Computer Science is DOOMED Because Humans are Bad at Programming (Video)

Wow, this video of Dr. Matt Welsh's presentation to the CS-50 class at Harvard really put the whole AI, large language model into perspective. It came at a very good time for me, because for past week or two I have been struggling with what was, essentially, a programming problem.  (Dean KK4DAS got it fixed, but the fact that it wasn't easy made me think about the software problems discussed in the video.) 

So much great stuff in this video: 

-- Shel Silverstein!  Yes!  We have discussed him before. The guy who wrote both kids books and articles for Playboy magazine.   The author of "The Cover of the Rolling Stone" for Dr. Hook and the Medicine Show ("The thrill that'll hit ya...") 

-- Pong.  Toy Story (Mark K6HX worked on this).  War Games. 

-- The connections between EE and CS.  Slide rules. 

-- "ChatGPT is like an eBike for the mind."  (But it seems like the jump from Rust to AI is a lot more significant and involved than the jump from bike to eBike.) 

Matt's presentation made me feel a LOT better about never having learned a programming language. And it made me feel a lot better about never having forced my kids to "learn how to code."  I always thought that there would be PEOPLE who were better at this.  As it turns out, the AI is better at that.  

Matt's admission that we don't really know what is going on in the large language models was really interesting.  Especially his description of the "empirical discovery" of the need to use the phrase "let's think step-by-step." 

His description of the cost economics behind today's programming vs. the cost of future (AI-based) software development should be quite sobering for many.  

I found the video on this article: https://thenewstack.io/if-computer-science-is-doomed-what-comes-next/

Apollo Core Rope Memory -- CuriousMarc Takes it on (video)

Here is another amazing Apollo video from CuriousMarc (AJ6JV).  Thanks to Bob KD4EBM for alerting me to this. We have recently been discussing the "Apollo rope memory" as I read Sunburst and Luminary by Don Eyles (ex K4ZHF). In this video Marc and his colleague Mike get ahold of some actual Apollo memory modules, develop a device that allows them to read it,  and they discover a design error.  Wow.  

My analog HDR head hurts after watching this.  Even Marc says he was approaching his limits in explaining all this.  

I had not heard of the bug they discovered in the Apollo 11 software just a month before launch, and how they had to climb into the Saturn V to fix it.  Amazing.  

Thanks Marc, thanks Mike and thanks Bob. 

Sunburst and Luminary -- A Poem about Transistors and ICs

 

Sunburst and Luminary by Don Eyles has a lot of the kind of color that helps the reader understand what was going on technologically during the 1960s.  For example, there is this poem about integrated circuits (you don't get to use "poem" and "integrated circuits" in the same sentence very often): 

The transistor's a marvelous invention

Replaced the tube convention

        Found its niche

        To amplify or switch

Whatever the designer's intention. 

But the breakthrough was the IC

Integrated monolithically

        It became pivotal

        As computers went digital 

With increasing complexity. 

Eyles tells us that this poem was written by hardware designer Jayne Partridge, and appears in Eldon Hall's write-up of the Apollo Guidance Computer and the decision to use ICs in it: 

https://www.amazon.com/Journey-Moon-Guidance-Computer-1996-09-03/dp/B01F9FSN7C  

Sunburst and Luminary -- An Apollo Memoir by Don Eyles (video)

Buy the book from the Amazon link on the right side of the page >>>>>>>>>>>>>>>>>>>>>>>

There is so much great stuff in this 2018 video.  I am definitely going to buy the book.  This is another of those things that reminds me (a hardware guy) of the importance of software (Sunburst and Luimary were the names of two programs that Eyles wrote to enable the LEM to land on the moon).  

-- Asked about one of the biggest ancillary contributions of the Apollo program,  Eyles immediately says, "integrated circuits."   They used three terminal NOR gates.  Lots of them. 

-- They never had a hardware failure in the Apollo computers.  Demonstrating a classic troubleshooting technique, when they discovered what they thought was a hardware failure, they ran the program on another computer.  The problem was also there, so they knew there had been no failure on the first machine. 

-- The LEM simulator was very cool. 

-- Eyles' ability -- in two hours -- to write the code for the automatic landing program that Astronaut John Young was asking for, then have it flown on an Apollo mission to the moon was very impressive. 

Thanks to the MIT Museum for posting Don's talk.  And thanks to HackADay for alerting us to it.

Also, take a look at this one: https://www.youtube.com/watch?v=oi4h04ZgQsQ

And this:  https://www.youtube.com/watch?v=cmAc7t7tAlc

More Ancient Technology Keeping Space Missions Alive

 https://www.bbc.com/future/article/20230815-the-ancient-tech-keeping-space-missions-alive

Three cheers for software.  Really. 

Another Model Rocket that Lands like SpaceX -- With a Great Description of the Flight Computer, Software, and Design

Wow, I have to give the devil his due:  this is NOT something that could be accomplished with my beloved discrete (not discreet!) transistors. 

As someone who spent a lot of time as a kid shooting Estes model rockets into the sky, this project really caught my attention.   (My simple rockets used parachutes, "streamers,"  or nothing at all to land.)  This guy uses onboard computers, software, and retractable landing gear.  Very cool.  

Thanks to Jenny List and Hack-A-Day for alerting us to this.  

Cargo Culting, Appliances, and Homebrewing

Hack-A-Day had an interesting short piece on cargo culting and computer coding:

https://hackaday.com/2022/12/03/cargo-culting-and-buried-treasure/  

Obviously there is a lot here that it applicable to ham radio.  Feynman admonished us to try to deeply understand what we are doing.   Do we risk cargo culting when we make use of gear that we really don't know anything about?  Or when we use a chip that we don't really understand?  (I'm looking at you, Si5351.)

I guess we can't really understand some of this stuff as deeply as Feynman would like -- can anyone describe the signal flow in a CPU chip?  I don't think so.  And Feynman would be the first to admit that no one really understands quantum mechanics. Still, as the author notes, we should be cognizant of the gaps in our understanding.  For there dragons lie.  Or opportunities to learn. The comments on the Hack-A-Day post are mostly pretty good. 

Model Rocket Lands Like a SpaceX Falcon 9

I had a blog post about this in 2018: 

https://soldersmoke.blogspot.com/2018/06/rockets-and-microcontroller-videos.html

He has finally achieved the very difficult goal of landing his model rocket just as SpaceX does with its Falcon 9.  Check out the video above.  

It was also very cool to see him building the rocket, using a very wide range of electronic, software and metalwork skills.  

Hack-A-Day has a good post explaining how he did this: 

https://hackaday.com/2022/08/05/bps-space-succesfully-lands-a-model-rocket/

Congratulations to BPS.space!  

Software, Hardware, and Rockets -- T-Zero Systems (videos)

There is a lot of really cool stuff in these videos.  I am a hardware guy, devoted to HARDWARE Defined Radios.  But these videos are a reminder of the importance of software, of the things we could never do with our older, analog technology.  

Watching him build his rockets, I even get ideas for my comparatively low-tech workshop: that small jig-saw that he used to cut the rocket fins looks like something I really need. 

This fellow is a professional rocket scientist who likes the work enough to take it home as a hobby.  It looks like he is working in Florida.  

Watching the videos and hearing him discuss the joys and frustrations of his endeavors reminds me a lot of what goes on in ham radio homebrewing. He often seems to have the same kind of haunted, obsessed look in his eyes that we are so familiar with.  That is what Jean Shepherd must have looked like when he couldn't get his Heising Modulator to work properly.  Oh, the humanity!   

Here it the YouTube channel: https://www.youtube.com/c/TZeroSystems

Here are some videos and stills of our 2017 rocket launches from Virginia's Shenandoah valley: 

https://soldersmoke.blogspot.com/2017/05/still-photos-and-slow-motion-video-from.html

Go WERRS! 

SolderSmoke Podcast #238 -- SolderSmoke Shack South, Cycle 25, Chiquita Banana Radio, RCA, HQ-100, Mate Mighty Midget, Sony SWL RX , Mailbag

SolderSmoke Podcast #238 is available:  http://soldersmoke.com/soldersmoke238.mp3

TRAVELOGUE:  

Cathartic decluttering:  Bill preparing for future winter travel to Dominican Republic.  Will build SolderSmoke Shack South.  Dividing everything up:  Rigs, parts, tools, supplies, antennas, test gear.  Everything.  

OUR SPONSOR:  Parts Candy.  

https://www.ebay.com/usr/partscandy  Premium quality test leads! Hand cut, hand crimped, hand soldered, these will become your new favorite test leads GUARANTEED!

PETE'S BENCH:

-- Cycle 25 better? – Out here on the left coast – it is not evident

-- Chiquita Banana and the US Navy in early ‘wireless” operations.  Why RCA was created by the US Navy in 1919.

-- Update on the MAX2870 –someone has written the code to make it work with the Raspberry Pi and the QUISK SDR software

-- Field Day prep

SHAMELESS COMMERCE DIVISION:

 

Bill needs your help:  

-- Please watch his YouTube videos.  The longer the better!  Success based on hours watched. Great to have on while you are working in the shack. Just go to YouTube and search for the SolderSmoke channel. Or:  SolderSmoke - YouTube

-- Please put links to the SolderSmoke blog on your websites and blogs. 

-- How to USE the SolderSmoke Blog: Propagation, shopping, other sites... 

-- Please put comments under the articles on the SolderSmoke blog.  We like comments and dialogue. 

BILL'S BENCH:

-- Repair of the Sony ICF SW1 shortwave receiver.  Bad electrolytics.  Number Station receiving device? 

-- HQ-100   Q-Multiplier. BFO Switch. AVC.  Noise Limiter limitations.  Dave K8WPE: Old Radio Lessons.

-- MMMRX: Detector circuit. Alignment. Muting. On the air (40 AM with DX-100) 

MAILBAG: 

-- Bob Crane W8SX -- Great interviews at FDIM. On the SolderSmoke Blog. Thanks Bob! 

-- Dave Bamford W2DAB -- Stickers on the Upper East Side of Manhattan.  FB Dave! 

-- Farhan VU2ESE:  LADPAC software now available through W7ZOI' site.  

-- Lex PH2LB: Stickers in a Netherlands pub.  

-- Dave K8WPE Michigan Mighty Mite links.  Old Smoke idea. On the SolderSmoke blog. 

-- Rich WB4TLM was in the electronics class of CF Rockey W9SCH. FB. 

-- Dean KK4DAS Working on  his dad's HQ-170A. VWS maker group on mixers. 

-- Grayson KJ7UM -- Mixology article in ER. 

-- Pete Eaton -- Farhan's new analog rig: Daylight again!  Standby for more info from Farhan. 

-- Will KI4POV New HB Al Fresco single conversion superhet.  FB. 

-- Alvin N5VZH. Shep's "I Libertine."  Yes.  I laughed, I cried,  It changed me. 

-- Chuck KF8TI.  Mr. Wizard!  

-- Steve N8NM on the mend after some routine maintenance. 

-- Ben AB4EN is listening and likes the podcast -- Thanks Ben. 

May 1939 QST

"Experimental Methods in RF Design" LADPAC Software Available FREE!

We've frequently said that is pays to check the W7ZOI web site.  Tony G4WIF did just that and pointed us in that direction, noting that the LADPAC "Ladder Package" software is now available for download from that site. 

Homebrewers will really want to have that package on their computers.  There are all kinds of useful programs in that package:  software for designing crystal filters and feedback amplifiers, a program that allows you to think systematically about receiver gain distribution and dynamic range, and many other useful things.  

You can get the program here: 

http://w7zoi.net/emerrata.html

Also on Wes's site is this May 22, 2022 picture of Farhan VU2ESE with EMRFD co-author Bob Larkin, W7PUA. 

Thanks Wes! 

Help! Can You Write the Software to Control the MAX2870 Board?

 

It is not every day that you get the chance to help a master homebrewer like Pete Juliano N6QW.  But today is just such a day.   Pete asks for some coding help: 

https://n6qw.blogspot.com/2022/03/march-18-2022-help-software-programmer.html

The Radio Gods would be pleased if you help.  The Mojo and Juju levels of all your projects would increase.   If you can help, please do so.  

Roy Lewallen W7EL Gives Us All EZNEC for FREE! Thanks Roy! (Video)

Free download here: https://www.eznec.com/

Thanks to Roy Lewallen W7EL for this great gift to the amateur radio community.  

And thanks to G-QRP for the excellent video (above) of Roy talking about antenna modeling.  

An Interview with Paul Lutus (Audio)

https://corecursive.com/remote-developer/

Thanks to Bob Scott KD4EBM for alerting us to this wonderful interview. 

Hack-a-Day: Has DIY become Click-and-Buy?

https://hackaday.com/2021/12/11/has-diy-become-click-and-buy/

Hack-A-Day today asks about the boundaries between DIY construction and the use of purchased, completed electronic components.   This is closely related to our long-standing discussion of what really constitutes "homebrew." 

Is it really homebrew if you buy a bunch of  already-stuffed PC boards and connect them together?  

Is it really a homebrew receiver if 90% of the components are inside one chip? 

Is it really homebrew if most of the signal processing is done in your computer (that you definitely did not build)? 

The comments below the article are interesting.  There we see some of the same arguments used by ham radio operators who are more inclined toward click-and-buy. They argue that since none of us are making our own resistors and transistors, we are ALL therefore click-and-buy people, so we should just get over it and pull out the credit cards. Some commenters carry this to extremes and ask if the real homebrewers are out there mining the copper for their wires.  

The debate seems to spill over into the software area:  One person asks if it is really DIY if you are using software libraries that contain code written by someone else.  Or to be truly DIY should you write all of your own code in assembly language?    

There is one very insightful comment about hams who are inclined to disparage the homebrewing that they did in their youth.  We often hear this:  "Oh, I used to build my own gear, but now-a-days I just buy commercial transceivers -- they are so much better."  As if homebrewing was a folly of youth, something that they grew out of (and up from) as they became able to afford the latest ham radio appliances.  As if homebreweing were a regrettable thing that was done only out of necessity.   This is, I think, sad.  

I think I'm a lot closer to the traditional concept of DIY than I am to click-and-buy.  I still prefer LC oscillators to Si5351/Arduino combos.  I prefer traditional filter rigs to SDR rigs.  And I prefer to make my own crystal filters.  I don't like to use ICs unless I really understand what is going on inside them (so I can be comfortable with an NE602 or an LM386, but I'm not comfortable with a CPU chip that may have millions of transistors in it).  But I am not homebrewing my own transistors nor am I mining copper. 

What do you folks think about this? 

https://hackaday.com/2021/12/11/has-diy-become-click-and-buy/

Mythbuster Video #12 -- Bandpass Filters

The really cool part comes at the end when I put the scope probe on the output, then on the input of the bandpass filter.  Exciting stuff my friends! 

Software for Computer Interface with NanoVNA: NanoVNASaver

In the UK, "brilliant" is high praise indeed.  So when Tony G4WIF told me that the NanoVNASaver software was brilliant, I took heed. 

Being able to use the tiny (nano!) device with your computer makes it a much more accessible and useful tool, especially for those of us with fat-finger syndrome. 

The photo above shows the NanoVNA measuring Return Loss and SWR  on my 40 meter dipole. You can see the identical Smith Chart traces on the two screens. 

Download and installation was very easy. I got my Windows PC version from here: 

https://github.com/mihtjel/nanovna-saver/releases/tag/v0.2.2

Other versions available from here: 

https://github.com/mihtjel/nanovna-saver

Coincidently, a fellow just today put out a video on how to use this software with the NanoVNA. 

https://www.youtube.com/watch?v=2S0dt0bYoLs

Thanks Tony! 

Digital Engineering vs. Analog Engineering

In  a book review Thompson makes this observation about the digital-analog divide: 

One difference might be that human beings can deal with ambiguity, and computers really can't. If you've done any Python [coding], you make the tiniest mistake, and everything stops immediately. That’s what makes it different even from other forms of engineering. When you are trying to fix a car, if you fail to tighten a bolt on one wheel as tight as it should be, the entire car doesn't stop working. But with code, an entire app, an entire website can go down from the misplacement of a single bracket. I think that's the one thing that sometimes scares writers away, because they are more accustomed to working with ambiguity.

https://arstechnica.com/gadgets/2019/04/how-the-new-art-form-of-coding-came-to-shape-our-modern-world/

I am definitely more accustomed to working with ambiguity. All of my rigs are filled with ambiguity. 

VU3XVR's Assembly Language 1K AtTiny-Si5351 VFO

Although Ram VU3XVR's project is in the digital realm, his barebones approach to the bits and bytes is, for me, very appealing. He takes a Si5351 and runs it with ATtiny13 with only 1k of space.  He makes intelligent use of every bit of that space.  He reveals his overall approach to rigs when he states in the video that his VFO will NOT have the traditional glowing numeral frequency readout because those bright lights can be so annoying and distracting.  I'm with your Ram!  Well done OM. Simplicity is a virtue. No more trouble with the Arduino and its fickle IDE.  No more agonizing visits to the Si5351 library.   

I see lots of applications for this little circuit.   Ram mentions beacon transmitters.  

He provides details here: 

https://vu3xvr.blogspot.com/2018/07/si5351-dds-clock-generator-using.html   

Understanding Antenna Directivity -- Help from Canada

I am in the process of repairing my beloved 17 meter fishing-pole Moxon.  It was taken out of service by the last Nor'easter of the winter.  This repair has caused me to review the theory behind antenna directivity.  I find there is a lot of "hand waving" in the explanations of how directivity happens:  "You put a reflector element next to the antenna.  And it REFLECTS!"   You are left wondering how that reflection happens.  

The Royal Canadian Air Force made a video that does a pretty good job of explaining how the reflector reflects.  See above.  Thanks Canada!  

While we are talking about antennas,  I wanted to alert readers to a really nice antenna modeling program that is available for free.  It is called 4nec2.  You can find it here: 
http://www.qsl.net/4nec2/   There is a bit of a learning curve, and I am still climbing it, but I can see how this software would be very useful.  It has an optimization feature that runs the antenna through many versions and tells you how to optimize for F/B,  gain,  SWR, or whatever you want to prioritize.  

I have discovered that my Moxon was resonant below the 17 meter band.  In other words, the antenna elements were too big.  About 3.6% too big according to my calculations.  This may be the result of my using insulated wire for the antenna elements.  Apparently the MOXGEN software assumes the use of uninsulated wire.  I'm thinking that an easy way to deal with this would be to use the frequency 3.6% above my target frequency and then use the dimensions given my the MOXGEN program.  Any thoughts on this plan? 

What a shame that Cebik's web sites have all disappeared.