Companion Robots and Maker Faire Season!

I’m super excited for Maker Faire Bay Area / Mare Island and Mini Maker Faire Rocklin.1  I’m not just excited to see everything, but to show all the things I’ve been working on for a while now.  It’s also time to pick up all the little dev boards I’ve somehow accumulated and see if I can make anything with them to show off.

  1. Project Boards
    1. Wemos D1 Mini.  A small insanely cheap (~$3?!) WiFi enabled dev board2 , which has 4MB onboard and can run Arduino.  I think it can also run MicroPython, but I haven’t tested this yet.
    2. Wemos 600 Pico.  An even smaller, even cheaper (~$2 when ordered from China) WiFi enabled dev board that runs… MicroPython?  I think??  I’m saying “I think” because I haven’t been able to get it to do anything yet.
      1. Since starting this blog post, I found a guide on installing MicroPython on Wemos boards that seems promising.
        1. Flashing MicroPython on an ESP8266
        2. https://github.com/espressif/esptool/tree/master
        3. Arguing with Python to let me use “esptool.py”
          1. esptool -p COM13 -c esp8266 flash_id
      2. As promising as that series of blog posts looked, I eventually scrapped the Wemos because it was just too much of a pain to get going with MicroPython.  I think I could have made it work, but for $7 I could also just use the Adafruit QtPy I already have.  The advantages of simply uploading code over a USB cable into a virtual drive just can’t be overstated.
    3. Other Boards
      1. I have a bad habit of picking up dev boards.  I’ve got several Adafruit QtPy’s, several Adafruit Trinkets, an Adafruit FX Sound Board, Raspberry Pi Pico (non-WiFi), various Digispark boards, a small handful of ATTiny85’s, and an even weirder assortment of VERY small programmable circuit boards (ISD1806B-COB) designed to go in greeting cards (just 6 seconds), etc.
  2. Companion Robot
    1. Background.
      1. I started this post at least a month ago when I only had a vague idea of what I wanted to make and even fewer skills.  After seeing my kid’s companion robot take shape, I wanted to get in on the action and make my own.  I decided to make a really small companion robot with just some LED’s, piezo, and small microcontroller unit.  I’d taken a stab at making a companion robot a few years ago, but set it aside for a variety of reasons and never went back.
      2. The idea for this new robot would be something a little less ambitious, make more use of NeoPixels than in prior projects3, with a little more interactivity, trying out some CircuitPython, and… let’s be real… more pizzazz!
    2. Idea:  Friendly Cloud/Vapor/Flame
      1. I still really like the copper-toned PLA I’ve been using since it has something of a steampunk flair to it.  I settled on repurposing a small plastic enclosure with a clear dome as the “body” for the robot.  I wanted it to look something like a small entrapped / captive / domesticated4 sentient cloud of vapor or perhaps flame held within a steampunk enclosure.
      2. As a very small, inexpensive board that could run either Arduino or CircuitPython, I decided on the Adafruit QtPy M0.  It could run NeoPixels, there were lots of cool guides on it, plenty of pinouts, and could definitely fit within the confines of my enclosure.
    3. Enclosure:
      1. I started the enclosure by trying to design and 3D print a part to mate with the clear plastic dome.  It took a few tries.

        This slideshow requires JavaScript.

      2. Once I had that, I extended the base so it could hold more electronics.  I could definitely have shoehorned everything into the dome, especially if I took up some of the space inside the dome, but even with an “elevated base” it was still plenty small and could use a battery pack rather than a rechargeable lipo.
      3. Once I had a good design for the enclosure, I tried to make it work with an existing 3xAAA battery pack.  In the process I yanked off the connector and ended up soldering the battery pack leads directly into the circuits.
    4. Internal Electronics
      1. I’m just not a great electrical engineer and am still copy/pasting from various guides, tinkering, changing bits of code, swapping out parts, and using “close enough” resistors.  Wiring up some LEDs or a piezo to a project isn’t very difficult – it’s some of the more fiddly bits.
      2. Piezo Element Speaker
        1. I wanted to use a piezo buzzer/speaker because they’re large and incredibly thin.  They’re not without their downsides.  The crystal wafer is also thin and a little fragile.  The piezo buzzer without additional electronics has the potential to act as a knock sensor and can generate a high voltage spike which can fry a board.  And, without additional electronics, the piezo just isn’t very loud.  There are some libraries for the Arduino that basically double the volume of a piezo by connecting it to two pins and then running each opposite of the other, doubling the voltage difference, but they only work for Arduino chips.5
        2. After searching for various ways to increase the sound of the piezo elements, I settled on trying to use the Adafruit piezo amp.  I bought two – and tried desoldering the terminal blocks.  This completely ruined one.  The other one worked great, but for the modest volume gain it was just too big in an already cramped enclosure.
        3. After searching around, I found some amplifier circuits using a small number of common parts.6
        4. Then I tried building an amplifier circuit using an NPN transistor.  After reviewing the datasheets for my NPN transistors (and PNP transistors), and breadboarding the circuit with resistors, I sketched it a few times, laid it down with copper tape, soldered it in place with SMD resistors, then pulled it off and placed it onto a piece of Kapton tape and put another piece on top – “laminating” it in place.
      3. Capacitive Touch
        1. Buttons are great and all, but with a capacitive touch pad, I could add metallic elements to my robot rather than a much bulkier button.  I bought a few brass upholstery tacks because they looked great – but they just would not accept molten solder.  I ended up cutting the prongs short with wire cutters, wrapping the stub with copper tape, then soldering the wires to the tape.  I’d also added a little piece of heat shrink tubing over the connection to help keep it together.  It’s been working well so far.
      4. LED Animations
        1. As we know from Phillip Burgess‘ incredible “power sipping NeoPixel” guide, we can conserve power and increase the impact of the LED’s by reducing the number of LED’s, keeping max brightness ~20% for a disproportionately large impact, running fewer LED’s at a time, and even running fewer colors at a time.  Between Phillip’s work, Todbot’s guide, and the specialized QtPy NeoPixel guide by Kattni Rembor, I was able to put together a few neat animations.
      5. Piezo Sounds
        1. I had a heck of time getting the piezo buzzer to do anything interesting.  Fortunately, with my kid helped convert the piano music for “Paint It Black” into tones for me.  I haven’t gotten all the note timings right, but I’m working on it!
  3. Future Modifications
    1. More Accessible Enclosure.  Right now the “lid” with a hole for the LED ring just sits on the enclosure with a light friction fit.  One idea is a hinged lid, either with a conventional hinge or perhaps a hidden swivel hinge.  The problem with that, of course, is it requires even more internal space.  Other ideas include a ring on top that screws down, holding the top down and in place.  I’m crap at designing screw threads, so I’ve avoided this.

      Hinged lid for enclosure

      Hinged lid for enclosure

    2. Piezo Knocks.  Perhaps the next version will include some kind of tap / double tap / knock sensors using one or more piezo elements.
    3. Knobs.  There’s not a ton of room inside the enclosure, but by including a gear within a gear, I might be able to rotate part of the case and have it manipulate a potentiometer.

      Offset gear within gear, manipulating an off-center internal potentiometer

      Offset gear within gear, manipulating an off-center internal potentiometer

    4. Motors.  A robot that just flashes lights and makes a few beeps can still be pretty interesting.  However, I have some neat potential features that could be added with just one or two motors.  There are some interesting limitations with the current incarnation of this robot and using a QtPy.  I’ve only got 10 pinouts7 , 1 for NeoPixels, 1 for the piezo, 6 in use for the capacitive touch sensors, leaving 2 for other potential tasks.8  However, space is already tight so one or two micro servos would be a big space commitment.  I’ve seen some really tiny micro servos that might work, but I have no idea where to source them.  One silly idea is a “weapons system” using a spring loaded projectile activated by a very small servo.

      A small spring loaded projectile launcher, actuated by a small servo

      A small spring loaded projectile launcher, actuated by a small servo

    5. Creating Tone Library.  The basic piezo tones are easy enough to play, but including the entire list of tones and the frequencies associated with them seems eat up the poor little QtPy’s memory.  I think compressing them into a library might be the way around this issue.
    6. Playing WAV files.  WAV files are bulky, but that’s the only sound file format a QtPy M0 can play.  However, with the extra 2MB from the SPI chip installed, this shouldn’t be a huge problem.  I used Audacity to mix the sound clip down to mono then to 22 KHz sample rate.  My preliminary tests worked – but it was incredibly quiet.  I haven’t run it through the audio amplifier yet, but I’m planning to.
    7. Sleep / Deep Sleep.  Ever since I swapped out the tiny LiPo for a 3xAAA battery pack, I’ve had a lot more battery life, so adding sleep / deep sleep functions haven’t been a priority.  However, this inclusion just couldn’t hurt.
  4. Other QtPy and CircuitPython Resources
    1. Adafruit’s QtPy CircuitPython PWM resource
    2. TodBot’s CircuitPython tricks
Companion Robots: Building Robot Friends
  1. Cephalopod Robot Friend, the story so far
  2. Cephalopod Robot Friend Progress
  3. CuttleBot Body and OpenSCAD Design Tips
  4. An Assembled CuttleBot Body
  5. Building the Monocle Top Hat Cat for #MicrobitVirtualConcert
  6. Companion Robots and Maker Faire Season!
  1. I just got a notice they’re no longer a “Mini”! []
  2. pinouts for my future reference []
  3. LED goggles and a Marvel Universe inspired set of “Infinity Knuckles” []
  4. OMG dome-sticated?! []
  5. This is just my very basic understanding of how it works.  I’m entirely positive this is far too simplified. []
  6. And one very long article about using a lot of parts []
  7. 12 if you want to count the onboard NeoPixel []
  8. Or 4… []

Beep Boop: Emotions for ED-E

So I’m trying to program emotions for ED-E.

Better ed-e dialogue at Fallout New Vegas - mods and community

In Fallout New Vegas, ED-E doesn’t use regular dialogue, but communicates through beeps. Basically, he gets <(Emotion) + beeping> as his dialogue. Obviously the only way to research his beeping for this project is to replay Fallout New Vegas and talk to everyone’s favorite eyebot. I came up with a list of some of the emotional beeps he has in the game (and some that I just want) to start programming into him. I have ideas for what I want him to sound like, but I just need to find the right pitches and durations.

By this I mean I am sitting in my room and saying “beep beep” to myself at different pitches in sync with a program I wrote.

Yesterday my dad was checking in on me and how my programming was going and asked how I was doing. I replied, approximately, “bleurrrrghhhhh.”
I was having trouble programming the emotional beeps because I have very specific ideas for what ED-E should sound like and no way to get that into a computer. I’m a huge music nerd, but I unfortunately don’t have perfect pitch or some superhuman ability like that to use to get ED-E to sound exactly like I want him to.1 Dad suggested using a loop with an array going up every time, which made everything so. Much. Easier.
This plays frequencies starting at 100hz and going up by 10hz every time. I started with 100hz but going up by 50hz every time, but I liked the specificity of 10hz. Once I get in the ballpark of where I want to be, I can then just run the program and correct the pitches if need be.
If you read my last post, specifically the spoiler section, you will know why I started with a shy beep. If you didn’t read my last post, go do that! If you didn’t read the spoiler section, that’s fine. What you really need to know is that I want ED-E to make noises.
I’ll admit, this took a while to get right. Not playing the tones or even figuring out what pitches they were or how long they should play for. No, getting the tones to play was easy. Getting them to stop… was much harder.
ED-E was shyly screaming at me on repeat for at least 10 minutes while I was slowly losing my mind and trying to make him shut up.
I turned to ChatGPT to help me fix this, but it was utterly unhelpful and I ended up fixing the problem myself by deleting a bunch of the garbage it generated. Now the problem was that I had to continually upload the code to make the ShyBeep function run again. I was happy that it wasn’t looping anymore, but I wanted to fine tune the beeps and making it upload again and again was a pain. That’s why I decided to make it run when a button is pressed.2
It took a little while to make it run when a button was pressed, but then it would only run once and never again, even if the button was pressed. I finally realized that this was happening because runOnce was set to false, and fixed that.
Now that I have this framework, it has been much easier to program more beeps. Now I have a sad beep, and I’m going to start working on a happy beep because I have had just about enough of ED-E’s negativity.3
  1. I almost decided to use the definitions of pitch from my Rickroll code so I could just ask the computer to start at middle C and go down or say I want the pitch to be a half note in 3/4 time. I then realized I was making this WAY more complicated than it needed to be and just used seconds and hz like a normal-ish person []
  2. I seriously love Circuit Playgrounds. They have everything. []
  3. Just kidding ED-E, I love you []

Coding with an LLM Sidekick

I fell down a rabbit hole recently which lead me to think about my experiences in the nascent field of “prompt engineering.”12

As a thought experiment, I was thinking about what I’ve managed to accomplish working with an LLM, the challenges along the way, and perhaps even where I can see the frayed edges of its current limitations.

After several starts and stops trying to hire someone to assist with a website I own, I turned to the idea of getting help from an LLM. 3 4  After all, some of them were touted as being able to actually draft code, right?  Besides, if the first step in even hiring a developer is just being able to describe what you need, and the first step of getting an LLM to generate some code is defining what I need, then…

There's no way this is going to work, right?

There’s no way this is going to work, right?

  1. 1. Task 1:  Pie Chart WordPress Plugin

    1. I started off with a simple and easy to define task.  My original plugin was a quick and dirty bit of code, so if ChatGPT could create a WordPress plugin, there was a chance it could do something simple like this.
    2. My first attempt was a wildly spectacular, but highly educational, failure.  A brief description of the plugin’s function was enough to get a WordPress plugin template file with very little functionality.  Then came the arduous LLM wrangling, my asking it for refinements, it losing track of the conversation, and the endless sincere heartfelt apologies from ChatGPT about forgetting really basic pieces of information along the way.  Some changes were minor, but changing the names of variables, functions, the plugin, switching API’s, forgetting requirements, etc.  It was constant whack-a-mole that spanned nearly 90 pages of text.
    3. My next attempt was more focused.  I created a framework for discussions, provided more context, goals, descriptions of workflow, and resources for examples.  The result was a lot better, with portions of largely functional code.  However, the LLM kept forgetting things, renaming variables, files, directories, etc.
    4. Next I created the directory structure and blank placeholder files, zipped these, and uploaded them as an attachment for the LLM to review – along with a description of the contents and the above additional context.  This was even better than before, but after a certain depth of conversation no amount of reminding could bring the LLM around to the core of the conversation.
    5. My thinking was that after a certain level of conversation, the LLM was not going to be able to synthesize all of nuance of our conversations plus the content of the code drafted.  To get around this I would begin a conversation, make a little progress, then ask it to summarize the project, the current status, and a plan for completion – which was fed into an entirely new conversation.  This way, Conversation N was able to provide a succinct and complete description which Conversation N+1 could use as a jumping off point.  My thinking was that the LLM would be best positioned to create a summary that would be useful to another LLM.
    6. This process of minor “restarts” in the conversation was one of the most successful and powerful techniques I’ve employed to combat LLM hallucinations and forgetfulness.
  2. 2. Task 2:  Blog Post Series Plugin

    1. After rewriting the above pie chart plugin using an LLM, I turned my attention to a slightly more complicated plugin.  The pie chart plugin is really just a single file which turns a shortcode with a little bit of data into a nice looking pie chart.  There’s no options page, no cross post interaction, database queries or anything.  It was really just a test to see if an LLM could really draft a basic piece of working code.
    2. The series plugin is still a reasonably simple piece of code, but it has several additional feature which require a settings page, saving settings, custom database queries, and organizing information across multiple pages.  It’s also one of the most used plugins on this website.
    3. I figured I would try feeding the LLM a description of my plugin, all the code in a directory structure, and then my initial “base” prompt which explains our roles, needs, resources, and scaffolding for a discussion.  I asked the LLM to summarize the function and features of the plugin, which it did quite nicely.  I added a few additional features I had previously worked on and asked it to incorporate this into the description.  Asking the LLM to simply “build this WordPress plugin” was met with a “you need to hire a developer” recommendation.  However, asking it to propose a workflow for building a plugin with these features was successful.  I was provided with a roadmap for building5 my plugin.
    4. This system worked reasonably well, allowing me to compartmentalize the steps, backtrack, retrace, revise code, working on a section, then another, sometimes going back to a prior sections at the LLM’s direction.  The LLM still tended to get lost, renamed variables/paths/directories/filenames, but it was less pronounced than before.  I did find it harder to use the “summarize and restart” strategy when dealing with a multi-step code development system.  However, it was still workable since I could upload all the code produced so far.
    5. The result was a new plugin, with better functionality than what I’d written myself 10 years before.  Here, the new strategy of having the LLM break the project into sections and providing a roadmap was particularly helpful.
  3. Strategy:  Conversational Scaffolding
    1. I mentioned “conversational scaffolding” and “frameworks” for discussing things with the LLM above.  This was an overarching and evolving strategy I use to help focus the LLM on the goals, keep it on track, and hopefully help it provide meaningful and useful replies.  The full text of my “prompt framework” file is too large to include here, but I’m happy to provide the highlights.
    2. Personas.  I assigned the LLM three distinct personas with differing backgrounds, strengths, and goals.  Their personas were defined in reference to one another, so the first would activate, the second would then review and interact with the first, after this process completed the third would be activated, perhaps interact with the first two, then it would move on.  I would say this process was rather successful.
    3. Myself.  I would describe myself, my goals, level of expertise, etc.  I found that I if I referred to myself as an expert, the LLM would not be as likely to offer me code proposals – but if I described myself as a newbie, it would recommend I hire a developer rather than tackle such a complex problem myself.
    4. Rules for Conversation.  These are a collection of 12 rules (at last count) which helped myself and the LLM interact.  The high points are:
      1. Answer Numbering, Answer Format, Eliminate Guesswork, Organize Assumptions, Conversational Review, Complex Answers, Context Refresher, Problem Solving Approach, File Structure, @Rules, and Personas.
      2. Each of these items were followed by a few sentences explaining something about how the LLM should be expecting to receive information and react.  My favorite of these was the rule “@Rules” which directed the LLM to begin it’s response by reviewing the Rules and following them.
    5. Knowledge.  There are a number of programming languages and technical topics I’m interested in and have used an LLM to address.  To this point, I solicited a list of useful resources from the LLM and started including a “Knowledge” section where I listed dozens of the most important resources for the languages and API’s I most commonly use.
    6. By beginning each prompt with the above “framework” (~10k of text) and following it up with a short description of my project or a file to consider, I found I was able to jump right into the project without having to provide additional significant background information.
  4. Task 3:  “Project Drift”
    1. This is a considerably more complicated task I will simply refer to as “Project Drift.”  This isn’t a real codename since the developer base is all of exactly one dude, but I don’t want to name the location/website for a variety of reasons.  In any case, Project Drift involves multiple user interfaces, numerous settings, database queries, data sanitation and validation procedures, administrator functions, and numerous other facets.  All of the above tasks and attempts were basically part of the run-up to this (ongoing) project.
    2. Using the LLM’s ability to open and read a ZIP file, as well as propose code, has been invaluable.  This in conjunction with my prompt framework allows me to get the LLM up to speed after a micro-restart – and it’s summarization procedures help me get back in the mindset after I’ve stepped away from the project for a few days.
    3. Since this project isn’t done yet, I can only give a progress report.  It’s going very well.  Much of the heavy lifting, scaffolding of the code, can be assembled for me, tedious database queries and chunks of code provided.  There are still large areas where the LLM is unable to be very helpful – and that relates to pinpointing a bug in the code (or between code sections).  This still requires a knowledgeable hand at the helm.
    4. As a solo-coder, having the assistance of another “persona” to keep me on track with a given section of code has been helpful.  I have only assigned three personas, but I could see adding a few more to fulfill different roles.

I would estimate Project Drift is roughly 30-50% complete, but this is still an incredible amount of progress in a very short time.  I would also estimate it has cut the amount of my development time by 90% (but on the easiest and most tedious stuff).

Software Development with LLMs

  1. Series Plugin Test for Illustrative Purposes Only
  2. ChatGPT WordPress Plugins
  3. Coding with an LLM Sidekick

  1. I know, it feels pretentious, doesn’t it? []
  2. I’ve got the same knee-jerk reaction to “visionary,” “thought leader,” “polymath,” and “futurist.” []
  3. Don’t get me wrong, some of the developers I’d hired simply disappeared while other relationships didn’t work out due to timing.  I don’t think anyone was malicious, just… busy, really. []
  4. Still, the job needs to be done. []
  5. Re-building? []

ChatGPT WordPress Plugins

This is kinda bananas.  Years ago I wrote a plugin to solve a problem I had.  I wanted a simple WordPress plugin where I could insert a shortcode into a blog post, specify a series title, and have it automatically search up all the other blog posts that used the same shortcode and series title, and then insert a nice looking list of blog posts in that series in chronological order.

It was one of my first plugins, still available on WordPress.org – just hidden since it hasn’t been updated in almost a decade.  It still works to this very day, if occasionally a little buggy.  After several WordPress versions, it no longer properly displays the series title, which is a real shame.

On a whim, I tried using ChatGPT to generate some plugins.

Here’s an example of my old plugin and the new ChatGPT written plugin (in this order):

Default Series Title

See how bad that was? It completely mangled the title.

Edit:  Since publishing this post, I realized that I would have to choose between

  1. Leaving the old defunct plugin in place just to make a point about how it didn’t stand the test of a decade’s worth of WordPress updates, but then also leaving broken series titles sprinkled through my back catalog of blog posts.
  2. Go back through nearly 10 years of blog posts12 to change them over to the new plugin shortcode.
  3. Disable the old plugin, but have the new plugin work with the old shortcode as well as it’s own new shortcode, at the cost of losing an example of how bad the old plugin performed.

I went with option 3.  Just take my word for it, it looked bad.

He makes a valid point

He makes a valid point

Now for the ChatGPT version:

Software Development with ChatGPT

  1. ChatGPT WordPress Plugins

It took me about an hour to whip up a working WordPress plugin with the same core functionality.  I would break down the time I spent as follows:

Time Spent Creating Series Plugin with ChatGPT

But, that’s not all!  You see, as I was writing this blog post, I realized it would be fun to include a pie chart to indicate the time I’d spent on this.  Unfortunately, the plugin I had written to do exactly this many years ago has apparently completely given up the ghost.  Thus, before I proceeded to this very sentence, I used ChatGPT to create a plugin for displaying custom pie charts!

Time Spent Creating Pie Chart Plugin with ChatGPT

Obviously, this plugin took a lot longer.  The first few versions were having all kinds of problems between the HTML Canvas code and trying to figure out how to make sure the javascript was not loading too early or too late.  In the end, I just asked it whether it was capable of even creating a pie chart – and it gave me a piece of workable javascript.  I told it to refactor the plugin using this same javascript, and then it was a matter of fine tuning the result.

If you don’t know anything about writing WordPress plugins, you could probably use ChatGPT to create a very simple plugin.  However, once it got slightly more complicated, it would likely require some troubleshooting to figure out what was happening.  In the series plugin it took me a while to root through the WordPress functions to figure out that apparently ChatGPT was trying to use a function in a way that simply did not work.  I explained to ChatGPT that that particular function could not operate in that way, explained how the data it was feeding into that function needed to be first modified, and then asked it to refactor the code.  From that point forward, it started to look a lot better.  There were some additional quirks – like putting more than one series title in a single post would only display one.  I suspect these problems of ChatGPT taking a shortcut to generate code, hardcoding certain variables and names, not considering that it might need to operate more than once on the page, may be difficult for it to anticipate and address.  Without some degree of WordPress development knowledge, I think a novice user armed only with ChatGPT would need to do a lot of refactoring, asking the program to generate the plugin all over from scratch many times, before arriving at a workable result.  Then again, a million monkeys at typewriters, right?

I think ChatGPT could be great for creating relatively simple plugins like a series plugin, a pie chart plugin, or even a table of contents plugin.  However, having seen how much time it cut out of the development process, I think it would be interesting to try developing an A/B testing plugin or more complicated plugin.

I think the next task to see if I can get it to generate QMK code for a keyboard, Arduino code, Raspberry Pi code, or a chrome extension.

I can already see some ways to improve both of the ChatGPT generated plugins used in this blog post.  My series plugin included two arrows at the bottom so the reader could navigate to the prior or next post in the series.  And I think it would be great if the chart plugin had a feature where I could specify the units, so the magnitude data would be included with the labels.  I may try getting it to shoehorn these updates later…

If you see these reflected in the charts above, I must have already done it.  :)

Software Development with LLMs

  1. Series Plugin Test for Illustrative Purposes Only
  2. ChatGPT WordPress Plugins
  3. Coding with an LLM Sidekick

  1. NGL, I can really be a lot some times. []
  2. Um, you’ve probably gathered that. []

Series Plugin Test for Illustrative Purposes Only

The only purpose for this post is to serve as a reference for a more interesting and useful post.

Software Development with LLMs

  1. Series Plugin Test for Illustrative Purposes Only
  2. ChatGPT WordPress Plugins
  3. Coding with an LLM Sidekick

Software Development with LLMs

  1. Series Plugin Test for Illustrative Purposes Only
  2. ChatGPT WordPress Plugins
  3. Coding with an LLM Sidekick

[custom_pdf_generator visitor_data=“John Doe”]

 

DIY Carved Eraser QR Code Stamp

After some modest success carving some neat designs into pink erasers, I tried making a QR code stamp.  It didn’t work well at all, with exactly just one impression working … sometimes.

The first attempt took a really long time and turned out terribly.  After a few days break, and some mental distance from the project, I returned with some new ideas and inspiration.

Here was my new approach and plan:

1. The Stamp

  1. Go Slow.  Proceeding slowly and methodically is always a good idea with sharp instruments.  I went fairly slowly the first time, but this time I would be even more methodical.
  2. Cutting.  Rather than using the carving blades for the QR code features, I switched to using a craft knife.  It was just too hard to cut precise lines with a V or U shaped blade, managing not just the direction and speed of the cut – but the depth as well – for both sides of the blade.  The craft blade let me focus on just one side at a time.  I used the blade to cut at about a 45 degree angle along one side, then other side.
  3. Don’t Cut Too Much.  I used calipers to measure the pixels cut into my first attempt as well as the stamped result.  I discovered \the stamp pixels were very slightly larger than their rubber counterparts.  This tells me it would be better to cut too little rubber – and cut more later if necessary.
  4. Removing Scraps.  Rather than sticking my big old fingers into the eraser or trying to pop it out with the blade, I used a pair of 3D printed tweezers to pluck them out.

2. The QR Code

  1. Optimize the QR Code.  There are several ways to optimize a QR code for eraser / stamp carving. 1.  I used as many of these methods as I could:
    1. “Pixel” Size.
      1. As you add more information into a QR code, the QR code generator will need to use more black and white units2 to encode the information.  After some tinkering it seems like the smallest QR code that can be generated is 441 total pixels, 21 wide by 21 tall.  The absolute largest QR code I could generate looks like one of those “magic eye” posters.  I didn’t even try to count how many pixels wide this thing was.  It’s 9,216 pixels, 96 wide by 96 tall.

        This slideshow requires JavaScript.

      2. I was having a hard time carving a stamp 21 pixels wide into a 24.5 mm3 wide eraser, so the idea of carving more than 21 lines into an eraser by hand seemed not feasible.  The very next step up from the 21×21 grid would be a 25×25 grid, so I knew I had to find a way to limit the data, find the best error correction, and find a way to cut these small pixels and thin features.
    2. Proper Error Correction.
      1. QR Codes have built in “error correction” that allows the user’s scanning device to scan information from a partially formed, damaged, or obscured code.  These settings range from L (low), M (medium), Q (quality), and H (high quality) able to error correct from up to 7%, 15%, 25%, and 30% damage respectively.  Lowering the error correction allows you to create a smaller QR Code, but it will also be less robust.
      2. I fiddled with these settings a lot to find the maximum amount of data I could put into a QR code while still retaining a maximum size of 21×21 pixels.  I was able to encode about 16 characters in a L, 13 characters in a M, 10 characters in a Q, 6 characters in a H.  The code stores numeral easier and requires more pixels to store letters and special characters.
      3. My first attempt used an error correction level of L, but was basically unusable as there must have been more than 7% distortion.  This time, I decided to try for a very high level of error correction with the Q setting for 25%.
    3. Reducing Data.  This is where I used some tricks you may, or may not, be able to replicate.
      1. URL Shortener.  A TinyURL link to my Instagram page requires 29 characters.  Looking above, this would immediately suggest a 21×21 pixel QR code would not be possible.
      2. Trimming a Link.  After some fiddling, I realized that as long as the data encoded looked like a URL (as in some characters separated by a “.”), the QR code scanner would interpret it as a link.  This means we can skip the “http://” and “https://”, saving 7-8 characters!  Unfortunately, this still doesn’t let me encode the shortest URL that TinyURL could give me which requires 20 characters after discarding the “http” stuff.
      3. Maybe Just a Domain?  Maybe you just wanted to point someone to your website and not a big long link, shortened with a URL shortener.  Let’s work the numbers backwards.  Most commonly used domains end with “.com”, “.org”, “.biz” – with 4 characters each.  Using the information above, this means we could use a domain name with up to 12 characters for an L encoded QR code, 9 for an M, 6 for a Q, and just 2 for an H.  While it would be easy to find a 12 character domain, you’re stuck with only a 7% margin for your error correction.  A domain with 6 to 9 characters for Q and M would allow for 25% and 15% error correction.  You can still find 6 character “.com” domain, but… they’re unlikely to be very memorable.  This isn’t necessarily a problem.  You might be able to find a good short domain with an unmemorable name, but forwards the user to your real website.  The problem, of course, is that no one is going to want to click on that link.
      4. How About a custom URL Shortener?  It’s still possible to purchase a short URL, but they’re pricey.  I happened to buy a good one several years ago and have hung on tightly to it.  I slapped a YOURLS install on it, and have been using it ever since.  Using my own URL shortener means I can keep the URL down to just 9 characters – including the TLD!

Okay, back to carving.  I grabbed my headphones, put on some music, and took it very slowly – a little under two hours.  Here’s some progress photos:

This slideshow requires JavaScript.

Here’s how it looked (with some additional shots to show the original design overlaid):

This slideshow requires JavaScript.

I stamped this design 9 times – and all 9 were more or less easily scannable.  The neat thing about this design is that it points to a URL shortener I own, so not only is it about as tiny as possible, but I can change the destination if I ever needed – without having to spend two hours recarving an eraser stamp!

Eraser Stamp Carving

  1. Carved Eraser Stamping
  2. Further Adventures in Eraser Carving
  3. DIY Carved Eraser QR Code Stamp

  1. I won’t get too much into the weeds on the actual method of generating QR codes, mostly because I haven’t studied the math in it, but I did find a great article which has a lot of good background info and explanations []
  2. I’ll call them “pixels” from this point forward []
  3. Just barely under an inch []

FIXED: Minecraft Launcher Stuck / Won’t Download / Slow Download on Windows 10

Sometimes you just gotta roll up your sleeves and craft your own fix

Sometimes you just gotta roll up your sleeves and craft your own fix

I recently purchased Minecraft for the PC (Windows 10) and tried to install it.  It was brutal, but I got through it.  I’m documenting my experiences here in the hopes it helps someone else.

The download was incredibly slow, would time out, get stuck, would give me the message “queued in position X,” or would simply not work at all.  Sometimes it would download 10 or 50 MB and then stop.  It was pretty frustrating.  I tried a number of things and eventually found a constellation of things that ended up fixing the problem.

  1. Things That Helped / Worked:  From Simplest to Most Complicated
    1. Making sure I was on a speedy wifi network.
      1. The easiest first step was to get off the wifi extended and onto the main wifi network.
    2. Ended all Minecraft processes / programs that were running.
      1. Press Ctrl + Shift + Esc to open the Task Manager.  (Right clicking on an empty space on the task bar can also bring this up as an option in the context menu.
      2. Right click on any “Minecraft” processes as select “End task.”
    3. Sign out of the Microsoft Store app.
      1. You should see a two letter (probably your first and last initials) circular icon in the top menu bar for this app.  Click “Sign out”
    4. Reset the Microsoft Store app.
      1. Start / Windows Button -> ⚙️Settings -> Apps
      2. Scroll down until you see “Microsoft Store,” select it and click “Advanced options”
      3. Click “Repair,” let the computer do it’s thing, then click “Reset.”
    5. Make sure the time zone and clock are up to date.
      1. This sounds kinda crazy, but it’s a legit reason why Minecraft might not be installing.  Many programs try to synchronize and authenticate each other across the internet – using agreed upon times as a basis.  An incorrect time might be used by a malicious person or program to breach a system.
      2. Anyhow, here’s how you apply this fix:
        1. Start / Windows Button -> ⚙️Settings -> Time & Language
          1. Turn on “Set time automatically” and “Set time zone automatically” and click “Sync now”
    6. Check to make sure you didn’t cripple the download speeds yourself.
      1. Start / Windows Button -> ⚙️Settings -> Update & security -> Delivery Optimization -> Advanced options
      2. When setting up my PC I had crippled Windows download speeds to 0.1 Mbps.  While this helps my day-to-day computer usage and prevents Windows from chewing up all my bandwidth to update itself, it also crippled anything I wanted to download through the Windows Store app.  I’d recommend removing all bandwidth caps while trying to download Minecraft.
    7. Notes:
      1. I can’t be sure, but Windows does a lot of stuff behind the scenes, but having checked the installed apps after Minecraft was fully installed, I noticed several other apps were also installed.  I believe the Windows Store needed to install a bunch of other programs and app / dependencies before it would actually allow Minecraft to download.  You may or may not see these pop up.
  2. Things That Did Not (Seem To) Help
    1. Restarting the computer didn’t seem to help.  I suppose it’s worth trying.
    2. Trying to install a prior version of Minecraft for Windows 7/8 didn’t work either.  They seemed download fine – but wouldn’t run at all.
    3. Trying to install the Minecraft Java version first.  I ended up uninstalling all versions and all launchers and starting from scratch.

I hope this helped.  If you’ve got some other fix that worked or idea how to help, I’m sure plenty of others would want to know.  Feel free to leave a comment.

PCB Design with KiCAD

It is pretty incredible that you can find a written or1 video tutorial on virtually any topic to learn anything.  Today, I’m particularly thankful to Shawn Hymel, Sparkfun, and Digi-Key for putting together their Intro to KiCAD video series on printed circuit board design.

This series took me from knowing nothing at all about PCB layout and design to ordering my very first board through OSHPark.  My first design isn’t anything amazing – it was basically a breakout board for an ATTiny85 to make it easier to build small projects. 

My first ATTiny hacked tap light was a mess.  I soldered wires directly to the microcontroller making it a real pain to update. ((I ask you – is this the work of a sane man?)) I soon realized my mistake and soldered an 8-pin socket in its place so I could reprogram the chip easily.

This is the alternative to a custom PCB – a rat’s nest of wires soldered to a chip

Mercifully, Shawn’s tutorial series got me up and running very quickly.  This post is not meant to be a tutorial for KiCAD, but more like a “lab notebook” for the workflow to create a board.  If you haven’t built a board yet, go check out Shawn’s series and follow along in KiCAD.  If you are a novice like me, you might find these notes helpful:

Eeschema

  • If you launch Eeschema separately from KiCAD, you can save different versions of a schematic.  Keeping old versions of design files is hugely helpful to me and if you launch KiCAD directly, the option to save different file names and versions is not available!
  • The keyboard shortcuts in Eeschema are great.  With just a few, it’s possible to really get around quickly.
    • “Shift-A” and left click to place parts 
    • “M” to move parts
    • “R” to rotate parts
  • It is necessary to add “PWR_FLAG” to both the power and ground lines.
  • Double check your connections work by clicking on the bug icon. 
  • Assign the parts you intend to use to match up with the symbols using the “Assign PCB footprints” icon.
  • Save your work and “Generate netlist” to have something the Pcbnew will be able to work with.

Pcbnew

  • First configure the Design Rules by going to Setup -> Design Rules.  Shawn pulled these KiCAD Design Rules from the OSHPark.com website.  KiCAD has apparently changed a little since the version used on the OSHPark website, but the settings are easy enough to identify and change.
    • Net Classes Editor
      • Clearance: 0.01.  Track Width: 0.01.  Via Dia: 0.03.  Via Drill: 0.015.  uVia Dia: 0.03.  uVia Drill: 0.015.  Diff Pair Width: default.  Diff Pair Gap: default.
      Global Design Rules
      • Minimum track width: 0.006.  Minimum via diameter: 0.027.  Minimum via drill: 0.013Custom Track Widths: Track 1: 0.03
  • Read netlist” to bring your design over from Eeschema.
  • Placing parts and drawing lines gets a lot easier when you fine tune the Grid.  I started with 5.00 mils at first, then smaller figures to place smaller parts and features.
  • Once the parts are arranged in Pcbnew, connect the ground and power lines using 30 mil traces and everything else using 10 mil traces.
  • Create the outline for the board cutout by clicking on “Edge.Cuts” and drawing with the “Add graphic lines” tool.  Starting with my second board, I began cutting the corners off, so that they were a little nicer to hold and 
  • Label things on the “F.SilkS” and “B.SilkS” layers using the “Add text”‘ button.  Since my boards are so small, I wanted the text to be a fair bit smaller than the default settings.  I edited the text settings by going to Setup -> Text and Drawings.  
    • Copper text thickness:  0.007.  Text height:  0.035.  Text width:  0.035.
  • Create a copper pour with Place -> Zone, then choose “F.Cu”2 and “GND”.3 and draw a box around your board.  Then repeat for the “B.Cu” and “GND.”

Again, I’m a total newbie at circuit design.  If I got something wildly wrong, please let me know.  :)

  1. More frequently these days []
  2. Front copper []
  3. Ground, natch []

Regex: Back reference and Self reference in Regular Expressions

This is going to be a very short and niche post.

Sometimes I have to use regular expression or “regex” searches to parse a bunch of text, but I can’t remember how to use the search function to find a particular sequence of character and then reuse those found characters in the text I’m trying to replace it with.  In Notepad++, this would be done as follows:

  • Search string:   (7//*[0-9]*[0-9])\r\n
    • This will find all entries with “7/3” or “7/14” or similar digits with a line return afterwards
  • Replace string:  \1/2018;
    • This will replace those entries with “7/3/2018;” and “7/14/2018;”, respectively

The trick here is that the first set of search information is collected together within a set of parenthesis, which are then referenced back by the “\1”.  If you forget the parenthesis, the “\1” term won’t “know” what it’s supposed to be repeating.

Like I said, very niche.

[SOLVED] Google Pixel stuck in a boot loop after update and full storage

If your Google Pixel is stuck displaying the “G” and forever looping through a boot screen, Google’s support will tell you there is no hope and you’ll either have to do a complete factory reset and data wipe of your phone or send it in so they can do it for you.  However, if you are adventurous or have nothing to lose, there may still be hope.

I’m not saying this will work for you, or that you should even try what worked for me.  But, if you’re as desperate as I was, you would try absolutely anything to get your photos back.1

While out of town on vacation, away from WiFi and cell carrier internet access, I took a lot of photos and a few short videos.  My phone, a lowly Google Pixel with just 32GB2 of storage ran out of space.  It claimed it was at “100%” usage of storage with only a few dozen megabytes of space left.  I leaped into action, right there on the beach, deleting some large documents and ZIP files I had downloaded and then asked the Google Photos app to “Free Up Space.”  Pats on the back!  Job well done!

Except the Pixel had other ideas.  I reached into my pocket later to pull out the phone and take a picture to discover that it was on the white boot screen with the colorful Google “G” logo and a light gray “progress” bar undulating underneath it.  I thought this might be due to the phone overheat and put the phone away.  When I pulled it out later, I found the phone was still showing the same white boot screen.  I don’t know what caused this problem.  I don’t know if it was the Android system update from a few days ago, the storage being mostly full, possibly being too warm from being in my pocket, or what.

I tried everything I could think of and everything I could find to no avail:

  • I was unable to boot the phone into anything resembling active status, not even the last measure of “Safe Mode.”
  • I could boot the phone into “Recovery Mode” which looks like the Google Android robot lying on its back with a front panel open and some text underneath.  However, attempts to “Start”, “Reboot”, and use “Recovery Mode” were all unsuccessful.
  • I tried following the recommendations of the mysterious “Techno Bill” on the Pixel Google Groups to install ADB and Fastboot and then apply a Rescue OTA to the Pixel.
    • If nothing else, this one possibility gave me hope that I could still interact with my phone, somewhat.  I was careful to try out the specific OTA3 for my phone.  As you navigate the various menus in the Recovery screens for Android, you’ll need to take note of the version of Android you’ve got – so you can place the same version over it.
    • As best as I understand this fix, you’re basically overwriting a possibly damaged or corrupted operating system with a shiny new factory fresh version – and that doing so leaves your data intact.
  • The thing that gave me the most hope was this post by Mac_DG on Reddit.  Basically, they said, “If you can get there [the Google recovery screen], and it’s a software problem everything is fixable.”
  • I even spent 43 minutes4 chatting5 with Google’s Pixel support team member, Marlene.  She was sympathetic, but assured me in no uncertain terms that my data was lost forever.  My options were to send my phone to Google or take it to a repair facility for a diagnostic.

Finally, after much searching, I found this thread on Google’s product forum where someone named Jintoku described an identical problem to mine – and someone named Kara Alexander described an identical problem … and a solution suggested by Samuel Cooksey!  If you think you’ve got the same problem – nearly full storage, frozen or looping “G” screen, and unable to enter Safe Mode, it is absolutely worth your time scrolling through this thread to see if this solution may work for you.  It is also worthwhile reading through Jintoku’s conversation with Google’s representatives to see how hard they tried not to do a factory reset on the phone – and just how certain the Google representative was that this was the only possible solution short of outright replacement of the phone.

I will not duplicate the efforts of Techno Bill, Jintoku, Kara Alexander, Samuel Cooksey, and Mac_DG by repeating everything they’ve already said.  Rather, you may use any of the above links from my search to follow along and, hopefully, set you on the correct path.  If you would prefer a condensed guide, with links, I hope this will help:

  1. Follow Techno Bill’s guide on installing ADB and Fastboot.
    1. If you’re on Windows, you either need to follow Techno Bill’s guide on updating your computer’s “Path” or, if you’re a little more comfortable doing so, you can skip steps 1-8 and instead open the command prompt and type
      1. “SET PATH=%PATH%;c:\adb”
      2. The benefit of doing it this way is that you don’t have to permanently add this path to your operating system, it’s a lot faster and easier than Techno Bill’s method, and you don’t have to restart your computer
  2. Download the Developer Beta Preview OTA for your phone6 .
    1. It almost goes without saying that with words like “Developer”, “Beta”, and “Preview” in the title, this version of the Android operating system isn’t for the faint of heart.  If you don’t know what you’re doing or, like me, don’t have anything to lose, you shouldn’t be trying this.
  3. Follow Techno Bill’s guide to “sideload” the Rescue OTA to the Pixel.

If this blog post, that I write as my phone is running the Photos application right now to back up my vacation photos, saves even one person’s data, I will be an incredibly happy person.

  1. After all, what’s the worst damage I could do?  In a worst case scenario, the phone would still be unresponsive or bricked permanently []
  2. !!! []
  3. I believe this refers to an “over the air” operating system update []
  4. I was emailed the chat logs after the session ended []
  5. Since I had no phone, natch []
  6. Pixel, Pixel XL, Pixel 2, Pixel 2 XL []