I loved always having access to a physical keyboard and an actual root linux shell out of the box without any hacking. I loved that my n900 behaved more like a computer and less like "mobile" device.
Too bad Nokia decided to end support for the n900. Most of its developers and lots of its users abandoned ship, and now I only use it as a camera because the n900 is stuck in 2009, and the web in 2016 is so bloated and heavy that nothing- not even the web browser or the unupdated apps- are usable now.
Things like this is what makes me feel that the mobile market came to a screeching halt somewhere around 2007, and has barely been picking up speed since.
Also, look into the neo900 project and Pyra. Neo900 is trying to upgrade the n900, and Pyra is the successor to Pandora, which is a pocket computer in a semi-similar vein to the old nokias. Neither is finished, but they're worth a look.
The neo900 makes me sad, because it's basically my ideal phone, but the price is just too much for me. I know it's not their fault, the low production quantities will do that, but still. It's an updated n900, with a fully FOSS stack and an isolated baseband to boot... awesome.
Plasma Mobile and Tizen are my only hopes for getting a non-Android Open Source OS on my phone.
The other project was called Moblin. It was an Intel project.
It's a shame the N950 never got a proper release. The N9 got praised highly on release, but the N950 would've been the true successor to the N900.
Reading this made me so sad. This is exactly how I feel. Nothing will ever be good enough again. :(
As mentioned, it'd be a usable device on its own to this day. If newspaper sites hadn't become unusable without an adblocker, if I didn't need browser updates every nanosecond and if the latest memory-hogging web trend language wasn't, well, what it was. I'd still use it.
Now I own two mobile devices. A BlackBerry Passport. Great UI, physical keyboard, indestructible. Dead software ecosystem, horribly closed source, last of its kind and while there's a lot to love It's not hard to see why. Then there's a rooted HTC phone. Great choice of software (termux alone makes up for so much). Hate hate hate ssh'ing with a touchscreen. Typing generally. Terrible battery life. Seems like it'll die from one harsh drop.
If my social life would not force mobile messengers on me, I'd have gone back to 20€ dumbphones and text/voice only. Though I bet even those come with a horrible browser and the accompanying remotely exploitable holes now.
All in all? Wonderful project. I'm very much hoping for a third player in the mobile game.
I used it as my main phone when travelling for years and it worked everywhere I needed it to when other phones wouldn't. I'd use it for getting unwacoughted access to WiFi when I was in a tough spot. I'd manage my servers with it. I'd use it to emit music through the radio on car trips and such (why is this no longer a thing?). I could even push my own programs on it without having to make a sacrifice to the phone gods. I'd also kill for a keyboard like that again.
But no, instead you have to get a hobbyist device or get one of the many shiny slabs running (poorly) one of three (and I mean, really, two) ecosystems.
So many of us miss that phone.
Android is 100% native code since that version.
Even the changes in Android 7, although they use an intepreter written in Assembly, the code is JITed with profile information, both are cached and it gets ultimately fully compiled to native code when the device is charging.
Also in case you didn't notice, Apple is pushing for LLVM bitcode for their newer OS versions.
There is even a BOF on this weeek's LLVM meeting about how to make LLVM IR fully hardware neutral.
Just about every Android device I've used has been painful in some way or other and invariably I've ended up rooting it and installing a custom firmware, generally the most light-weight one I can find.
Unfortunately, despite these antics Android remains annoyingly prone to sluggishness, freezing, reboots and various other random problems.
Ultimately it seems that with Android you're either forced to use a stock ROM which imposes a heavy memory footprint due to all the "wonderful" tweaks Samsung/Sony/HTC/Xiaomi have made or drop down to a user-built AOSP-based variant that is inevitably plague by its own issues.
I'm horrified to think this kind of thing is the height of the mobile user experience.
But on my Galaxy S4 and my current Galaxy S5 (which is updated to 6.0.1 I think), things work quite well.
It's not perfect of course: I'm using the regular Samsung/Sprint ROM which is absolutely loaded with crapware. I disabled all the crapware, but it's still there wasting a lot of valuable space out of my 16GB. It could be a little faster at times, like with the camera. But freezing and reboots just aren't a problem the way there were on old Android versions that I've used.
I do plan on trying out CyanogenMod sometime when I have some free time. But for now, the carrier-pushed version seems to work decently well. I don't have any giant complaints.
I have seen pretty sluggish, full of memory leaks and dangling pointers, C code written at enterprise level, by developers that couldn't care 1s about code quality.
Your points are still relevant, don't get me wrong, but the system isn't running Java bytecode specifically.
Google just used the Java language and APIs. But never it's byte code or VM.
Nowadays, the next best thing is the Pyra, but that still has't been released.
By the way, have you seen the neo900 project? It sounds like you would be interested, although it is also incomplete.
The n900 is everything that I thought the g1 was going to be, but wasn't. In a way, it was my fault for blindly creating my own hype over a first generation product I knew nothing about. The g1 turned out to be my bad and disappointing foray into smart phones.
Also, yes I have heard of the neo900 project, but am sad that the project is so far from its goals. There will never be enough n900 users to meet a crowdfund. :(
Another interesting project which might fill my desire for a true mobile computer (if only I could afford it) is the Pyra, the more powerful successor to the Pandora. But at $~500-$~600, it may not be worth it, and it still hasn't actually released yet (although the company did make the Pandora, so they have a history of Getting Stuff Done).
If you're going on the cheap, you can always hack together a Raspberry Pi and a tiny screen, keyboard, and battery into something semi-workable, and use a phone for the rest.
You can get a GSM board from Adafruit fairly cheaply, which would give you 3G data and voice/SMS on a Pi. You'd have to do your own wiring, but if you're building something like this that wouldn't be a problem. Throw a Ting SIM in there (or your other favorite carrier) and you're good to go.
Basically a mobile phone version of Bunnie's Novena laptop, though not quite as open source (at least not until Broadcom finishes opening up the RPi's processor).
On top of that, over the years, the kickstand broke, the screen got scratched badly, my belt clip broke in multiple places, the SIM slot became really finicky, and even the plastic around the keyboard was starting to come out.
And a lot of the problem was performance: a 600 MHz OMAP3 was great in 2009, but by 2014 it was just too slow to handle modern websites, and the browser based on Firefox 3.5 (maybe 3.6?) couldn't even handle some newer websites. And to be honest, the RAM was low even for its time... only 256 MB, operated as a unionfs with the rootfs, plus some of the (very, very slow) eMMC was formatted as swap. Over time, most apps stopped working: I remember using Facebrick to access Facebook for a while, and it was nice, but eventually the author stopped updating it, and Facebook changed their API, so no more Facebrick. It got to the point where I'd only use the browser, terminal, camera, and dialer, because that's all that was useful.
About a year after getting my Nexus 5, I booted up my N900 to get some photos off it, and I found I had difficulty emailing them to myself because Gmail had changed their API and flagged my N900's connection attempts because they used the old API that was deprecated for security reasons. None of the emails went through, and when I got back to my desktop, Google took me to a page telling me it thought my account had been compromised.
My batteries are swelling. I suspect it's no longer safe really.
:-?
Universal battery chargers are a life saver if you have multiple types of batteries to charge.
I'll give you the keyboard though. Nothing beats a full keyboard.
How much extra power do we have to cram into the device just to achieve the same performance?
Back in the late 90s/early 2000s, you could just run Linux on your x86. Sometimes half your hardware didn't work, but as long as you could get Ethernet and display you were in pretty good shape usually. The thing is, you could install it. You could try it.
With ARM, everything is so specialized. Looking at the forum posts for this thing, half of the threads are about porting to devices.
The manufactures have no incentive to standardize the platform. They customize to a device and it's better for them if you just throw it out in two years when they drop support.
A limited number of hobby devices now use the device tree configuration, but the ARM platform is far from being as standard as Intel or Power.
If the Ubuntu Edge had been funded, I think we'd be in a difference space with at least a handful of devices that had some standardization (if only to accommodate Canonical). As it stands, ARM fragmentation is a huge problem for hackers/devs.
Also SOC manufacturer usually has the drivers reference code but he might not publish it.
Generally not. Linux kernel modules generally can't be loaded by kernels with different versions, as the ABI constantly changes. It can be hard to use them even with the same version, if it was compiled differently. This is why they're always pushing for the drivers to be open-sourced and submitted to the kernel project for inclusion, because maintaining them separately is a giant PITA.
Phone makers don't bother because they don't care: they're not going to maintain the drivers very long anyway.
ps: I was actually looking into FDroid to free myself a little from Android locks, wishing for a simple linux + ssh + any signed code I wish.
The polish it has is unbelievable.
The question is: how? The website doesn't make it very clear to developers (after a fairly quick skim on my part, at least) as to how they can help with the initiative. I have a Nexus 6p, and it seems like my device isn't supported, but how can I maybe help with porting to a new device?
It had most of the downsides of android (closed drivers, no updates after 6 months, buy a new device or get exploited, beholden to vendors, OEMs and telcos which invariably suck, central control over software distribution), and zero of the upsides (abysmal performance [even the dialer was a laggy mess], lack of software)
Except for the fact you might trust Mozilla more than google, it wasn't anymore meaningfully open than android, and suffered crippling drawbacks in comparison. Nobody except a few developers cared that you can write apps in html/js.
The leadership was caught in a innovators dilemma (Symbian was aging but had a massive install base, Maemo has future potential but could massacre their Symbian earnings in a transition), and the board in the end brought in a Trojan horse that made a deal with Microsoft.
Also, what you are currently seeing happening in desktop Linux land, wayland, systemd, etc etc etc, is because of lessons (X11 and sysvinit is crap) various devs "learned" while working on Maemo in some capacity or other.
yes they may not be the perfect fit in a power constrained mobile device (My N800 seemed to last me just as well as the Android devices i have around me right now though), but why the hell are we turning the whole DE stack into a mobile phone in response?!
And the lie keeps being repeated, it was the board that sold Nokia by having a contract that would grant Elop a huge bonus if he managed to do sell the company to someone else!
http://www.forbes.com/sites/terokuittinen/2013/09/24/nokia-a...
"According to changes implemented in 2010, Elop was entitled to immediate share price performance bonus in case of a “change of control” situation… such as selling of Nokia’s handset division."
When this was known it was all over the place in European news.
However Nokia as a company did not go under; it is still very much there, particularly in telecom infra business.
Some customers, particularly very many in the USA, are not so happy to have Huawei or ZTE core switches etc, but in emerging markets they are less worried about Chinese intelligence infiltrating them, and also in Europe there seems to be less protectionism and concerns about national security. So the Chinese are catching the global market, without the Americans noticing much.
That explains a lot about you: although mostly your preference for strong typing. Working for a phone company will do that for you.
No, what explains that is that I used C for the first time in 1993, yet I was already coding since 1986.
So I was doing systems like level programming in Timex Basic, Turbo Basic, Turbo Pascal, Z80 and 80x86 on those in between years.
Also had friends using 68000, AMOS and GFA Basic on their Amigas.
Also I majored in systems level programming, with emphasis in compilers, which meant I have spent quite some time researching old programming environments, and still do. Better than reading newspapers sometimes.
So I learned that were better ways to do systems programming without having the compromises of C.
Plasma sits on top of the Android HAL like Firefox OS did. Nothing different there. Go buy you QC license and see for how long you'll get blob updates.
You're totally wrong about the performance penalty of FxOS compared to Android on the same hardware, but I'm done arguing that with trollaways.
In 2012, at the Akademy (KDE conference) in Tallinn, I spoke with Plasma developers who were, at the time, working on a tablet called "Vivaldi", which was supposed to ship with Plasma Active (the KDE 4 version of Plasma Mobile) based on Mer (the community fork of Meego). And they really struggled with the hardware part: They carefully assembled a BOM of parts for which open-source drivers were available, ordered test devices, built the whole software stack. And when they were ready to start mass production, the manufacturer was like: "What, that chip? That's soooo last month. We have this new chip here, which is completely undocumented, but we have an Android driver for you." Last I heard, they never got an actual device shipped because of that mess.
I don't know if it's gotten any better since then, but I can easily imagine how that situation led them to base the new Plasma Mobile on the Android HAL.
As this catches on, it only helps improve linux hardware (ARM) support, telephony applications for linux, and the like. As well as being able to have (or make for others) a truly remarkable user interface and apps. This is a mobile computer, running linux, running a desktop environment (not webkit, or Java VM, or closed source code) that only limits what you can make by your skill and imagination.
Yes it looks pretty mediocre now, but that's because its waiting on you to get involved.