Bruce Perens

This is about a ham radio issue, and probably won’t be comprehensible to non-hams.

To give you a very brief background, I was instrumental in elimination of the Morse Code test for Radio Amateurs, the founding of Open Source, and the creation of Open Source digital voice for Amateur Radio. I still create, evangelize, and code today.

This piece is my opinion. I am on the advisory board of the M17 project, but am not speaking for that project, or anyone else.

My issues with ARDC go back long before their recent, very sudden, decision not to accept applications for grants from the M17 project any longer. But I should make this clear: Everyone at ARDC means well, and wants to do the best for Amateur Radio. Even if, IMO, they are failing at that in some ways.

ARDC’s decision to part ways with M17, coincidentally I think, was followed by the the sudden removal of M17 from MMDVM over what seems to be an interpersonal conflict. Wojciech Kaczmarski, M17 key developer, has responded very positively to my criticism that this came about due to his behavior, not just that of other developers. And we’ve formed an official M17 advisory board of 5 people, so that he can have more support.

M17 responded to the removal by creating a new WSPD (a next-generation digital voice software suite for Amateur Radio) fork restoring M17 and continuing advanced development, which I am told now has more vitality than other projects.

Wojciech and his collaborators seem to have architected a triumphal exit from M17’s involvement with ARDC and the MMDVM fiasco, releasing not just the WSPD fork, but an amazing hardware project called LinHT, which you can see here. What is LinHT? It’s a handheld transceiver with a 500 KHz I/Q modulated SDR based on Semtech SX1255 RF-to-digital transceiver IC, capable of all modes and as much as 2 megabit per-second communications, and including a Linux system-on-module. Because production of good HT cases isn’t something we have down yet, it fits in a Retevis C62 case. [We could use a good project with industrial designers who just make HT cases, visually desirable ones, and injection molded, please, not jaggy 3D prints.] Developers can use GNU Radio flowgraphs or C/C++ on the built-in development platform in the HT! Prototypes work today and an alpha production run is being sponsored by PCBWay. If you want to start working on the platform before boards are available, the SXceiver is a good way to learn SX1255 and develop for it. There is a SoapySDR driver for it, so you can get it running with existing SDR software, or GNU Radio, immediately.

LinHT is the most important hardware project in Amateur Radio today. Among other things, it is the perfect platform to run a variant of RADE on VHF/UHF. RADE may just be the future of voice, all voice, on Amateur Radio. It integrates what we used to use a CODEC and MODEM for into one unified component powered by machine learning (the non-hype version of AI). In tests, it works at a significantly lower signal-to-noise ratio than analog voice modes, even SSB, and legacy digital modes like D*STAR and DMR, while providing better fidelity, noise immunity, and fade resistance.

Even if RADE isn’t the next big thing, LinHT should get us there, because it provides a hardware platform capable of running new, experimental, and modified modes, and isn’t limited to just FM and FSK. Your smartphone uses complex multi-carrier modulation, HF digital modes use it, your HT should too.

I was involved in an effort to create a similar, though less capable thing, called Whitebox, with the technology of 12 years ago. The Whitebox effort failed because the design had too much electrical noise, and we spent too much time building the computer and too little working on RF. LinHT uses a pre-made single-board-computer with an ARM CPU containing two ARM A55 processors, a Mali GPU, and an Ethos-U65 microNPU to accelerate machine-learning applications, all combined on to a “stamp” form module, which is soldered onto the main PCB, and costs as little as $32 in quantity. Thus, very little of LinHT effort is invested in the computer side of the design, and the design can easily change as better single-board modules become available. And Wojciech and his crew are probably better RF engineers than we were 12 years ago.

I think ARDC has shot themselves in the foot, because they won’t be involved in the LinHT project. The reason, as stated at their new policy, is that ARDC will not accept projects where the community fractures, where they don’t work with their prospective partners well, or where they don’t communicate well with their users.

What’s wrong with that? Several things:

In the case of M17, the project released a few members who submitted their own grant request for M17 without the cooperation of the overall project, confusing ARDC about the project leadership, and then departed, by their own written admission, with equipment that was property of the M17 project and funded by ARDC. M17 is much better off without those folks, in my honest opinion. But the response to M17 asking ARDC what to do about the non-return of equipment was a reply that ARDC had no advice and would no longer consider M17 grant applications. Wojciech and the M17 project are proud of having approached ARDC in an open and transparent manner about this issue, despite the outcome. But the lesson learned is that it might be best for other ARDC-funded projects to hide such issues from ARDC. Any conflict, project forks, etc. Make your own moral decisions. And that’s even though we know well that forks are healthy for Open Source communities.

The second issue is that the things cited by ARDC as reasons for rejection are all issues of interpersonal relationships. This ignores a key distinction of Radio Amateurs and Open Source Developers: a lot of us aren’t neurotypical. I have noticed this especially in the most creative and productive among us. Many of us, diagnosed or not, exhibit traits associated with autism spectrum disorder, level 1. It used to be called “Asperger Syndrome”. We can be very focused on one topic, to the exclusion of others. We can be very persuasive in speech and writing, and of normal or greater intelligence, but we have problems with interpersonal relationships and the perception of emotional cues and nuance. We may be more comfortable talking with you on the radio, or using a key or keyboard; than we might be if we met face-to-face. And most critically: We’re not always good at getting along with others. I personally am not neurotypical, and took two decades to speak clearly and not walk on my toes.

Unfortunately, some of the ARDC staff obtained their ham licenses during the onboarding process, and can’t be expected to have a long familiarity with Radio Amateurs and our quirks. So, perhaps they missed this point about our community.

Thus, we have some mainly neurotypical people who wish the non-neurotypical folks would just act like them. And non-neurotypical people try. But emultating behaviors that don’t come naturally to them isn’t a simple thing, and they fail. ARDC could approach them with more sympathy and understanding, rather than the sudden, outright rejection that we have seen.

We also have a potential law violation. Since ARDC’s rules on rejecting projects for reasons based on interpersonal relationships are very specific to the developer community member’s disabilities, they may violate Title III of the Americans With Disabilities Act, and other law requiring handicapped accommodation in organizations; and also in connection with employment, since developers in question were funded to be employed full-time. Thus, ARDC might be vulnerable to lawsuits based on civil-rights law by any developer rejected under their policy who can show a diagnosis.

And I think there’s also a degree of conflict-avoidance in ARDC’s policy. But projects should stand for something! M17, for example, stands for the rejection of black-box technology like the AMBE codec used in all commercial digital voice systems, and its replacement with Open Source technology that hams own; like CODEC2, and later, RADE. That’s something worth fighting for, and about.

Other issues with ARDC:

There’s where the money came from. It’s from selling 1/4 of the IPv4 address block allocated to Amateur Radio to Amazon Web Services, for about USD$134 Million dollars. ARDC continues to hold the remaining 3/4, which is very sparsely used by hams. This address block, Net 44, contained 1/249 of all usable public IPv4 addresses worldwide, and was issued to Amateur Radio by the American Registry for Internet Numbers (ARIN), for free, in the early days of the Internet.

I am on record, on the ARDC mailing list archives, of having asked ARDC to return this IP block to ARIN once IPv4 addresses became scarce, and it was clear that there was no significant use of Net 44 on the air any longer. Amateur use of 1200-Baud TCP/IP, which vies with carrier pigeons for the slowest transport prize, had mostly ended with the rise of the Internet. Other responsible agencies, like Stanford University, HP, Xerox, APNIC and ARIN itself returned their similar blocks or assigned them to be redistributed, to hold off IPv4 address exhaustion until the world could switch to IPv6, which is still incomplete today. My request was refused by Brian Kantor and other people on the then ARDC mailing list (ARDC was not as organized when they only managed the moribund Net 44). So, IMO, Amateur Radio did not act responsibly, to the general public, in the face of worldwide IPv4 address exhaustion, with a very few people making the decision. This very few people making the decision would become a recurring motif with ARDC. Once the opportunity to make money came along, ARDC did release part of those same addresses, again not involving a broader Amateur community. And the income from those addresses, again, is managed and allocated, with more people than before because they have opened up committees, but still by a very small group.

Did ARDC damage Amateur Radio by the sale of its addresses? No. They didn’t ask your permission, but you weren’t hurt. Of course all new Amateur Radio development should use IPv6, which won’t exhaust its address space as IPv4 has. ARDC still can dispense very many more public IPv4 addresses than there are hams worldwide. And each of the remaining IPv4 addresses that ARDC can dispense comes with 4 billion IPv6 addresses. You won’t run out.

But the proceeds from that sale have made ARDC the 800-Pound Gorilla. They can do whatever they want, and they aren’t responsible to anyone but themselves and IRS. They have tons of money, and too many people want it. You may be able to influence ARDC by joining one of their committees, if they let you in.

Unfortunately, the 800-pound gorilla hurts people and projects, and not just with things like the sudden and unexpected termination of M17. One way is that they insist that your project be a non-profit, for their tax compliance reasons, but they don’t tell you about the bad things that will happen to your non-profit if you get most of your funding from just one donor, ARDC. That would be true for most projects that operate their own non-profits. It’s a part of the IRS rules called the public-support test, which requires that non-profit funding must come from multiple donors, not mainly one, and may result in your organization losing its 501(c)3 non-profit status.

Why doesn’t ARDC tell you about this? Their lawyers told them not to. So, off their funded organizations go, blithely and uninformed, into a pit of non-compliance with IRS regulations that may span years and be complicated and expensive to exit.

However, it is perfectly possible for ARDC’s lawyers to tell them how they could make sure their funded organizations are well-advised by a third party, even before applying for grants, without placing ARDC at risk. Structuring your organization to pass the public-support test is a pretty easy process, if you know what to do early enough. Making sure organizations know how to do that should actually be part of ARDC’s due diligence. That they let this issue persist, IMO, indicates that ARDC might need better legal counsel ( see this ) unless the lawyers have competently advised and there’s some leadership problem.

And finally, I think ARDC fails by exclusively funding projects, not people. This limits how much “R” they can really have in “R&D”, as you have to go into your grant application with a complete plan of what you are going to build. But real researchers start by exploring ideas, and don’t necessarily know what will come out of them. So, the act of creating something really brand new can’t be funded by ARDC within its own guidelines, and the “R&D” they do fund will really just be development, to build out something already known. I’ve discussed this, and the alternative, in more detail here.

So, what am I doing about this? I’m not being quiet about it, and neither should you. I have often sent my criticism to ARDC, and met with them. They’ve told me they circulate my comments, but I can’t say I’ve ever seen any advice put into effect. So, it became time to bring you into the discussion. I continue to work on Post Open so that Open Source developers will not have to continue to be supplicants, dependent on corporations, and grantors like ARDC, for all of their funding; but will receive fair compensation for their work.

I haven’t written this to facilitate burglary everywhere, but to complain about the poor security on my recently-purchased Forest River trailer.

Locks sold in the United States are generally poor. If you’d like to learn how bad, watch the Lockpicking Lawyer on Youtube. They are short videos that you may find shocking for how easily he can open the locks that are protecting your home, etc.

My trailer is a Forest River Rockford Geo Pro 15fbs. This is one of a few dozen different brands manufactured by Forest River, probably the most popular manufacturer of travel trailers in the United States.

The trailer body is a sandwich of fiberglass, styrofoam, and very thin wood; on an aluminum frame, but it’s mainly styrofoam. When discussing the locks, one should keep in mind that anyone with a pry bar can open one of these trailers easily, and bears have been known to tear off the doors. So, there is a limit to their security.

The entry door and outside cabinets of the trailer have locks manufactured by Global Link, which come with one of 90 different keys. The keys have codes stamped on them, in the range G-301 to G-391. Just 90 different key codes seems like a really small number to me, and key rings containing all 90 can easily be purchased online. But this isn’t the biggest problem of these locks.

For the convenience of Forest River and its dealers and service people, the locks are master-keyed: there is one key which opens all of them. Although Forest River and Global Link claim that master keys are available only to locksmiths, RV dealers, and RV service companies, you can purchase 10 of the master keys for $13 on Amazon, here: https://www.amazon.com/Camper-MK9901-Compatible-Global-Motorhome/dp/B09BBDZT84

If your RV lock has the letter “M” on it, it can be opened by the above master key. My RV had an “M” on all of the outside cabinet locks and the main entry door lock. A second lock, below the main one, on the entry door, which RV dealers call the “deadbolt” although it is not made of the heavy material of a real deadbolt, is missing the “M”. But most RV users only lock the top lock, and don’t even know what the second lock is for.

So, for $13, you can equip 10 thieves to enter many popular RVs without breaking in, and look like they’re legit: they got in with a key.

Global Link is not the only maker of RV locks with poor security, see https://framon.com/update/RV-2023.pdf for a discussion of a number of differently branded, through rather similar and equally bad, RV locks.

So, what should you do? First, don’t replace your RV door lock with an electronic combination lock. While these add a combination with up to 10 digits, that you would think might be more secure, they also include the same poor-security mechanical lock as a backup, which can entirely bypass the combination keypad. If you look at the replacement keys they sell, they generally offer no more than 60 different key codes. Hopefully these locks don’t admit the master key.

So, what to do? First, use the cylinder change key: https://www.amazon.com/Creative-Products-GLOBAL-LINK-CHANGE/dp/B08CD7BFTF to remove the top cylinder on your RV’s door. Order a non-mastered version of the same lock code on eBay here: https://www.ebay.com/itm/276120210472 and replace the old cylinder. Your entry door will now no longer admit the master key, but it will still open to one of 90 keys.

The Global Link lock is a “wafer” design which is probably easy enough to defeat with a raking attack, something that amateurs can do in a minute without learning lock picking. So, even master-proofing your lock might slow down a thief a bit, but won’t defeat every one.

Here are slides from my new presentation Mitigating Fading at the Antenna. I will do a video, but you can understand the whole thing from the slides.

https://perens.com/static/PerensFadingPacificon2.pdf

I am following up on the ARRL payment of USD$1M ransom to computer criminals. This would not happen with competent IT management. The defense is to have good off-site or off-line backups and to be able to identify the vectors to be blocked and restore your systems from bare metal when something like this happens. Not to pay someone a million dollars to leave you alone, and wait for the next criminal gang to come along for more.

The guilt belongs to ARRL Executive Director David Minster NA2AA. Before the breach, he had terminated the head of IT and was not able to hire or retain competent IT staff. Under his management, staff turn-over exceeded 50%. Many of the staff were in it for love of Amateur Radio, and ARRL, as a non-profit, was unable to pay salaries commensurate with the market. Thus, when the environment became toxic, many of the staff chose to leave.

The ARRL executive board, or at least a majority of them, are also at fault for their continuing support of Mr. Minster in the face of these issues.

I am on a well-known broadcast/podcast, I think Friday. If you are concerned that I haven’t given someone credit or that history has been misrepresented: I was recorded for three hours, and what they called me to fact-check would be just a few minutes on the air. The producer was really nice and did his absolute best, and I probably did say what you wanted, and because of time constraints of a 15-30 minute program where I was only one of several guests, very little made it on the air. History was also simplified by the program to make it comprehensible to the audience, for example USENET became a bulletin board and postings became tweets.

I do my best for the community with the press that I frequently have to deal with. Much of the community rarely deal with the press and aren’t aware of the issues, the main one being that I can only very rarely get them to say all that I want, the way I said it, and that my words generally pass through a reporter, an editor, a proofreader, and maybe other people before you read them.

Amateur Radio operators like to put antennas outside and in the air. Once there, we might not be able to reach them for a long time. But someday they will need to be serviced or disassembled.

Not everyone learns right away: you should use anti-seize compound when assembling your antenna. It should go on every mating of two fasteners, and every place where a tube slides into another tube. Anti-seize compound is a special lubricant, sometimes mixed with microscopic copper, zinc, aluminum, or nickel powder, which keeps metal fasteners from getting stuck. I once disassembled an antenna that had been left in the air for 15 years, at a site a couple of miles from the ocean and its salt spray. Because the assembler had diligently used anti-seize compound, the entire thing came apart easily, and all of the fasteners, probably stainless steel, could be reused.

You should be using stainless steel fasteners for things that live outside, as these will best survive an extended time outdoors without becoming a lump of rust. But stainless steel has special needs: The friction of two threaded stainless steel fasteners mating causes galling, microscopic debris that can lock two fasteners together, or they can cold-weld, which causes the two surfaces to fuse together. So, stainless steel fasteners always need anti-seize compound to preserve their ability to be disassembled.

Similarly, you should use anti-seize when you mate antenna parts like nesting tubes together. Or don’t expect them to come apart again in a decade.

Some people are concerned about the conductivity of the anti-seize compound, even going to the trouble of measuring its DC conductivity with an ohmmeter. This is a little naivé, because antennas carry RF, not DC. The metal-filled kind of anti-seize compound is more conductive, obviously, but in general you don’t need to be concerned about its conductivity. The metal parts of the antenna will generally DC-couple even in the presence of a non-conductive lubricant, the contact surfaces just push it aside and the lubricant fills gaps. And even if they didn’t, the capacitive coupling between the metal parts would be sufficient. Metal parts are in general not bare metal: both stainless steel and aluminum resist rust because of a protective oxide coating that they form immediately when they contact air. But these parts maintain DC electrical contact even though their oxides are insulators. They rub it off, or it’s too thin to matter.

What is the best anti-seize compound to use for antennas? Most anti-seize compound is designed for other applications, in particular use in automotive engines. So, you’ll see it rated for use with temperatures in the thousands of degrees Farenheit. Obviously we don’t need this for our antennas that probably won’t exceed 150° F.

However, our antennas are outdoors in a harsh environment. Acid rain is everywhere, and birds drop guano on antennas that becomes phosphoric and nitric acid through bacterial processes (for the bio heads: denitrification, glycolysis, probably others). When you have two different metals touching each other, and acid, you have all of the parts of a battery: a cathode, an anode, and electrolyte; and the result is galvanic corrosion, a form of electrically and chemically accelerated rusting. Even mildly impure water is sufficient to be the electrolyte, so the potential for galvanic corrosion is everywhere that dissimilar metals touch.

It’s a characteristic of the electro-chemical reaction of a battery that the anodes donate electrons to the circuit and the cathodes receive them. Anodes corrode as they lose electrons, and the cathodes may get coated with some chemical by-product, but generally cathodes don’t corrode. In our antennas combination of two metals, one will be the cathode and one the anode. Which is which? You figure that out with the galvanic series, a table of which metals are better, or worse, electron donors. But why would you need to know?

A very common strategy for defeating galvanic corrosion is the use of a sacrificial anode. For example, most boats and water heaters contain one, and you’ll sometimes come upon a marked cover for one in the street, as they’re used to protect most underground infrastructure. The sacricifical anode is generally zinc or magnesium, as those are on the bottom of the galvanic series. That means that other metal connected to the sacrificial anode will be a cathode and won’t be subject to galvanic corrosion, while the sacrificial anode will be, sacrificing itself for the good of the cathode!

So, this brings us back to the anti-seize compound. As I mentioned, these are available with zinc filling, aluminum, copper, or nickel. Our antennas are generally aluminum. Copper and nickel are both above aluminum on the galvanic series, and thus in combination with them, our aluminum antenna part would be the anode, and would be subject to galvanic corrosion in combination with copper or nickel anti-seize. This might be micoscopic, but the result is going to be rougher surfaces when we try to disassemble two nesting tubes. What if we use zinc-filled anti-seize? Zinc is below aluminum on the galvanic series, and thus the zinc would be the anode, and maybe some of the zinc in the anti-seize will sacrifice itself to keep our aluminum antenna parts pristine as acids intrude into the connection. So, I’m going to use zinc anti-seize on my antenna, even though a lot of hams swear by Jet-Lube SS-30, a fine brand of copper anti-seize sold by, among others, DX Engineering.

An alternative is the aluminum-filled anti-seize sold by InnovAntennas. In the joining of two aluminum elements there would be no dissimilar metal issues. However, entertainingly their own catalog page shows the elements secured with a stainless steel pipe clamp, making it clear to me that there would still be dissimilar metals issues. I’m going to stick with zinc where the pipe clamp would be, and with any dissimilar metal contact, but maybe this aluminum paste is a good idea for between the elements. Permatex also makes an “Aluminum” anti-seize, but they describe it as a blend of aluminum, copper, and graphite, so I assume it’s called “Aluminum” because it’s intended for aluminum engine blocks.

When would copper anti-seize be appropriate? Well, I think it’s made for car engines, where heat is a problem and the main material used is steel. Jet-lube might have been designed for aviation, back before aluminum was in as much use. Steel and iron are above copper on the galvanic series and thus copper could be a sacrificial anode. Engines often have a large sacrificial anode, or more than one, somewhere within both their oil and cooling circuits, and this will often be zinc or magnesium.

Copper anti-seize is also appropriate for stainless steel antenna parts, at least the ones that don’t touch aluminum. Zinc would work at least as well, including for parts that touch aluminum.

There are places that anti-seize doesn’t belong: don’t put it on moving parts, and don’t bridge things that are supposed to be insulators with it. That means definitely don’t put it on your coaxial connectors! Don’t put it where it might collect lots of dust, it’s meant for where metal is permanently mated with other metal, and won’t help elsewhere.

The new version is here. This is still not the result of a lawyer, and shouldn’t be used on your work.

I received a two-page “offer” in the mail of $47K for some real-estate I own, with a contract ready to sign. It wasn’t a real offer. I’ve done some analysis so that other folks can understand what’s going on when someone out of the blue sends you a letter offering money for your property.

The offer was from Richland Acquisitions of Dallas, Texas, owned by one Armaan Premjee. You can see a YouTube video of Mr. Premjee explaining his business here. He claims to be a “wholesale” land flipper. It seems that sometimes he actually does buy land. He just doesn’t buy all of the land for which he sends offer letters. Or maybe not much of it. Which seems deceptive to me.

What you read here is all my opinion, and I am neither an attorney nor a real estate expert. I’ve been sued for defamation before, so I am careful to stay on the right side of the 1st Amendment and the California anti-SLAPP law. (I won my case, and the court made the guy who sued me pay $315,000 for my defense.)

Mr. Premjee and his company’s offer included a contract that I would have been required to sign by April 10 if I wanted the money. It was March 20 when I received it. It had a few interesting things:

• It was binding on me, but it allowed Mr. Premjee and his company to back out any time right up to the final closing date, for any reason.
• It had a maximum 112-day delay between contract acceptance and closing, when I would supposedly get paid.
• It included a term allowing them to advertise my property before they actually paid for it.
• It included the fact that there could be a “double closing”. I think this means that they may sell your property to someone else at the same time that they buy it.

I wrote to Mr. Premjee asking for the name of his escrow company and why there was a 112-day delay between the contract acceptance and closing. I quickly received an email back from one “Brian Smith” stating “After taking a closer look at the property we have decided it would not be a good fit for us.”

So, the offer wasn’t real. I can guess a few things from this:

From the video referenced above, it’s clear that the letters are sent out in bulk. He talks about making offers in “entire counties” and scaling to “entire states”, for a set of properties in those areas that fit his initial parameters, stated in the video, things that make the property more easily buildable and more likely to sell.

Although the letters appear to be offers, I don’t believe that Mr. Premjee and his company have performed their full due diligence on the property before making the “offer”, as indicated by Mr. “Smith’s” email quickly rejecting my property once I inquired about it. They make no commitment to actually go through with the offer. So, for all of the people who return signed contracts, it sounds to me that Mr. Premjee and his company then take a look at the properties and decide which ones they are interested in, and maybe at what price.

And what about the “double closing?” One of the contract terms allows Mr. Premjee and his company to advertise your property before he actually pays for it. So, it sounds to me like the “wholesale” part is that he actually gets his own buyer for your property before he pays you for it, buys and sells at the same time, and pockets the difference. And perhaps if it doesn’t sell in those 112 days, he doesn’t buy it at all.

OK, clever “system”, as he calls it in the video, but in my opinion somewhat shady, because it’s not explained to the land owner what is really happening, and that the offer isn’t really an offer. Probably nothing illegal but I will check with the California Department of Real Estate.

At the link below is the first draft of the Post-Open License. This is not yet the product of a qualified attorney, and you shouldn’t apply it to your own work yet. There isn’t context for this license yet, so some things won’t make sense: for example the license is administered by an entity called the “POST-OPEN ADMINISTRATION” and I haven’t figured out how to structure that organization so that people can trust it. There are probably also terms I can’t get away with legally, this awaits work with a lawyer.

Because the license attempts to handle very many problems that have arisen with Open Source licensing, it’s big. It’s approaching the size of AGPL3, which I guess is a metric for a relatively modern license, since AGPL3 is now 17 years old.

Send comments privately to bruce at perens dot com.

License Text

February, 2024.

I purchased an HP Sceptre x360 16′ laptop with 3840×2400 UHD+ display. I got this because it was affordable and had about the highest resolution display unless there is some exotic and unaffordable 8K laptop out there. I am mostly doing 2D web graphics development.

For some reason this laptop came with the display disabled in BIOS, and at first presented a disappointing black screen when turned on. I think it was Fn-F12 that finally turned on the lights, or it might have been holding down the power button for long enough for the machine to clear the BIOS parameters. Perhaps this is why Best Buy offered an $800 “open-box” discount on what was obviously not an opened box. Once it booted into Windows and ran it initial software update, it immediately went into an HP BIOS updater and reflashed the BIOS.

Debian 12 installed and ran.

The processor is a 13th Gen Intel(R) Core(TM) i7-1360P. This is a hybrid CPU, the cores all run the same instruction set but they aren’t all the same. There are 4 P (for performance) cores which have hyperthreading and run at high frequency. Because of hyperthreading, they each appear as 2 CPUs to Linux, but the CPU is only running one of the 2 threads at any time. Hyperthreading reduces the overhead of multiprocessing when there are more processes to run than there are CPUs. Without hyperthreading, switching processes would otherwise require flushing caches and reloading the registers and the virtual memory system. There are 8 E (for efficient) cores which run at lower frequency and are power efficient. The E-cores don’t support hyperthreading and each appears to Linux as 1 CPU. The Intel Thread Director hardware reports statistics on these processors to the Linux scheduler, which can then schedule them as appropriate. You can manually choose which are used with the taskset command.

There are several issues with this laptop because Intel has not yet properly supported its own hardware on Linux.

The keyboard can be folded back and the laptop used like a huge, heavy tablet. Gnome/Wayland recognize when the display has been turned on its side, but don’t recognize when it’s right-side up again! You’re stuck. I’ve not yet found what is missing and where.

The Xe graphics work well enough to stream movies and run WebGL. There is also a discrete ARC 370 graphics controller which is listed as a PCI device, but I’m not at all clear if it’s being used. A driver is available for Ubuntu (because some Dell laptops are certified to run Ubuntu) and may not yet be upstreamed. I will investigate further.

The Intel IPU6 integrated camera is an incredible dumpster fire. I’m not kidding. It’s an integrated 5MP camera. Sounds simple. doesn’t that? All of those use UVC and work with existing drivers, right? Well, not this one. In order to save a few dollars at most, this is a non-UVC camera, and a lot of what would be in embedded firmware on a UVC camera has to run in user-mode libraries. Sounds bad, right? It gets worse.

There is a repository full of proprietary Intel firmware and image processing libraries for the camera, even though there could be no business-differentiating value in a 5MP integrated camera design which would justify making it proprietary at this late date, and camera image-processing, including autofocus, is well-known art and available today for other cameras as Open Source.

There are four large Intel git repositories in total required for the camera drivers, one of which has the main branch empty, just to give you a hard time, and you have to figure out the branch to check out and then figure out all of the dependencies to build this stuff. This is at least an hour’s work. Another manufacturer might have listed all of the dependencies, put everything in one place, and made it work by just typing “make install”. What Intel has done is make it work only where they were obligated to do so. Which is Ubuntu and Windows. Everyone else might be able to figure it out from that. This isn’t good community member Intel that I’m used to.

Once you get all of that working, it is not so simple as just loading a driver. The camera needs a user-mode pipeline to do its processing. That connects to a Video4Linux loop device so that the camera can interface with existing software as if it were actually a kernel device. It depends on V4l-relay, which doesn’t appear to be in Debian so I’ll have to find that. You can configure udev to start this pipeline, but it doesn’t appear that Intel actually documented how to do so.

After you do that, there is some additional configuration necessary so that the browser can see your device, which Intel has not documented. Some folks figured it out for other Linux distributions, and it looks non-trivial, with a lot of files modified.

In the interim, I bought a UVC camera on Amazon which I will hang on the back of the display, until some other poor soul in the Open Source world works this out entirely. Which Intel really should be doing with paid staff for more than just Ubuntu.