Nokia N900 Necromancy(yaky.dev) |
Nokia N900 Necromancy(yaky.dev) |
> We don't ask "why?", we ask "why not?"
It appears that in this case, as the original battery aged, its internal apparent resistance (ESR) increased beyond the original design expectations, to a point where the phone won't work when plugged in to a charging cable because despite the charging cable most likely being able to deliver sufficient power at DC, it had too much impedance to supply it quickly enough. When current is demanded from a source that has too high impedance to supply it, the voltage drops. This will result in significant voltage ripple to the power supply of the digital circuits, which can cause logic to not function correctly.
Adding a large capacitor basically replaced the filtering and stabilization role of the original battery.
Interestingly people often intentionally remove capacitors for side channel measurements and glitching attacks.
I recall having to spin with my finger in a spiral to zoom in.
They're lies, but the batteries work. There were Chinese lines manufacturing knockoff BL-5Js, and there may still be one or two, or just a bunch of crates filled with old ones. Source: still use an N900, but just for podcasts.
> 10-year-old if you count compatible Lumias
Worse, those Lumia BL-5Js aren't actually compatible. The slots cut into the battery aren't wide enough to fit into the N900. Unless you're willing to cut apart the battery itself, they're useless.
Even after all these years, my n900 has been my favourite phone.
Maemo Leste is still around, I just tried it on a PinePhone with a keyboard not too long ago.
It's not postmarketOS with a popular DE nor Android, but has a terminal, browser, media player, et.
What made them interesting wasn't the hardware, though the full keyboard on the N900 was useful but the fact that it was Linux phone based on Debian complete with the ability to apt-get install whatever the hell you wanted. Including compilers, developer tools, openssh, and whatever you could think of.
The big difference between the N800 and the N900 was that the N800 was more of a tablet form factor (released years before the ipad or the iphone were a thing). The N900 was smaller and had a proper phone built in. It could connect to mobile networks and make phone calls. The N800 was by operator decree basically crippled to be wifi only. Operators absolutely hated the notion of an open platform like Linux running on devices connected to their networks.
The victory Steve Jobs imposed on operators was basically getting them to beg on their knees if he would please allow iphones on their networks. He completely turned the tables on them. The first iphones were exclusive to some networks only and people cancelled their subscriptions if they were on the wrong network. That's why iMessage is a thing and SMS texts are a thing that is no longer generating meaningful amounts of revenue for operators. There was no off switch for iMessage. Steve Jobs basically told operators to take it or leave it. Likewise MMS was not a thing on the first iphone. Nokia mistook that as a fatal omission. A missing feature. The truth was that MMS was dead as a doornail the moment 3G internet connectivity became a thing. Why have operators act as a middle man? Steve Jobs had no patience for any of that.
Anyway, Nokia still obeyed the operators and it ended up crippling anything with software potential. There were big discussions about having Skype on these things. The N800 had that. And a webcam. You could make video calls with it. In 2007. The N900 did not have Skype. And it was positioned as a developer phone. Worse, it had to compete internally with Symbian and the Symbian team was in control of the company.
So, it was positioned as a developer phone and the N9 was launched (2011) similarly crippled in the same week that they shut down the entire team working on the OS. That was around the time Symbian lost out to Windows Phone and the beginning of the implosion of Nokia's phone division.
The key point here is that Nokia had an Android competitor long before Google launched the Nexus. Before they had the Nexus, they were dual booting N800s into Android. I flashed mine with a development build at some point even. Nokia screwed up the huge chance they had there long before the iphone was about to launch. The N770 launched late 2005. The iphone wasn't even announced until 2007.
Nokia did not understand what it had and crippled the platform instead. And then it dropped out of the market completely.
Such phones exist today, too. Sent from my Librem 5.
Like any other large company, Nokia needed time to adjust its strategy, whether it would have happened or not we will never know, because Nokia got Trojan horsed by Microsoft killing any plausible competition to windows phone, and the whole phone division as a result. I still find it incredible that it happened just like that.
There sorta was and still is, just it uses the same endpoint as APN/APS so if any telecom wanted to block it all iphone users on their network would lose push notifications too.
I'm glad telecoms got put in their place, used to be a give them an inch and they'd take a mile type situation. I dont miss their games.
I'm no genius, but I'm reasonably sure I'm not a slouch either. I've got a masters in theoretical physics, I've worked with and written software for four years, I take an interest in anything techy I come across. I've picked up the basics of population genomics and molecular genetics without assistance.
I still find that projects like this are essentially black magic to me. Why are supercapacitors necessary to emulate a battery? How the hell does someone know how to mess with a bootloader in order to get past an internal partition corruption? How do you even tell if an internal partition is corrupted?
This is all stuff that I find massively impressive and enviable, but unlike essentially every other topic I've turned my attention to, there doesn't seem to be any readily identifiable path to mastery.
I picked up Linux at 13, fortuitously just in time for the release of the Nokia 770 (later getting, and still owning the N900 too).
At that time, getting real dirty with the kernel, hardware, cross compiling etc was necessary, so 1) there were more resources 2) it was seen as mundane, busy work rather than mystical and difficult.
If I were to say how to learn the same things today, I’d probably say Gentoo is ideal - it’s insanely flexible in tinkering, has good resources on compiling the kernel and packages, and I’m a fan of crossdev for cross compiling.
Getting real dirty with hardware and electronics, the obvious answer would be one of the Raspberry Pi lineup, but if you’re very tenacious, patient and a touch unhinged, then I would actually say now’s the time to get in on RISC-V.
It’s still early days, so there’s lots of resources that have very thin abstractions between hardware <-> tooling <-> code. Devices are cheap and exciting. You’ll be on the same footing as most other people so you won’t feel like a dunce.
The cons are that a lot of RISC-V devices get shipped out with very little documentation (and sometimes only in Chinese), binary blobs making mainstream kernels difficult, and you’re learning at the same time, so you might feel you’re ice skating uphill.
Wrt to the bootloader and partition corruption; towards the twilight years of the life of the N900, when it became clear N900 had been abandoned and the N950 was still only available to select few, a bunch of smart people on the Maemo forums started reversing and writing open drivers (uboot bootloader, wifi, camera iirc), so they became pretty documented.
TBH, I didn't / don't exactly have a path. I started with Raspberry Pi (and Linux for the second time) 10+ years ago, which led me to Arduino, which led me to low-voltage electronics in general. At the same time, I had an unreasonable dislike of google, which led me to flashing LineageOS on a test phone, which then became my main phone, which eventually led to PinePhone, which didn't work out, but was fun.
Can you grab the current boot partition? Once you have it can you decode it? Do you have a reference boot partition? Can you extract the bootloader from the boot partition? Can you read those binary files? Maybe turn them into readable assembly?
Can you clip a multimeter onto a PC trace? Can you do the same with a scope? Can you decipher what the 'scope capture means? Maybe use a bus pirate instead?
It's all about observability.
I have friends who started with porting Sailfish OS to their old Android phone and now they are designing their own PCBs for their home automation system. They custom built their own RC cars, 3D printed their own ergonomic keyboards, designed their holiday decorative lights etc..
I have seen a lot of people in my local FPV Drone Racing community, who started with building their own custom drone and then moved on to 3D printing their modifications, tweaking their firmware, building their custom lithium battery packs, designing their ergonomic keyboards, and now fiddling with their home automation software/hardware.
Also installing Arch Linux onto some random piece of hardware, regularly following Hackaday like blogs seems to help.. lol.
Booting from an SD card, while possible, is rather impractical on N900 because it gets disconnected whenever you open the back cover.
The N900 that lays next to me right now still works as a phone. I have to replace its screen though, as recently it took some damage in my pocket and got a small crack in its bottom middle. Touch still works perfectly though, so I'm not in a hurry :D
It soon won't be. 3G and 2G network are being depreciated quickly around the world
Maemo wiki states that Maemo Leste should be run from SD card. I am actually surprised that the phone can use the SD slot at high enough speed.
I went to see District 9 in the cinema in Helsinki. Uh oh, the alien parts are only subtitled in Finnish and Swedish and my Finnish is not up to that.
I installed a BitTorrent client, found the release on Pirate Bay, successfully torrented just the subtitle file, and used an editor to read the subtitles for scenes with a lot of alien.
The N9 had much better UI, but there was something of the cyberpunk “deck” idea in that thing, it was great.
I did my master’s thesis on that device. I had a custom hypervisor running a guest kernel, virtualized networking, and a buildroot userspace. I could SSH into the host N810, then SSH into the guest. I even virtualized the framebuffer at some point and got the “dancing baby” animation playing from the guest. It only ran at a couple frames per second, but it was _amazing_.
Had to use a barrel plug to charge it.
Spent a very nervous and sweaty day figuring that out when I bought one used with no warranty or returns and it didn't boot properly =)
Planet Computers Astro Slide 5G also sounds like it could work.
Picture me in 2007. "The iPhone. Psh. Like I'm going to switch to Cingular and pay thirty dollars for a data plan!" (Keep in mind that's $47 in today's dollars!)
I would use my N800 and Bluetooth-tether to my Verizon flip phone when on the go. It was mildly useful for things like LiveJournal and I'm sure the Twitter of that time would have worked on whatever browser Maemo had. But I had to admit by 2008 that I wanted a smartphone instead of this second device with a stylus.
In those days though, browsing the web as though you were on a desktop was thought to be the goal to aspire to. Even the iPhone launched with the default behavior in Safari being showing whole desktop webpages, and you zoomed in to the parts you wanted to use. It took a year or two for people to figure out 'responsive' and within 4 years most sites were starting to be designed for small portrait screens. At that point the landscape N800 style was at a disadvantage since the mobile sites being designed to be a little leaner, were the wrong layout, but the desktop sites were pretty heavy for a mobile device to handle. And as "apps" ate the world that probably put the final nail on our little N-series.
That was the death of mine. I had an external battery charger that I could use to charge the machine overnight, but it was too much of a hassle so it got recycled and I moved on to a Galaxy Note, which everyone laughed at for being enormous but now look at us, the base iPhone 17 is around the same size...
The N900 was a great little device, it was like having a tiny computer with full keyboard in my pocket. It's just a shame the built-in FM transmitter didn't work reliably, because I used it to listen to music in the car a lot.
It was also amazing to be able to download the whole world's map data (such as it was in 2010) to the device, so the GPS navigation still worked off-grid (deep-outback Australia in 2010 was not always that good for data connections).
https://blenderartists.org/t/blendersito-is-a-blender-clone-...
Good read. I did the exact same modification to my N900; it was my internet router for several years. In those days, a mobile carrier in my country offered a relatively cheap data plan for dumb phones: unlimited data, but restricted internet connectivity with a WAP proxy. Fortunately, thanks to curl and busybox utils, I figured out that the proxy allowed CONNECT requests, and a tiny C program on my N900 was able to transparently tunnel all TCP connections to the internet.
BTW, my N900 had another mod: it featured a Type-A female port :)
Yet still there is not true successor, although I would expect that producing things like this became cheaper in the recent years.
The project used PayPal to gather downpayments, PayPal decided to lock the funds for months (almost a year or maybe longer IIRC) because they saw money coming in but no confirmation of goods coming out. And, you know, when it comes to big companies, no explanation is sufficient, you are guilty of something because some heuristic said so, so the funds were locked, legal threats didn't work (try threatening a company with the power of a small/medium country), and by the time they got their money back, key people who were going to work at a discount to cover key milestones had moved on.
Sadly, that day never came ...
> Nokia N900 enjoying its new life as an online radio device using Open Media Player.
But I agree with your sentiment. Using supercaps seems overengineered to me if the device is connected anyway.
The caps / supercaps are necessary to provide enough current during boot or more resource-intensive tasks.
Does it? I don't recall mine doing so.
I wonder if one could do Anti-Secure-Attestation, like, only allow connections from rooted devices? Back to proving root by running a service on the good old sub-1024 ports?
As it was basically like Debian Linux inside I could do what I usually do - write hobby projects and run it on the N900. I had my minicomputer emulator running. Nice to see my old favourite minicomputer editor on my N900.
Surely this is getting close to realizable by hobbyists or a niche company?
It wasn’t until I was at a friends home who had it playing in the background, I glanced at the TV and jokingly said I wish we knew what the aliens were saying…lo and behold, there’s subtitles.
It had the best slide-out keyboard of all the phones, nice and rubbery keys. Super smooth sliding motion.
It also had a FM Transmitter (not just Receiver), so I could blast audio in my first car back then without struggling with bluetooth kits & audio cable (neither was standard).
It also had an infra-red transmitter that was programmable, so you could use it as a remote in certain circumstances.
It the time, the 32GB storage was absolutely massive for a phone.
It also had stereo speakers & a kick-stand, so you could watch a movie on it without issues.
I really miss this phone & era. Maemo OS could've owned the market today, as at the time it was much better than early Androids. Nokia messed up so hard after this, the N9 was shitty in comparison.
I had high hopes for the Creative Labs Zii Egg back in those days, it seemed to me to be a better Linux-based phone-like device. What a world it was...
But Nokia did one massive mistake and it was to bet on Linux for this device. Even when they already had lot of Symbian experience, which also was week though when it came to user apps.
The modified BSD on the first iPhones was simply blazingly fast.
What killed these Linux phones was Microsoft doing a hostile takeover of Nokia. The owners of Nokia felt they couldn't compete with Apple's iPhone and decided to scuttle their business and transfer out as much money as possible to their own offshore accounts in the Pacific before the company going belly up. I think they could have competed if they weren't such cowards.
BSD and Linux are the same thing. That's the whole point of Posix.
What made the difference for the iPhone was that Apple's most expensive part of the whole device was the design. At the point it came out they had something like 23 years of very high end UX under their collective belts. It's one of the reasons why the little 128k Mac that came out the same year as the clunky old IBM PC AT was so expensive, too.
Good design is expensive, and it's the most important thing you'll spend money on.
Remember earlier in the week, all the discussion of Damn Small Linux and how a lot of the conversation around its UI was along the lines of "But I like it without all the wasteful whitespace" contrasted with "The whitespace at least needs to be consistent and the widgets need to look like they weren't thrown from the far side of a barn"?
The iPhone was out two years ago before the N900.
Nokia was already fucked because it had set up a system in which internal divisions designed competing phones, as a result it had flooded the market with similar but-not-quite-the-same handsets with overlapping features, and it had missed out on usability advances that iOS had made.
Symbian was undergoing an overhaul which would eventually lead it to be ’good’ again, but by then it was too late as Android and iOS were already eating its lunch. And around the time of the N9 launch (touchscreen-only Maemo/meego phone), Stephen Elop took the helm and issued the famous “burning platforms” memo which put Nokia on the path to windows phone exclusivity, purely to the benefit of Microsoft, who delivered the killing blow by first forcing the doomed “Windows Phone” onto them, then buying the mobile phone division so MS could churn out more doomed handsets for their stillborn mobile platform.
tl;dr - The company was a clusterfuck riding on name recognition and then an MS plant killed it.
Linux on the N900 was neither here nor there. It was a skunkwork effectively, a niche device for nerds (and a great one). But it neither sank the company nor could have saved it.
... while you were on the movies? That's "Mr Robot" level, kudos!
That's probably what I would do but that's also why the iphone beat the crap out of Nokia, because that example of what you did with the N900 is a 1% of 1% of what users would use their phones for back then, and Steve Jobs knew it so he won consumers over with a pleasant and simple UX that lacked features instead of piling on Power User features that nobody would use.
You're not gonna sell too many phone if your target userbase is those who know what BitTorrent is and how to use it on their Linux phone.
Such Power User focused niche devices are only financially viable for small companies to develop and sell, but you can't keep a company the size of Nokia in business by only catering to Linux phone enthusiasts.
Their demise was inevitable at that point no matter what they did.
They even include an owner in-joke, which means someone in the production must have owned one of these phones. Everyone I lent the phone to would pick it up the "wrong" way -- they would put the external screen to their face, like every other phone. But the mic and speaker were on the back. I had to quickly find the scene in the movie here:
There are some weird bits of the 900MHz band that cross into the fairly free-to-use ISM bands in some countries, and I recall a CCC talk where someone demonstrated a SDR setup doing mobile phone base station stuff by sneaking into what were ISM bands in Germany where he was that handsets would talk to because they were allocated as cellular phone spectrum in other parts of the world. Here in Australia we are limited and can't use the upper end of the 920MHz ISM band with LoRa devices, because Optus bought that spectrum for their phone network.
(Here in Aust4ralia we have other cellular spectrum and phone network problems, where a lot of older devices that support some 4 and/or 5G cannot reliably call 000 (our equivalent emergency number to the US 911), because the fall back to 3G when roaming onto other networks... A few people have died recently, and all the telcos are busy blocking a growing list of phones, mostly older Samsung ones if the noise in mainstream media here is accurate. I know my old but still otherwise functional Galaxy S6 Edge is not on the banned list.)
https://www.amazon.co.uk/Vodafone-075375-Sure-Signal-V3/dp/B...
Edit: same as already posted hackaday, oop!
The fact you could run apt on it did not help that much for the regular user.
> The underlying OS makes no difference. perhaps you've never experienced the bliss after setting up a BSD that just works 10 years after... and have never experienced the incredibly stable and snappy multiprocessing this miracle of a kernel (and OS) exhibits for decades now.
Let me tell you something - 20 years ago Linux was slow and unstasble as shit, and even slower on embedded. On the other hand FreeBSD and other BSD-derivates were super stable, but took more effort to setup and work with. They did not have the UI though, what Apple did was to wire their half-baked NeXT-inherited GUI on top of it and it flied.
I'm not even going to comment on the abomination called ObjectiveC, but matter of fact - the underlying OS workings were done in a brilliant way, WHICH, more than everything else enabled all the glitter tossed over the UI that you guys love so much. Like, there's a reason for game engines being written in C++ and not Python, right? Still a programming language though...
Sure lot of people adore what Sir Jony Ive did to the overall look and packaging of these products, and for a reason. But what truly distinguishes all these Mac products is what they can get out the hardware.
Sorry, but win3.11 did not work well on a 128kb RAM device. I've followed everything MS released since DOS 3.30 and witnessed firsthand the evolution of Linux and many of the distros. Nothing comes close to what Apple could do and is still doing with their hardware/software. No matter if you like Tim Cook (me personally - not) or Steve Jobs (very inspiring guy).
One of the reasons MacOS could draw attention from developers, who now form very important part of the user-base, is the fact they have a Unix-like thing at their disposal, and a very fast unix-like thing with some sort of a not-so-disgusting UI (wait for MacOS 26 though).
Nobody cares about darn window shadows, edges, or the unreasonable animation effects that we'd be turning off sooner or later.
If there was actually a holdup of funds that killed the project, and eventually the funds were released, that's an even worse story. I didn't think there could be a worse story. It would mean that the project fell apart while they were waiting on cash, then when they got it they just treated it like a personal windfall. IIRC I ended up out $1.5K on the thing.
I don't know who you are or where you were but the paypal problems were pretty well announced and _I_ was there. The project was already facing delays because a key person who was needed to make the board layouts was held up with the also now dead DragonBox Pyra handheld.
I have no idea where you think the money went, how much you think there was, but it was a constant game of trying to drum up enough activity to gain attention of potential customers to bring in enough down-payments to pay the salaries of the small number of people working on the project while having enough money to buy parts. The money went into salaries, parts, nobody bought hookers or blow with it, and certainly nobody got anything but stress out of that project in the end. I believe Joerg Reisenweber paid a lawyer out of pocket to try to get PayPal to release the funds.
Most of the dates stated are just plain wrong.
The UK dates are completely wrong - by 5 years in most cases.
All of the UK's 2G networks are still running, and the last won't be switched off until at least 2030.
Just ask the person to say a naughty word, I guess?
Many countries/carriers in europe have already shut down 2G, many will shut it down in 2027. A few will keep it a few years more.
My fancy new 5G smartphone doesn’t work in rural parts of the country. We are going backwards.
The whole point is to free up spectrum, how would that work if that spectrum is still in use for the American carriers in countries that shut down the service for domestic use? Why would service be maintained for such a niche usecase?
please note that the list is not fully up to date, eg. in Germany Voda and Telekom have said that they will sunset 2G in summer 2028.
As for the N9, it still has the most modern and beautiful GUI of any smart device, 15 years after being discontinued. It will take at least 15 more years for iOS or Android to reach that level, if ever. The physical design was also very nice and refined.
For those who don't believe me: https://www.youtube.com/watch?v=iCFNXhiFnKY
As for the N900, I just wanted one or two more hardware iterations (the design flaws were annoying, and a couple mentioned in the OP.) The N9 looked great, but I couldn't get over the loss of the keyboard (although of course that was our dictated future.) The Meego transition seemed unnecessary and annoying (not the UI changes, but everything else), especially the move to rpms from debs. They were just hostile to Debian mainline for some reason; if they had been less hostile, their work would have survived without a break even after Elop intentionally tanked the company. That proprietary moat is just irresistible.
iOS and Android literally grabbed the designers of Maemo/Meego and WebOS to update their horrible UIs. Back then, they were still even refusing to multitask. Android copied WebOS almost exactly.
Going by the username having a near slur in it. Probably that.
My BSD statement stays, though, MS did some very good work with the WinPhones, and in fact they were super snappy and useful, very close to what iPhones were at the time. And let's not forget, that the flat looks of (not sure which macos) was directly influenced by these winphones...
I don't think I ever saw N9 or any of the N9XX phones in real life.
That was never the problem. The problem was how could Nokia stay competitive to Apple and Google when their focus was selling Linux devices aimed power users.
Not good enough to save it from the iPhone onslaught.
People on HN make this mistake to assume that they represent the "average user", the same mistake Steve Wozniak made. No, the average user wasn't interested in all the features of the N9. They much preferred the simpler iPhone and the proof is in the pudding.
>and it seems it sold very well in the few markets where it was actually sold
How do you know it sold well when Nokia never release official sales numbers for the N9?
Estimates put the N9 at less than 1 million sales in the 23 markets it was sold in. A drop in the ocean compared to total iPhone sales of the same timeframe in same 23 regions which estimate at 50 million total. Face it, the N9 was a sales flop no matter how you try to spin this, and launching in more markets would not have moved the needle significantly to make a dent in the iPhone.
>(sadly no major markets)
It launched in 23 markets mate, mostly EMEA and Asia. Not NA because even Nokia leadership realized the N9 won't stand a chance to compete with the iPhone and Blackberry on their home turf.
>A decision which many people called out for being stupid before the consequences had fully unfolded.
Nokia was already dead man walking even before that. Even their own employees said so when they got to play with the first iPhone in their HQ. The N9 was the band playing on the decks on the Titanic.
Blaming Nokia's inevitable failure on Windows Phone is historical revisionism. They would have failed either way since they lacked the software ecosystem beyond the phone that Apple and Google offered their users.
Fanboys praising the N9 as something that would have magically saved Nokia even they have done X or Y or Z with various Linux spins, are huffing some top end copium.
In contrast the explanation I have for Nokia's failure gives a logical explanation: They panicked, prematurely declared thir existing phones obsolete, cancelled there next-gen development such as N9, and instead offered a poorer product (Windows Phone) at a later time. It is difficult to see how this can lead to anything else than failure.
Whether N9 and co. would be successful enough to save them in the long-run is pure speculation, but I see no fundamental why it could not, and it was ready at a time where Nokia was still big enough to get some app developers on board.
For me that is not even in the same league than the N900.
There was another phone with keyboard around the same time, but I forgot the name. That was claimed to be very much in the spirit of the N950 and its cancelled follow-up, the Nokia Lauta.
All with Sailfish, the spiritual successor of Meamo/Meego from Nokia.
You need to operate this as a business first, with the freedom part being a nice bonus. Nobody cares how free your thing is if it's dead on arrival and gets beaten by an entry-level smartphone.
Make a competitive product. Nowadays that could very well just mean Android with manufacturer-sanctioned root access and preinstalled terminal & X/wayland server for those who want to run desktop apps.
The Jolla phone someone linked below actually looks like a decent product. The Android app support means it's actually usable in the modern world, and the specs look competitive.
Or do you remember the Beepberry/Beeper?
OTOH, your phone is more than capable, so maybe a small bluetooth keyboard is all you really need. There are apps like iSH on iOS or Termux on Android that give you a Linux shell.
Not only that, but you should not get suckered down into overcomplicating things by chasing complex novelties, e. g. integrated slider- or clamshell-implemented keyboards, silly and outdated form factors (clamshell UMPCs, OQO already showed the way), etc.
You want a good, small keyboard? Design it to be attachable. This is possible in a variety of ways and can be adapted to your manufacturing expertise. It also leaves open third-party hardware support for your device. Not to mention maintainability/repairability. It's utterly puzzling to me how many hardware start-ups already fuck up the basics.
And never forget: In a satured market, even catering to a niche, means you should go for a somewhat unique feature set. How many ultramobile devices are out there that are truly accessible and usable? That goes beyond just safety or repairability.
OLED screen? I'd rather prefer something PWM-free. Precision control? Digitizer/stylus support. You don't even need to house the stylus in the device. But it would be very useful to have at least one. Audio? Yeah, 3.5 mm is a must. Dedicated, easy-access mSD (Express) card slot? Yes, please. Exchangeable batteries? Good idea, as long as it's a standard design in good supply. Kill switches. Maybe a modular camera set up like those Chinese flagships that are otherwise rather useless. Full-feature connectivity (1-2 x USB 4). Etc.
Get one of the BB + USB-C keyboards available.
But then again, experience shows I'm wrong.
> There was another phone with keyboard around the same time, but I forgot the name. That was claimed to be very much in the spirit of the N950 and its cancelled follow-up, the Nokia Lauta.
Probably F(x)tec and their successors. Those have a similar small keyboard as Nokia N900 and Nokia N810
There's also the Hackberry. This device uses a real Blackberry keyboard, with custom firmware. It works together with a 3D printed case, and a RPi CM5. This keyboard, while small, is very ergonomic.
If you do decide to buy one, the hardware (of the Pocket 2 at least) is OK. Easy to open and such. You can even buy replacement battery. Buy one from Amazon in EU if you can, they have to provide 2 years warranty, at least.