strings-attached relief checks to every person in the country, forgiveness of all federal student loan debt, cancellation of rents and mortgages (in the latter case, if necessary, by extending the mortgage terms by the number of months cancelled) and prohibition of eviction and foreclosure, massive federal investment into developing good practices for remote learning, even more massive investment — and worldwide collaboration — into coordinated research efforts for treatments and vaccines, federalizing production of masks and PPE if necessary, not scuppering the plan the USPS already had to deliver masks to everyone, and mobilizing the National Guard to distribute supplies, food, and medicine all over the country, to make sure everyone could safely and comfortably stay home until it was safe.
We didn’t do the right things, and a crisis became an ongoing, enormous catastrophe, and hundreds of thousands of people are needlessly dead—over 320,000 officially, but that’s well understood to be an undercount. If you look at general excess-death figures, and consider how many people died of COVID without getting diagnosed, and how many died of other causes because the health care system was overloaded, or because they didn’t seek enough care soon enough because they were afraid of COVID or because they’d lost their jobs and health coverage due to COVID, it’s hard not to conclude that the true number is probably already over 400,000. People don’t need “stimulus”, and they don’t need bare “survival”, they need relief.
I’ve just pushed v1.0.0 of Cproj up to Github. Usage is very simple: source cproj.sh (I have a line in my .bashrc to do that when my shell starts up), then type cproj <projname>. Cproj creates a <projname> directory and populates it with a ready-to-build project skeleton including source and header files, Makefile, and basic test suite using Scuttle. If Scuttle isn’t installed on the system (Cproj checks the default installation location /usr/local), Cproj attempts to download it from the main Github repository.
You can immediately cd <projname>; make to build the project and unit tests and run the test harness.
I wasn’t getting through episodes fast enough to even keep up with new releases, let alone approach the present.
At some point I became one of those people who listens to podcasts on 2x-speed, because I’d subscribed to so many — and in many cases they had long backlogs I wanted to catch up on — that I wasn’t getting through episodes at normal speed fast enough to even keep up with the new releases, let alone asymptotically approach the present.
We’re in something of a Podcast Era now, and have been since the early 2010s. Long ago, before I even started my first blog (remember blogs?), there was a First Podcast Era, which began shortly before Ben Hammersley named them “podcasts” in 2004. (Before that, of course, there had been “internet radio” — mp3 streams, which we called shoutcasts after the popular server software produced by Nullsoft; listened to at 96 or even a luxurious 128kbps in Nullsoft’s ubiquitous Winamp media player; and which, as I recall, consisted mostly of European ambient and techno — but you had to be at your desktop computer for those, because in 1998 the iPod was still three years away.) Back in the First Podcast Era, of course there was no Spotify to offer far-right bigots a hundred million dollars, so you had to have a podcatcher app, or eventually iTunes, and figure out how to copy and paste an RSS URL into it. And there was no Patreon to funnel millions a year to “far-left” bigots, and as I recall not much in the way of podcast networks or available sponsorships, so most podcasts had what we would now consider fairly amateurish production, and were strictly side gigs.
Anyway, I used to listen to a lot of podcasts in 2005–2007 or thereabouts, but I sort of fell out of the habit, and by the time the Second Podcast Era got started in earnest around 2012–2013, I didn’t consider myself “a podcast guy”. But eventually I had friends who were doing podcasts, and I wanted to support my friends, and one thing led to another, so here I am with a paid Pocket Casts account and, uh…76 subscribed feeds.
Some of them have been limited runs, or have just ended, the way things sometimes do, without having had a planned ending; or work on a seasonal schedule and are between seasons currently; or are just on some kind of hiatus; so only about two-thirds of those are still releasing new episodes with any kind of frequency, but it’s still a lot to keep up with.
All this woolgathering was by way of establishing why I’m only just getting to the fourth season, “Twilight Mirage”, of Friends at the Table, an “actual play” (i.e. episodes are recordings of gameplay sessions of what we used to call a “pencil-and-paper role-playing game”) podcast that I believe is one of the best of the genre. FatT alternates (roughly) fantasy with (roughly) science fiction, so Twilight Mirage is the second sci-fi season, and the last episode I listened to was the post-mortem Q&A for “Winter in Hieron” and its prequel “Marielda,” which formed the second fantasy season.
The, I suppose, impresario (and also gamemaster) of Friends at the Table is Austin Walker, a critic and author, the former EIC of Vice‘s former Waypoint games vertical (now reduced to “Vice Gaming” because corporate decided there was too much individuality, though the podcast — remember podcasts? — Waypoint Radio lives on with Austin as host). Austin is a genuinely brilliant person, and Disney even let him write some Star Wars stories, and his talents as a GM are matched only by the FatT cast, and in particular Jack de Quidt’s stunning work composing the scores for each season. So when, around the midpoint of the Marielda/Winter postmortem, I heard Austin describe the season 2 (“COUNTER/Weight”) episode “An Animal Out of Context” as “the best thing I’ve ever made”, I decided, well, I remember that being great, but I should go listen to it again.
I don’t know if that episode, which intersperses small vignettes with the other main characters among longer stretches of Jack and Austin playing a GM-less, two-player storytelling game of their own design, recontextualized for the COUNTER/Weight setting, would have the same impact for someone who hadn’t followed the story up to that point, so I hesitate to recommend listening to just that episode alone. But there’s a moment, three-quarters or a little more of the way through, where in one of those side vignettes, Art Martinez-Tebbel mentions casually that “an animal out of context” (in a future zoo, as it happened) is a hard thing to understand — giving, presumably unknowingly in the moment, the episode its title and also perfectly summarizing the alienation Jack’s character feels. I had to pause and take a breath when I heard that again, because it crystallizes a lot of what I think is so great about FatT: that Austin and Jack were telling such a powerful story; that Art, separately, got at such a crucial idea in a different (as it were) context; that Austin and Alicia Acampora, the producer and also a cast member, caught that brief phrase and realized how well it evoked the episode’s themes.
This is a very rambly weekend post, but the short of it is, I’m glad that, in this Second Podcast Era, it can be feasible for a show like Friends at the Table to run for over six years, and if you think a collaborative longform fiction radio show “focused on critical worldbuilding, smart characterization, and fun interaction between good friends” sounds interesting, you should try giving it a listen.
Scott Madin 
You must be logged in to post a comment.