Jon architects and implements modern technology for a living. Our home is a (usually) seamless integration of vintage tech and newer stuff — although at any given time, there’s probably a table somewhere covered in circuit boards waiting to be assembled or repaired.
Also read: Restoration Projects | Dev Projects
Switching my home computers to LTSC was the single best decision I’ve made for the health of my network — and for my sanity. If you don’t need all the latest bells and whistles — and more importantly, if you’re driven insane by the constant feature updates that are often more painful to install than they’re worth, give serious thought to getting a hold of LTSC. In particular, on my Mac Pro, every Windows feature update was a battle to keep it stable. The only feature I care about is that it starts up when I need it.
However, there is one (dubious) down side: LTSC does not include the Microsoft Store for getting access to apps. Lots of apps are available from alternate channels, but occasionally there’s one — like Microsoft’s To Do app — that’s only available in their app store. But don’t panic: you can still get these apps (assuming they’re free). It just takes a little more work…
Update: A reader made a comment below that provides an alternate approach — I haven’t tried it yet, but its a fantastic idea, so I’m adding it here. Thanks, Oliver!
If all goes well, you’ll get a list of one or more packages. For To Do, there were 5 packages that I needed — but the results had almost 10 times that. Don’t worry, you won’t need them all. Start with the .Appx or .Appxbundle that looks like the app you want. Note, from the possible choices, you’ll want the latest version number, and the correct processor architecture. If you’re running 64-bit Windows, you’ll want the x64 version of the app. Download it somewhere memorable.
When you run it, you’re likely to get a scary red error dump. Inside that message are helpful tips telling you that the app needs one of the other files from Step 3. Go get it — again paying attention to version number and processor architecture.
Step 5 – Repeat
Run the Add-AppxPackage command on each download, then re-try Step 4. Each time it will tell you about another file it’s missing. If you’re feeling confident, you can guess ahead — but to be on the safe side, go file-by-file, grabbing exactly the one it complains about after each attempt at Step 4.
Eventually you’ll have all the dependencies installed, and the app you wanted will actually install — assuming its compatible with your version of Windows, you can now use it like normal!
The company that started with the motto “Don’t Be Evil” has spent the last decade or so flirting with ideas that are awfully close to evil. That doesn’t mean that the organization is bad — any more than a hang nail means a human being is dying — but Google sure could use a pair of nail clippers.
When GMail first came out, I was ecstatic to get an invite. They were transparent about the trade-off at the time, and we all accepted it as reasonable: Google has automated systems that read your mail so that they can personalize advertisements to your interests. If you send an email to someone about how you burnt your toast that morning, seeing ads for toasters in the afternoon seemed fairly innocuous — even a little amusing! At the time, though, Google’s coverage of your digital life was just search results. Adding e-mail felt natural and not really that intrusive.
And let’s talk about YouTube, a virtually unavoidable Google property, full of useful content, and a site that historians might one day determine was a leading cause for the end of our democracy. YouTube is awful — and its entirely by accident. Google deflects privacy concerns by pointing out that the analysis of all this data is done by algorithms, not people. There’s probably no person at Google that actually knows how to gather all the information you’ve given them into a profile of you personally. But there doesn’t need to be: their software is sufficiently empowered to manipulate you in ways you aren’t equipped to resist.
YouTube’s recommendation algorithm has been disowned by its own creator as reckless and dangerous, and while its been tweaked since it was launched on the world like SkyNet, the evil AI from the Terminator movie franchise, and now has human over-seers to guide its machinations towards less destructive content, its still a pernicious and outsized influencer of human thought. Look no further than 2020’s rampant embrace of conspiracy theories for proof positive that recommendation engines are not our friends.
Google set out to do none of these things. I’ve been to their campus, and interviewed for jobs with their teams. To a fault, everyone I’ve met is full of idealism and optimism for the power of the Internet to empower individuals and improve society. I actually still like Google as a whole. But if the Internet is Pandora’s box, Google is the one that pried it open, and can’t quite figure out how to deal with what was inside. Humanity is not inherently good, and accelerating our lesser qualities isn’t having the positive outcome Google’s founders might have hoped for.
So, how do you throw the bath water out, but keep the baby? Can you use Google’s awesome tech, without contributing to the problems it creates? I don’t know, but here’s a few of the ideas we’re trying:
Diversify Your Information Holdings
I’ve said this before, and it bears repeating: don’t put all your eggs in the same basket. If you have a Google Mail account for work, have your personal account with another provider. If you use Google Classroom for school, use OneDrive for your private documents. If you have an Android phone, don’t put a Google Home in your bedroom. This isn’t just good security practice, preventing an attacker from gaining access to everything about you from a single hack, its good privacy practice. It limits the picture of you that any one service provider can make. Beware, though, of offerings that appear to be competitive, but are actually the same thing under the hood. The privacy browser Brave may tell a good story about how they’re protecting you, but their browser is based on Google’s Chromium, so its effectively the same as just using Google’s own browser.
I’ve been involved in software development for 20 years — data really does make software better — but did you know Google will willingly relinquish older data they have on you? All you have to do is ask. Whether you’re an active Google user, in the form of an Android device or one of their enterprise offerings (like Google Classrom), or just an occasionally searcher with an account, you should take them up on this offer and crank up your privacy settings.
Google is still pretty much the top of heap as far as search results go, but they’re far from the only game in town — and the deltas shrink daily. Bing is remarkably close in the search race, although its backed by an equally giantic corporation that is probably no more altruistic with their data acquisition, and DuckDuckGo does a decent job most of the time. Why not switch your default search engine to something other than Google, and switch back opportunistically if you can’t find what you need?
Check Who’s Watching
Just like Facebook has its fingers in most of the Internet, Google is everywhere. A service called Blacklight lets you plug in the address of your favorite website, then gives you a report on all the data collection services that website is cooperating with. The scariest ones are probably the ones you trust to give you news and information. Use RSS where possible, anonymizers, or different browsers for different purposes… which brings me to my final suggestion.
If I sound like a paranoid old man by now, I’ve earned it. I’ve literally been working on the Internet my entire career — my first experiments in web development date back to 1996. I love this thing called the web, and generally Google has been good for it. But a democracy isn’t democratic if its ruled by dictator, and the Internet isn’t open if its entirely controlled by Google. As citizens of cyberspace, you own it to your community to help it stay healthy, and as individuals, you owe it to yourself to practice safe web surfing.
Recently someone posted a great idea in a Facebook group for users of old Macs — but apparently wasn’t interested enough in the community to describe how it was accomplished. The idea is to have an external USB drive with multiple Mac OS installers on it, so you can restore or recover a wide range of Macs. Although all the instructions are online, they’re scattered across different sites that have to be pieced together. Here’s my attempt to collected everything you need to make a multi-OS recovery disk for almost any Mac made in the last 15 years.
Things You’ll Need
A working Mac with a relatively modern OS (I used a 2008 MacBook Pro running a patched High Sierra)
The install media (disk, disk image or installer app) for each OS revision you’ll want (see links at the bottom)
At least 100gb external USB drive (an actual drive — not a USB key)
Optional, the DosDude patchers for any Mac OS you may want to shove on an unsupported Mac
Optional, but recommended, the latest Combo update for each Mac OS you may be installing. (I’ll include links to download these at the bottom of this post)
Moderate proficiency with Apple’s Disk Utility and just a little bit of Terminal comfort
Note: You’ll notice that these instructions stop at Mojave. So do I. You could argue that Catalina’s murder of 32-bit apps was necessary for the ARM-transition — and maybe that was right for Apple, but you can read why I don’t think its great for consumers.
The Delete Facebook movement has been around for a while now, and I have to admit, the idea is tempting. The downside of allowing a single company to have such an outsized view into our lives has become increasingly obvious, while the benefits have dwindled. By design, Facebook is more than just a social network – its evolved over the years to become something of an Internet hub. Sure, there’s a lot less people playing Farmville, but it’s still the closest thing to a ubiquitous messaging platform we have on the Internet, so it’s hard to just turn it off. Short of writing a letter and putting it in the mail, Facebook is the one place where I can get a message to most of my extended family. And there are things to be said too (both good and bad) about Facebook Groups, where strangers with common interests can meet and create connections — most of my hobby projects have been significantly helped by members of one Facebook group or another.
So quitting Facebook might be going a little too far for most of us, but maybe putting some limits on Facebook’s reach can help. Here are some easy steps you can take to control Facebook’s visibility into, and impact on, your digital life.
Delete the App from your Phone… Then Put it Back
Facebook’s mobile app, whether on Android or iOS, has a staggering privacy impact. Except on the latest OS versions, most of these permissions, once granted, are permanent, and accessible in the background. Recent improvements to underlying platforms have revealed numerous “bugs” that have all the appearance of spying on users – even while the app is not in use. For example, Facebook helpfully asks for access to your Address Book to facilitate “finding friends” but can use that information at will to quietly strengthen its social graph (the powerful database that makes Facebook so interesting to advertisers and political parties.) Recently a former engineer reported that Facebook experimented with uploading all your pictures in the background to “improve performance” when you chose to post a picture on their site.
Obviously, it’s nice to have your social network in your pocket – it’s convenient and helps pass the time. But, giving away all your personal data seems foolish. Fortunately, there is a work-around, and its actually quite nice. By design, your mobile web browser is a “sandbox” – websites can’t get the same permissions as Apps can, so they’re intrinsically safer. And to make it more convenient, both Android and iOS allow you to “pin” a website to your home screen so that you can launch it just like an App. The experience is slightly diminished from the full App, but its remarkably elegant, and significantly less intrusive.
The process is slightly different for each platform, but it amounts to:
Open Facebook in a web browser
Find the browser’s menu, and choose the option to Pin to your Home Screen
Find the new Facebook “App” icon on your Home Screen and launch from there
Use Facebook more-or-less as normal
A nice side effect of this change is that Notifications go away. You can always launch the “App” to see what’s new, but you won’t get things pushed to you constantly. Facebook Messenger is a separate app, which seems to have less privacy issues, so it can remain installed to allow message notifications.
Put Facebook in a Box
This tip applies to both your phone and your laptop or desktop computer, although the process is a little different. It requires you to get used to having multiple web browsers – and keeping Facebook in a secondary one.
My strong recommendation is to use Firefox as your daily driver – it has an extension that can limit Facebook’s reach automatically. Chrome and Edge both are reasonable for privacy, Brave is better, but in other ways all of these browsers contribute to Google’s unreasonable control over the evolution of the Internet – but I’ll get to Google in another post. Suffice it to say, choose your main web browser and make sure you’re signed out of Facebook (and Instagram) completely on it. When you visit facebook.com from that browser, you should get prompted to sign-in – otherwise, assume Facebook is tracking you all over the web.
Once you’re confident that your primary browser is Facebook free, install and setup a secondary web browser that can be signed in with Facebook. Use this secondary browser for your Facebook community, and limit other web surfing. On a computer this is really easy – your computer comes with a web browser that should be your secondary browser:
On a phone this is a little harder, because you can’t completely change the default browser – the built-in engine will still handle embeds and links no matter what you do. But you can still follow the same pattern – create the Home Screen shortcut “App” using the built-in browser and install another browser to do most of your surfing.
Prune Your Timeline
Aside from its privacy issues, Facebook also functions as sewage run-off for some of the Internet’s worst information pollution. Political viewpoints turn angry during an election year (or pandemic) and sometimes it gets to be a little much. You may learn things about your social network that you wish weren’t true – or maybe you just need a break from all the memes.
Sometimes you have no choice but to just remove connections (de-friend people) if they won’t listen to reason. But often a genuinely decent person has just listened to a little too much Fox or NBC News and you need to take a break from the partisanship. It’s OK to “snooze” people or unfollow them. This allows you to stay connected, without having to get inundated with their ideology.
I don’t mean to suggest we shouldn’t hear ideas and perspectives that are different from ours – in fact, I believe it’s healthy to hear both sides of a debate… as long as both sides are rational, thoughtful and based, at least in part, on objectively verifiable reality, or reasoned interpretations of events. But not all opinions are created equal, and not all sources of information are valid. I’d advocate first for a loving attempt to reason, out of concern for a friend, but I’d also advocate (especially as my kids are moving into an online world) for a limitation of the pollution you expose yourself to online.
The Facebook timeline algorithm is tweaked for engagement (sucking you in) and for maximizing advertising impressions (keeping you on the site so you see more ads). It’s not a good source of information, any more than if everyone in town went to the same park and all started shouting our opinions at each other. Prudently manage who and what shows up on your timeline, or ignore the timeline entirely, in favor of personal interactions or Facebook groups that are healthy for you.
Set App Timers
If you use the Facebook app, or a dedicated browser, both Android and iOS will allow you to limit your time in those apps. You can use this for any App that you find yourself mindlessly scrolling through more than you want to. In iOS, it’s called “Screen Time”, in Android it’s called “Digital Wellbeing”, but in either case you can find it in Settings, and easily set a timeout in minutes per day. Of course, you can over-ride it if you need to, but it’s a good reminder to manage what you’re consuming in a given 24 hour period, and make sure you’re including other interactions and sources of information.
Protecting Your Brain
We don’t let our kids use social media yet – their brains are still forming, and they don’t have all the tools they need to discern what they may read online. But adults aren’t immune from the cognitive biases that can trick our brains into unhealthy patterns. Facebook is a relatively new kind of media – one that empowers peer-to-peer sharing and information dissemination much faster than what we had a generation ago. It has many incredible benefits but inherits all the same problems of previous kinds of media, while introducing a slew of others that humanity isn’t really equipped yet to understand. There are efforts underway to understand and improve how this kind of media works, but until those things mature and inform the evolution of the Internet, it’s up to us as users to think about and manage how we interact with technology and other people using it.
Even though we knew it was probably going to happen, when the lock down order came in from the governor, we didn’t really get a lot of time to adjust. The kids were in school one week, and at home the next. Church was meeting in person on one Sunday, and exclusively online the next. A series of probably-Providential events had happened before this, none of which were deliberately timed by me, but all of which turned out to be helpful in getting our little country church online in time.
At the start of 2020, we didn’t even have Internet in our church building — we would upload sermon audio using a 4G hot spot. That audio was recorded on a 2008 iMac that I found on Goodwill Auctions for $140, and I had just replaced the 2006 Windows Vista eMachine that was in the pastor’s office with a 2009 iMac that I got for $80. With this “new” hardware installed, we decided it was time to petition the church leadership for a stable Internet connection. No one was opposed philosophically — they’d just never had a need before. At the February board meeting, they agreed to my proposal, and later in the month, I camped out at the church for a day and a half to wait for, then help, the Internet installer figure out how to connect our 160+ year old church building to the digital world.
The lock down order came only a couple weeks later. The Internet was unreliable, because the rural infrastructure near our location had issues, the more-than-a-decade-old iMacs were far too under-powered for their new task, and a decent webcam was suddenly very hard to find either online or in near-by brick-and-mortar stores… but in a little over a week, we managed to cobble together a streaming system, do some basic training for the pastor’s family, and hold our church’s first online service. All while hoping this would be a very temporary situation. It was not.
At some point in the summer, it became obvious that the system was too fragile for reliable live streaming, and it became more pragmatic to have a pre-recorded service in-the-bag. This created a more fault-tolerant, less stressful experience, but it didn’t change the chewing-gum-and-bailing-twine nature of the system: it all just barely worked, because none of the pieces were ever intended for the tasks that had been thrust upon them. While it became safe enough for most church members to attend an out-door service this summer, others who are at higher risk to the virus, could not attend — and with colder weather looming and no vaccine in sight, it became apparent that online church was going to be a reality for at least a little while longer. The pastor asked for some options for a more permanent system.
I priced out three bundles — good, better and best; cheap, not-as-cheap, and spendy. The elders settled on a combination of pieces that straddled the mid-range. I got to switch from trolling Goodwill, to a picking out a choice refurb from Backmarket (a great place to get used high-end hardware). The new Mac Pro is commonly called the “trash can” for its cylindrical design — Apple later admitted the look left them “designed into a corner” then basically abandoned the high end market for most of 7 years. Its from 2013, but was way over-powered for the time, and still outperforms most of the stuff you’d find at Best Buy today.
We switched from the commercial Windows software, vMix, to an open source package that runs native in macOS called OBS — a favorite of video game streamers and YouTube stars. It starts up in a fraction of the time, and handles virtually limitless inputs with ease. Switching to an HDMI camera with optical zoom, instead of the cheap USB webcam, allowed us to position the rig at the back of the sanctuary — instead of consuming the front half of the room — which made it significantly easier to connect to the sound board, and have an independent audio mix, which will improve both the in-house and online experience, once we tune it.
Ben helped me set everything up, and we even rigged up an iPad based remote control, so if needed, one person can run both the in-house video screen and the online stream.
This week, we’ll do some training on the new system, and probably work out a few kinks, then I’ll report back to the hospital for a follow-up surgery for a blood-clot related issue I’ve been dealing with all summer. Next Sunday, I hope to be worshiping from bed at home while recovering from this thing once-and-for-all!
I recently restored a NeXTstation computer — the grandfather of Mac OS X computers (and therefore the great grandfather of iOS). It joins a network of historical Mac computers in my basement, but was woefully disconnected from them. A crude file transfer between a G4 Cube running OS X 10.4 could be established relatively easily using FTP, but I wanted NeXT to really fit into the neighborhood.
Every computer in the house can reach a common file share running on a Raspberry Pi, which serves up a folder over SMB and legacy AFP (AppleTalk), with an ethernet-capable OS 9 machine bridging to the LocalTalk-only Macs from the early days. Unfortunately NeXTStep and OpenStep support neither SMB nor AFP (technically one version of NextStep had a crappy AppleTalk implementation, but not the version I’m running.) What Next did support was NFS — Network File System. And fortunately so does the Raspberry Pi…there’s just a little modification required to make it work with old versions.
After following the steps here to enable NFS and establish a share (using the same folder as AFP and SMB), a fellow nerd on Facebook provided some steps to force support for older clients (NFSv2):
Edit the file /etc/exports that you made when you enabled NFS, and decorate your share with some less secure options (at your own risk — obviously don’t expose this to the Internet!) I also had to assign an fsid. Here’s what my export looks like: /srv/A2SERVER/A2FILES *(rw,fsid=1,all_squash,insecure,sync,no_subtree_check,anonuid=1000,anongid=1000)
Edit the file /etc/default/nfs-kernel-server as sudo
At the bottom, add the line RPCNFSDOPTS="--nfs-version 2,3,4"
Run exportfs and make sure no errors are reported.
Important note: Your server should have a matching DNS or Hosts file record, since NFS does a reverse lookup when you try to mount a share.
On NeXTStep 3.3, the NFSManager.app GUI was not able to successfully mount the share — I had to do it from the su command line:
Type su and hit enter — provide a password if needed.
Enter a mount command like: mount serverhost:/pathto/yourshare /Net/localmountpoint
So, for example, my server is on a Host named NetPi and the shared path is /srv/A2SERVER/A2FILES (since the path is also used by A2Server’s AFP share) and I want to mount it to a local folder on my NeXTstation called NetPi, so my command is: mount netpi:/srv/A2SERVER/A2FILES /Net/netpi
If you want this to run at every boot, use Edit.app as root to add your mount command to the end of the file /etc/rc.local But be 100% sure your mount command works before you do — this can prevent booting if its wrong. To be on the safe side, include a time-out and limit retries like: mount -o rw,bg,mnttimeo=8,retry=1 serverhost:/pathto/share /Net/localmountpoint/
In OpenStep 4.2, the NFSManager.app GUI did work, and Sophie’s blog shows how to use it. And just as a point of interest, following the same HostManager.app steps to tell NeXT about a LaserJet 4 compatible printer on the Network let’s it print too!
I put quotes around “their own” because, despite their announcement, everyone knows that “Apple silicon” is derived from the ARM processor — a family of chips most often used in phones and other mobile devices. ARM has been around a long time, and Apple invested in the company back during the Newton era. Intel has obviously been around even longer, but Apple’s use of Intel chips is the stuff of relatively recent history.
This marks the 4th processor migration for Apple, from the Motorola 6502-based Apple I and Apple II computers, to the Motorola 68000 family in the early Macintosh line-up, to the Motorola (and IBM) PowerPC of 1990s processor-war infamy. With each generation, Apple struggled to position themselves against the WinTel (Windows + Intel) hegemony. It wasn’t until 2006, when they transitioned to Intel, that Apple finally found their footing.
Since Steve Jobs’ hostile take-over of the Macintosh project in the early 80s, Apple’s philosophy on computing has been fairly “closed.” Jobs envisioned the computer as an appliance for average people, not a tinker toy for nerds. In the original Mac, this meant unusual screws and an absence of hardware expansion slots. On the iPhone, it meant a “walled garden” where only Apple-approved apps on the Apple-hosted App Store could be installed (unless you were willing to do some serious hacking.)
It took a long time to prove this philosophy out — it was almost a full generation before non-nerds were doing most of the computer shopping. But in many ways it paid off. Macs have a reputation of being stable, reliable machines, and iPhones are the mobile device most people want to own. iOS really represents the logical outcome of Apple’s trend toward locking things down: its an operating system that users aren’t supposed to know anything about, on hardware that customers aren’t supposed to be able to open.
On the Mac, though, there’s always been another layer: under the simple, friendly veneer of the user interface is a powerful Unix shell. And under the sleek case is fairly standard, commodity hardware. The implications of this for the Mac is that despite Apple’s attempts to end their life prematurely, people with a little know-how can keep their Macs running for years. Unlike phones, where people feel compelled (either by fashion trends, or security concerns) to buy a new one every couple of years (don’t do it!), an Intel Mac can last a decade or more as a useful, performant machine. Obviously this is a problem for a company that primarily sells hardware…
Apple zealots will tell you that the move to ARM will let Apple build smaller, faster machines with better battery life. They’re not wrong — ARM rocks for mobility. What they won’t admit is that the move away from commodity hardware will let Apple control the lifecycle of these new computers the same way they intentionally keep the lifecycle of their phones shorter than necessary:
With an Intel-based hardware platform, upgrades made for Windows PCs mostly “just work” in a Mac
It remains to be seen whether the heroic hackers of the world will be able to bring these benefits to new ARM-based Macs, but if Apple’s plan is to make Macs more like iPhones (which it evidently is), you can bet they won’t help us.
The move from PowerPC to Intel was a painful one for the Mac community. Software we owned stopped working, or had to be run through short-lived and poorly performing compatibility tools. Then there was the swallowing of our pride as we collectively had to admit that Intel really did outperform the G4s and G5s we were so proud of. But ultimately, the benefits for consumers outweighed the costs: it was the right move. Arguably, the move to ARM is significantly less urgent — granted, Intel’s track record over the past few years hasn’t been great, but they’re still putting out decent performance at a reasonable price point. Besides, the average Mac user doesn’t care what kind of silicon they’re running on — and they shouldn’t need to. But they should care if a company is deliberately steering them toward a platform of aggressive planned obsolescence and a treadmill of re-buying things they don’t really need.
I’ve put more than two dozen used Intel iMacs and MacBooks back into service for churches, students, teachers and missionaries — all well past the date Apple would like them to be running, and all stable, reliable and with half of them running Windows 10 at least part of the time. They’re really great machines, and I mourn the end of this era. Maybe Apple’s new products will be better than I think; I’m sure they’ll be sexy pieces of hardware. I just hope they don’t become sexy pieces of garbage in a couple years…
Last year, our church got a donation of an air hockey table, for use by the youth group. It was in nice shape, save for the electronic scoreboard, which wouldn’t keep score. It would light up and make sound when powered, but the score was stuck at “88” to “88.” We resolved to see if we could fix it.
The electronics were wrapped in a plastic banner that crossed the middle of the table, and had only a few wires running to it — one from a puck “catcher” on each side of the table, with a simple switch that was pressed when the puck was present, and one modified ethernet cable running to a small control panel that let you start a timer, or reset the game. Our initial hope was that it would just be a wiring problem, that Ben and I could fix together. It turned out to be much more complicated.
All the wiring was fine, and via various traces, ended up connected to a small logic board — this turned out to be the cause of the problem: it was dead. The board had a part number, but no amount of Internet searching could find a source for a replacement part. We theorized that the actual logic being performed was fairly simplistic, and that we might be able to replicate it with a Raspberry Pi Zero — roughly the same size, and equipped with sufficient GPIO pins to map to the existing wiring. Connecting the puck catchers was easy, and with a little Python code, we could count score and show it on a connected SSH Terminal. The control panel was a little more difficult, but we managed to find a solution for some of its basic functions. The remaining problem was the seven-segment LEDs that actually show the score.
For those we called in some help from an engineering student we know, and managed to come up with the logic to light the LEDs by reverse engineering an array of bit values that could be toggled via the Pi’s GPIO. In theory it was going to be possible to restore displays, and 90% of the functionality of the system. In practice, it didn’t work out that way.
The seven-segment LEDs work in pairs-of-pairs: two LEDs for each player, and double that to mirror the value on the other side of the board. 4 LEDs, with 8 wires each makes for 32 tiny wires that needed to be run. Only half that needed to go the GPIO, since the other half were just mirrored values, but that’s still a lot of wiring — turns out there’s a reason most electronics use a printed circuit board. I briefly entertained designing such a board, and paying to have it printed, but we were still going to run into voltage problems. Once the LEDs were running, very little voltage remained in the little Pi for the blinking lights and buzzers that make the experience fun.
After literally months of soldering, brainstorming, and frequently ignoring the now very-messy project out of frustration, we decided to abandon the banner, and go all in on a Raspberry Pi 3B+ with an add-on display. The logic still worked fine for score-keeping, although I had to come up with a new routine for displaying the score in ASCII characters that filled the screen. The girls each composed a little ditty that gets bleated out by the buzzer when someone wins the game, and Ben designed a mounting shim and 3D printed it. We removed one of the two support poles for the scoreboard, and mounted the Pi to the other — neatly running the wires up the pole. I used plastic cement to attach a reset button and the buzzer to the side of the Pi’s case, providing the key features of the original control panel.
After a successful beta test, we refined the design and improved the crude graphics a little, then installed our new system in the Air Hockey table using most of the original wiring. Its not perfect, but its quite elegant — and I was pleased by how much we all learned putting it together.