If you're an insomniac, I recommend David Brinkley's book "Defining Vision : How Broadcasters Lured the Government into Inciting a Revolution in Television"
It's an account of the incredibly messy process that got us the ATSC standard in the USA. A little dry but a little interesting.
The upside is that the group took so long to define a standard that digital compression and broadcasting technologies matured enough so that HDTV in the US could be fully digital and not analog like the Japanese had pioneered and proposed as a world standard.
Well also not really NTSC either. Did you see the part where he talks about them trying to get a recording of gameplay for the commercial?
> "We can't get it to play on the videotape recorder, why?” I say, "Well, it's not really video! There’s no interlace… Treat it as though it’s PAL, just run up through a standard converter.”
Not exactly related to the topic at hand, but that opening paragraph really caught my eye:
"I had my father sign me up for an RCA correspondence course on radio and television repair. So, by the time I got to Berkeley, I was a journeyman TV repairman and actually paid my way through college through television."
Sometimes it's had to believe just how much the cost of college has changed since the 60s/70s.
Really highlights the workings of "Baumol cost disease" in general: TVs themselves have got cheaper, so everyone can have one for a fraction of their income. The relative cost of repair labour has gone up, as well as the devices getting less repairable, so TV repairmen are no longer a major source of employment. While college - which entirely consists of paying for the skilled labour of others, as well as buildings - has gone up hugely by comparison.
Someone should calculate the historical cost of college in terms of how many televisions you could have bought instead.
It's true, and a huge shame, particularly as it's driven appliances to be made with parts not designed to last beyond a 10 year lifespan.
That said, it's been interesting to witness the recent rise in YouTube repair guides coupled with online parts stores. Obviously not for everyone, but I've had good results repairing two different washing machines, a stove, a coffee grinder, my hot tub, a laptop, a video game controller, etc. For a lot of those I would have felt completely overwhelmed and not known where to start without a video tutorial.
> The relative cost of repair labour has gone up, as well as the devices getting less repairable, so TV repairmen are no longer a major source of employment.
I think you're missing another important factor. TV repair was a major source of employment because early TVs were highly failure-prone! A TV was usually the most complicated piece of electronic equipment in the house. It would often have dozens of tubes that could burn out, and older models were assembled with point-to-point wiring (instead of circuit boards) which was prone to short circuits. It wouldn't be unusual for one to need repair or tuneup work a few times a year.
Modern TVs, by comparison, rarely require repairs. A new flatscreen TV is likely to provide continuous, problem-free service until it is obsolete. That's the real reason why TV repairmen are rare now, not because TVs are cheaper or harder to repair.
I started college in EE at UT Austin in 1978. The rent on my two-bedroom apartment was $180/month. My textbooks cost about $300/semester. My tuition was something like $10/credit hour/semester. I don't remember the exact number because even on my very limited budget, it was in the noise.
I stayed in Austin for grad school. Pre-turtleneck Steve Jobs came to campus in 1984 and gave a talk about how great his new computer was, and told us he'd let us all buy one for half price. I bought one of those "Macintosh" things, plus a Lisa to develop for it, and the rest for me was history.
Interesting tidbit: Because Macs were cheap for students in Austin, a lot of great Mac developer talent sprang up there. Jeff Kodosky prototyped a really cool visual programming language on the first Mac, then hired several UT students to productize it into a commercial engineering tool. He called it LabView.
According to https://www.usinflationcalculator.com monthly apartment rental of $180/month would be $713 today, and the textbooks of $300/semester would be $1,188 today.
At $600 per year (1978 dollars, two semesters), inflation correction puts it at about $2400 per year in 2020 dollars. That's on the high end of what domestic students pay in Europe today for a full-time bachelor or master degree program (per academic year).
You might not realize it, but some of the textbooks are less expensive in other parts of the world partly because they are subsidized by sales in the US. It’s a complicated industry.
I’d rather pay more to use better textbooks. My experience is that for any given subject, there are often only a few really good textbooks. Sometimes you run into professors who have their own materials which substitute for a textbook, but that’s rare and where I’ve seen it, the notes have been built up over many years of teaching the same courses. Sometimes those notes will get turned into a textbook, which is what happened with Kolmogorov.
I’ve known some professors who’ve written textbooks and it’s basically a second job that takes time away from both research and teaching. It’s a very labor-intensive process with low reward.
The place I’ve seen it where it really made sense was where a foreign language professor I had also ran a foreign language education department.
> My experience is that for any given subject, there are often only a few really good textbooks.
Yeah, but they don't stay really good if they have to go through a mandatory new edition every couple of years to keep students from buying used textbooks. And the writers who get sucked into that treadmill generally aren't the ones who are capable of writing the really good textbooks in the first place — writing textbooks doesn't pay enough to make it a profitable alternative to practicing in a highly skilled research field.
Really good textbooks also rarely stay really good when they're translated from English into Spanish, Chinese, Arabic, or whatever. There isn't a free market in translation — whatever dipshit landed the contract to sell a local translation of Horowitz & Hill or whatever, that's whose shitty translation you have to buy, unless you fancy struggling through electrical engineering and an overcomplicated foreign language at the same time. Copyright is a scam to keep poor countries poor.
So generally the textbooks you have to pay more for are the worse ones.
> Yeah, but they don't stay really good if they have to go through a mandatory new edition every couple of years to keep students from buying used textbooks.
Buy an older edition on eBay and get problem sets from classmates or the library. This works well unless the publisher has some skeezy online component.
> So generally the textbooks you have to pay more for are the worse ones.
Not generally true. Maybe true for some subjects. My experience is that in upper-level STEM classes, there might be only one or two good books that you’d want to use.
For example, take algorithms. There are a couple good algorithms textbooks, one of them is The Algorithm Design Manual by Skiena. I can see that it costs $33 on amazon.com for a digital copy, but if I go to amazon.in it’s available for ₹866, about $11.50 at current exchange rate.
That isn’t even a specialized subject. If you’re in an advanced subject in an engineering field or one of the sciences there might be even fewer choices available.
Skiena's book is universally acknowledged to be great, but as you point out, it costs US$33 in the US, not US$150 (though Amazon tells me it's US$52.06, so I guess the price isn't very stable), and it isn't one of the ones that comes out in a new edition every two years with the problems slightly tweaked; it's in its second edition and its 23rd year. So it's an example of the dynamic I'm talking about, not a counterexample to it.
I'm talking about things like https://www.amazon.com/Computer-Science-Overview-Brookshear-... (Brookshear & Brylow), which Amazon tells me is US$155.49 for the Swindle or US$79 in paperback (new) and is in its 13th edition after 35 years; the first edition was evidently in 1985, so they've averaged roughly three years per edition. As far as I can tell, the only novel work Brookshear has ever published (not counting textbooks and papers about teaching) was a two-page paper in 1978 on projective prime ideals; and Brylow is hardly a more accomplished researcher, having published a couple of papers about formal methods for real-time software in 2001–3 and a few less-cited papers up to 2012. However good the book was in 1985, I wouldn't consider teaching from it today.
Skiena's book is an example of a really good textbook. Brookshear's book is an example of paying more. I did try to find a book like that about algorithms and data structures, but unsuccessfully; the closest I could find was Brookshear's chapter 5, which spends 50 pages to cover Euclid's algorithm, sequential search, insertion sort, and binary search (as well as, lamentably, bubble sort), all in terrible Python and 1958-style flowcharts, while giving terrible programming advice and performance numbers so wrong they weren't even right in 1985.
It's true that this kind of thing is limited to entry-level courses and more popular majors, as far as I can tell. Upper-level STEM courses don't attract enough students for it to be worth the effort to rip them off this way, even for world-class ripoff experts like academic publishers.
India has a significant advantage over Perú, Egypt, or mainland China, in that English is a second language to most of its population but not a foreign one, so the translation obstacle isn't nearly as serious.
The US textbook lifecycle where they are updated unnecessarily each year but cost hundreds of dollars is basically just an add-on fee that circumvents the university organisation, like a budget airline charging for food and hold baggage.
There are some measures to make sure you don't buy used books either. Things like including software licenses in the purchase or having a new edition every year or so. Sometimes these are even written by the professor teaching, so it is then part of their income.
So tuition and texts could be paid with 150 hours of on-campus labor. Let's call that 15% part-time. Someone on 20 hour/week 50% part-time would therefore have wages left over for things like food, clothing, and housing, and time left over for actually being a student.
Now, tuition and text can only be paid with 100% or less of a full-time wage job if you are an in-state student at a public university, or attend a community college. And then, no time or money left over for lectures, study, or basic survival.
In terms of student labor hours per semester, college tuition and textbooks have increased an average of 4%-5% per year from JoeAltmaier's 1976 numbers to 2020. Due to advances in efficiency, infrastructure, and technology, that metric should be decreasing every year for most goods. Some see large downward jumps as machines are invented that allow one person to do jobs that formerly required more laborers.
What the hell have universities been doing? Do they not all have economics departments and business schools?
Labor costs (faculty etc) should remain the same - people are the part that should be getting more of the pie, not less. And Universities are run on people (and buildings and football teams, yes).
Back in my day, you were able to take the "First Class Radiotelephone Exam" at the local FCC office and be able to get a job at radio stations easily. I worked my way through college this way.
This license no longer exists. It was several hundred multiple choice questions about operating commercial radio stations.
I am not sure what relevance they have these days. The FAA seems to handle everything involving aviation radios. I know nothing about shipping. Everything else seems to be computer-based (so you don't operate the radio, you operate a computer that operates the radio in accordance with FCC rules).
The licence that remains is equivalent to the old "General Radio Telephone" licence. This was a much easier, shorter test and wasn't considered the equivalent of a "degree" by employers.
"When the Apple II was being done, I helped them. I mean, I actually loaned them my oscilloscope, I had a 465 Tektronix scope, which I still have, and they designed the Apple II with it. I designed Pong with it."
AA: In one repair shop, there was a real cheap, sleazy color bar generator [for testing televisions]. And instead of doing color properly by synthesizing the phases and stuff like that, it simply used a crystal that was 3.58 megahertz [the carrier frequency for the color signal] minus 15.750 kilohertz, which was the horizontal scan frequency. So it slipped one phase, 360 degrees, every scan line. You put that signal on the screen and you’ve got a color bar from left to right. It really was really the cheapest, sleaziest way of doing it!
AA: Anyway, part of the way into the design, Nolan said, “Oh, by the way, it has to be colored.” But I knew he was going to pull this stunt, so I’d already chosen the crystal [that drove the chip] to be 3.58 MHz, minus 15.750 kilohertz.
SC: So, in the home version of Pong, the graphics would simply change color from one side of the screen to the other?
AA: Right, the whole goal for doing this was just to put on the box: “Color!”
I'd never seen Pong in color, but having a simple color gradient across the screen works very well for the game. You can see the ball changing color as it crosses the screen, which is a bit weird, but nothing else looks out of place: https://www.youtube.com/watch?v=F73cgR63Jn4
If you look at the video signal on a waveform and zoom in on the front porch area of the signal, you can see the color burst. It looks like a group of 9 (i think) spikes. The first spike is supposed to go up, but it was possible to have the color burst "flipped" so that the first spike went down. (Maybe I have it backwards, been a looong time.) If that happened during a broadcast, the FCC would issue fines. There was a very expensive piece of video equipment that was meant for doing video effects during post production (3DO maybe?). If you had a flipped burst, you could pass your signal through it without using any of the effects, and it would get corrected. For a time, the post house I worked at said they made more money using it for that purpose than what it was designed to do. That was all before my time though. Supposedly, some poor engineer would have to watch all incoming video tapes with the waveform zoomed in to the color burst just to watch for these flips. If it occurred, they would reject the tape.
I remember this being used in games on the Atari computers, but only those that were ports of Apple games and Apple’s crude graphics. By skipping horizontal pixels, they’d get composite NTSC to smear the area into a color.
But the Ataris had S-video outputs (in 1978!), which of course is not composite. The color resolution is too high to produce artifacts when using separated video, so those games were essentially black & white on a good monitor like the Commodore 1702 hooked up with separated Y & C signals.
The Ataris were great, forward-thinking computers.
> In those days, in Silicon Valley, we didn’t keep secrets.
That's how Silicon Valley was created from the 1950s through the 1980s, and still to a significant extent through the 1990s: knowledge was not treated as property, but rather as a commons. This is the foundation of the professions: judges and lawyers do not keep their reasoning secret, nor do priests, nor do professors, nor do medical doctors — to the extent that the knowledge they create belongs to anyone, it belongs to the profession, not to the individual professional or their client. When Silicon Valley started abandoning that principle in the 1990s and especially after 2000, the epicenter of innovation shifted to Shenzhen, where that principle was still practiced. And today with the covid pandemic we see the fallout: the US has been left far behind technologically and is completely unable to deploy the measures needed to stop the pandemic.
That isn't the only factor, of course, but it's a big one. The profession of engineering computing machinery has been subordinated to the quarterly goals of business, which can be more effectively promoted if information flow is controlled by management and money, not by engineers and sharing. But the innovation that remains vital in Silicon Valley today is underpinned by open-source software, which is based on the same kind of promiscuous openness that got Steve Wozniak debugging Atari machines for free on the late shift. Too bad free software can't tell if you've got a fever.
"Analog NTSC televisions generate color by looking at the phase of a signal relative to a reference frequency."
[...]
"And instead of doing color properly by synthesizing the phases and stuff like that, it simply used a crystal that was 3.58 megahertz [the carrier frequency for the color signal] minus 15.750 kilohertz, which was the horizontal scan frequency. So it slipped one phase, 360 degrees, every scan line. You put that signal on the screen and you’ve got a color bar from left to right."
55 comments
[ 3.4 ms ] story [ 76.7 ms ] thread*Committee stands for National Television System Committee
Officially, this stands for "Advanced Television Systems Committee".
Unofficially, the "A" stands for "always".
It's an account of the incredibly messy process that got us the ATSC standard in the USA. A little dry but a little interesting.
https://www.amazon.com/Defining-Vision-Broadcasters-Governme...
The upside is that the group took so long to define a standard that digital compression and broadcasting technologies matured enough so that HDTV in the US could be fully digital and not analog like the Japanese had pioneered and proposed as a world standard.
> "We can't get it to play on the videotape recorder, why?” I say, "Well, it's not really video! There’s no interlace… Treat it as though it’s PAL, just run up through a standard converter.”
"I had my father sign me up for an RCA correspondence course on radio and television repair. So, by the time I got to Berkeley, I was a journeyman TV repairman and actually paid my way through college through television."
Sometimes it's had to believe just how much the cost of college has changed since the 60s/70s.
Someone should calculate the historical cost of college in terms of how many televisions you could have bought instead.
Human rights violations in countries like China is to blame for reducing the value of skilled labor.
That said, it's been interesting to witness the recent rise in YouTube repair guides coupled with online parts stores. Obviously not for everyone, but I've had good results repairing two different washing machines, a stove, a coffee grinder, my hot tub, a laptop, a video game controller, etc. For a lot of those I would have felt completely overwhelmed and not known where to start without a video tutorial.
I think you're missing another important factor. TV repair was a major source of employment because early TVs were highly failure-prone! A TV was usually the most complicated piece of electronic equipment in the house. It would often have dozens of tubes that could burn out, and older models were assembled with point-to-point wiring (instead of circuit boards) which was prone to short circuits. It wouldn't be unusual for one to need repair or tuneup work a few times a year.
Modern TVs, by comparison, rarely require repairs. A new flatscreen TV is likely to provide continuous, problem-free service until it is obsolete. That's the real reason why TV repairmen are rare now, not because TVs are cheaper or harder to repair.
I stayed in Austin for grad school. Pre-turtleneck Steve Jobs came to campus in 1984 and gave a talk about how great his new computer was, and told us he'd let us all buy one for half price. I bought one of those "Macintosh" things, plus a Lisa to develop for it, and the rest for me was history.
Interesting tidbit: Because Macs were cheap for students in Austin, a lot of great Mac developer talent sprang up there. Jeff Kodosky prototyped a really cool visual programming language on the first Mac, then hired several UT students to productize it into a commercial engineering tool. He called it LabView.
https://www.ni.com/en-us/about-ni/leadership/kodosky.html
According to https://www.rentcafe.com/average-rent-market-trends/us/tx/au... the average cost for an apartment in the neighborhood of the University of Texas at Austin is $2,581.
:-)
But I earned $2.75/hr at the college computer lab as a lab monitor. So yeah.
$100 USD in 1976 -> $462 today (per CPI)
$100 invested in S&P500 in Aug 1976 -> $8890 today.
No, that's not how inflation works.
You have to pay for college? And buy your own text books?!
Why don't they just... build campus libraries..? With all the required books there?
I’ve known some professors who’ve written textbooks and it’s basically a second job that takes time away from both research and teaching. It’s a very labor-intensive process with low reward.
The place I’ve seen it where it really made sense was where a foreign language professor I had also ran a foreign language education department.
Yeah, but they don't stay really good if they have to go through a mandatory new edition every couple of years to keep students from buying used textbooks. And the writers who get sucked into that treadmill generally aren't the ones who are capable of writing the really good textbooks in the first place — writing textbooks doesn't pay enough to make it a profitable alternative to practicing in a highly skilled research field.
Really good textbooks also rarely stay really good when they're translated from English into Spanish, Chinese, Arabic, or whatever. There isn't a free market in translation — whatever dipshit landed the contract to sell a local translation of Horowitz & Hill or whatever, that's whose shitty translation you have to buy, unless you fancy struggling through electrical engineering and an overcomplicated foreign language at the same time. Copyright is a scam to keep poor countries poor.
So generally the textbooks you have to pay more for are the worse ones.
Buy an older edition on eBay and get problem sets from classmates or the library. This works well unless the publisher has some skeezy online component.
> So generally the textbooks you have to pay more for are the worse ones.
Not generally true. Maybe true for some subjects. My experience is that in upper-level STEM classes, there might be only one or two good books that you’d want to use.
For example, take algorithms. There are a couple good algorithms textbooks, one of them is The Algorithm Design Manual by Skiena. I can see that it costs $33 on amazon.com for a digital copy, but if I go to amazon.in it’s available for ₹866, about $11.50 at current exchange rate.
That isn’t even a specialized subject. If you’re in an advanced subject in an engineering field or one of the sciences there might be even fewer choices available.
I'm talking about things like https://www.amazon.com/Computer-Science-Overview-Brookshear-... (Brookshear & Brylow), which Amazon tells me is US$155.49 for the Swindle or US$79 in paperback (new) and is in its 13th edition after 35 years; the first edition was evidently in 1985, so they've averaged roughly three years per edition. As far as I can tell, the only novel work Brookshear has ever published (not counting textbooks and papers about teaching) was a two-page paper in 1978 on projective prime ideals; and Brylow is hardly a more accomplished researcher, having published a couple of papers about formal methods for real-time software in 2001–3 and a few less-cited papers up to 2012. However good the book was in 1985, I wouldn't consider teaching from it today.
Skiena's book is an example of a really good textbook. Brookshear's book is an example of paying more. I did try to find a book like that about algorithms and data structures, but unsuccessfully; the closest I could find was Brookshear's chapter 5, which spends 50 pages to cover Euclid's algorithm, sequential search, insertion sort, and binary search (as well as, lamentably, bubble sort), all in terrible Python and 1958-style flowcharts, while giving terrible programming advice and performance numbers so wrong they weren't even right in 1985.
It's true that this kind of thing is limited to entry-level courses and more popular majors, as far as I can tell. Upper-level STEM courses don't attract enough students for it to be worth the effort to rip them off this way, even for world-class ripoff experts like academic publishers.
India has a significant advantage over Perú, Egypt, or mainland China, in that English is a second language to most of its population but not a foreign one, so the translation obstacle isn't nearly as serious.
Now, tuition and text can only be paid with 100% or less of a full-time wage job if you are an in-state student at a public university, or attend a community college. And then, no time or money left over for lectures, study, or basic survival.
In terms of student labor hours per semester, college tuition and textbooks have increased an average of 4%-5% per year from JoeAltmaier's 1976 numbers to 2020. Due to advances in efficiency, infrastructure, and technology, that metric should be decreasing every year for most goods. Some see large downward jumps as machines are invented that allow one person to do jobs that formerly required more laborers.
What the hell have universities been doing? Do they not all have economics departments and business schools?
This license no longer exists. It was several hundred multiple choice questions about operating commercial radio stations.
I am not sure what relevance they have these days. The FAA seems to handle everything involving aviation radios. I know nothing about shipping. Everything else seems to be computer-based (so you don't operate the radio, you operate a computer that operates the radio in accordance with FCC rules).
Read to the bottom of the page. It lists the "First Class Radiotelephone Licence" as "no longer operative" https://www.fcc.gov/wireless/bureau-divisions/mobility-divis...
When I renewed mine, it was "downgraded" to the General Radiotelephone license.
https://www.dailycal.org/2014/12/22/history-uc-tuition-since...
Damn, that scope is a serious piece of history.
AA: Anyway, part of the way into the design, Nolan said, “Oh, by the way, it has to be colored.” But I knew he was going to pull this stunt, so I’d already chosen the crystal [that drove the chip] to be 3.58 MHz, minus 15.750 kilohertz.
SC: So, in the home version of Pong, the graphics would simply change color from one side of the screen to the other?
AA: Right, the whole goal for doing this was just to put on the box: “Color!”
I'd never seen Pong in color, but having a simple color gradient across the screen works very well for the game. You can see the ball changing color as it crosses the screen, which is a bit weird, but nothing else looks out of place: https://www.youtube.com/watch?v=F73cgR63Jn4
https://int10h.org/blog/2015/04/cga-in-1024-colors-new-mode-...
But the Ataris had S-video outputs (in 1978!), which of course is not composite. The color resolution is too high to produce artifacts when using separated video, so those games were essentially black & white on a good monitor like the Commodore 1702 hooked up with separated Y & C signals.
The Ataris were great, forward-thinking computers.
That's how Silicon Valley was created from the 1950s through the 1980s, and still to a significant extent through the 1990s: knowledge was not treated as property, but rather as a commons. This is the foundation of the professions: judges and lawyers do not keep their reasoning secret, nor do priests, nor do professors, nor do medical doctors — to the extent that the knowledge they create belongs to anyone, it belongs to the profession, not to the individual professional or their client. When Silicon Valley started abandoning that principle in the 1990s and especially after 2000, the epicenter of innovation shifted to Shenzhen, where that principle was still practiced. And today with the covid pandemic we see the fallout: the US has been left far behind technologically and is completely unable to deploy the measures needed to stop the pandemic.
That isn't the only factor, of course, but it's a big one. The profession of engineering computing machinery has been subordinated to the quarterly goals of business, which can be more effectively promoted if information flow is controlled by management and money, not by engineers and sharing. But the innovation that remains vital in Silicon Valley today is underpinned by open-source software, which is based on the same kind of promiscuous openness that got Steve Wozniak debugging Atari machines for free on the late shift. Too bad free software can't tell if you've got a fever.
"Analog NTSC televisions generate color by looking at the phase of a signal relative to a reference frequency."
[...]
"And instead of doing color properly by synthesizing the phases and stuff like that, it simply used a crystal that was 3.58 megahertz [the carrier frequency for the color signal] minus 15.750 kilohertz, which was the horizontal scan frequency. So it slipped one phase, 360 degrees, every scan line. You put that signal on the screen and you’ve got a color bar from left to right."