Well note I what I was replying to. What I said is that as long as you keep doing this then _anyone_ will be able to just put regular Android on it and do the same for a fraction of the price or even for free.

So yes, Android, AOSP and Hybris are hammers. I'm thinking there's a huge world out there of hardware where running Android would make no sense, but this Android-as-hammer requirement means people just don't see it.


That said, I obviously also have a huge anti-Android agenda. So here comes the off-topic part...

It's a bit ironical that the quote you mention comes from a person who has succeeded at exactly that.

"Ruinous compromises" for me.

Last time I was actively trying to port the Linux kernel to anything, it was the day I fried my Palm T|X*.

Well, that was in 2009 more or less, so I can tell you something: things have not improved. Running the software you want on the devices you own is _at least_ as hard today as it was 6 years ago, probably harder since nowadays the minimum amount of supported hardware for a device to be considered "usable" has increased dramatically.

Android has done _absolutely nothing_ regarding hardware freedom.

It just happens that it is a POSIX-like operating system and thus running most programs is much easier... but so is iOS. But from a hardware point of view? No improvement at all. Same as Windows CE days. Even words from WinCE I believe should be considered _blasphemy_ are still in use today (e.g. BSP, ROM), thanks to Android.

* When I fried my Palm T|X, I suddenly found myself without a working PDA and without money. Alongside many Palm PDAs, I found a cheap used N810 on eBay, so after a bit of thought I decided to take a risk and try it, instead of getting another Palm. That was 2009. The rest is history

I would abstract it even further and say user experience. You do not need all the services - even Apple did not try to reimplement Facebook, Twatter or Gmail. Besides, if you want to disrupt services, it would be more effective to do it directly, not around the long way by developing a new mobile platform.

It is not easy to define a good customer experience. It is much easier to define a bad one. For example, a mobile phone that cannot even create a contact from a received call or SMS. Or another example, I am typing this already the third time. I lost the first two because 1, my phone rebooted (again) in my pocket during a forced break in typing and 2, I foolishly wanted to check the weather forecast and the browser reused the same tab in which I was typing this. I am not going to name and shame the device in question but it does not shout "good user experience" to the world, does it?

Dave is right, the users do not care (much; I will get back to that) about the OS. They care whether the gadget does the job and whether it does it better than the competition. The only time the OS comes into the equation is for the availability of apps.

This is so obvious that I feel embarrassed to even mention it. Yet I keep reading about how important it is to open the sources and similar insignificant details. The user does not care! What good is it to have the most advanced and completely open OS if the gadget does not do what it is supposed to do? Or, if it does, but another, cheaper gadget does it just as well or better, why should you choose this one?

In my fictional world, a brave new entrepreneur decided in 2011 to make just such a device. He came from the background in a large company and knew how such companies work. He took great care to avoid the arrogance, hubris, lack of flexibility and other ills such companies are usually plagued with.

After 2 years of hard work, my fictional entrepreneur released a new mobile device at the end of 2013. The device was fresh, it was new, it challenged the way we were used to seeing mobile computing and our relationship with the producer, yet it was pleasant to use to the point of filling Apple users with envy. It did not come with all bells and whistles right away, but there was a clear direction and a clear roadmap of what missing features would be implemented and when. After two years of listening to the feedback and ceaseless polishing, polishing, polishing, most of the missing features were finally implemented.

My fictional device's use was completely anonymous, you did not need to create an account to use it. It had updates for individual components delivered as and when ready, without bunching them together in big monolithic updates. It was, in a word, something different. You might even say unlike.

As I said, this is just a fiction. Luckily, I do not have to include the usual disclaimer since there is no device that my fiction could have resembled, even by accident.

As someone who at some point predicted the N900 community would die away as much as the previous NITs communities did, I've been surprised to have been proved completely wrong. I hereby admit that.

In fact for a shitton of time attention on this forum has been half split between N900 and Jolla. Not the 770, not the N8x0s, not N9/50, but the N900.

Really, it seems that the N900 is something to a lot of people. Maybe we should open another thread to discuss exactly _what_ it is, since I'm not sure I get it.

Doesn't seem much of a surprise to me -- it works as a phone, and as a portable computer. Really, no device even tries to do that today -- modern "smartphones" are phone + portable media-player + portable gaming device.

I really wish Jolla had tried to release something like that, but no, they're explicitly going for yet another iPhone, just like everybody else.

EDIT: Except, yeah, as javispedro points out below, Microsoft is giving it another try. I hate to say it (being as I have been ignoring Microsoft for so long), but I may have to take a look at their offerings again...

(That, and take another look at cobbling together my own RPi mobile device...)

While that actually resembles an "hybrid" laptop more than a dockable phone,
i am generally against anything that is "big-slot-in dockable" like those early Asus phones. Several reasons:

- Completely proprietary docking port,
- You can't carry the dock easily,
- When the phone is docked in, you cannot use it independently. Crap!

I think I like approaches like just plug in a "USB Type-C video cable" and you get full desktop video out better.


By the way, these days it is MS who is trying to revive the "phone as mobile computer" again:



Doesn't the above image ring a bell?



Yes, that's my N900 connected to a 20'' inch screen, in 2010, and in the second screenshot showing a native 1080p Hildon desktop totally independent of the on-screen one. We were so close....

Will. Never. Happen. "Fill Apple users with envy?" How in the heck are you possibly going to do that? The iPhone is the baby of Steve Jobs, and Steve Jobs was a true believer with an unlimited amount of cash and an army of brilliant engineers -- he's created the most user-friendly, feature-complete, bug-free smartphone the world has yet seen. So long as you use it in the way he meant you to use it, the thing is unassailable, at least in terms of being pleasant to use.

You, personally, may have some strange aversion to the device, but let me tell you, the iPhone users I know will hand over their devices when you pry them out of their cold, dead fingers.

I dropped by in the DNA-store, which used to be the main Jolla retail seller in Finland and there is one - yes one - Jolla phone on sale in whole Finland. How they think to make money if they don't sell the product?

Simple -- the phone is not Jolla's product. Sailfish is their product, and they're trying hard to sell it to various manufacturers. Jolla doesn't want to be in the business of selling phones...

Let me try to describe something in my mind for quite some time.

The mobile phone only has one purpose : to help its owner in doing some tasks. Its goal is not to provide access to some ecosystem.
Each person will have different tasks he would want the device to do.

Mainly, the kind of tasks done actually revolve around communication, media creation/sharing/playing, localisation and navigation, gaming.

Each time there is a need, the solution usually chosen is "create an app". I think this is where it starts to create problems in usability.

If I try to follow Jaakko's articles, we have to think how to answer the needs without taking the old paradigms as the only ones. In this case, the application is a desktop paradigm of the way the underlying system works (running programs). In any case does it mean that this low level constraint must be visible on the UI level.

There has been some works in several ecosystems to blend the application together (like adding a share on twitter/facebook link in the camera app, or the contact hubs).

This should be taken on a higher level, removing completely the applications from the user level, and increasing a lot the intergration between those services.

Usually, on the phone, the user follows a task flow, and doesn't work in a single app.

For example :
* receive SMS. Get notification
* Open the SMS app : A friend asks if you want to go see a movie
* Open the movie app (or browser), search for the current movies and time
* Go back to the SMS app : answer the friend

An other flow:
* Have an alarm reminder of the coming meeting with a friend
* Open the mail app, and search the mail were a friend gave you the restaurant website's URL for tonight
* Open the website, find the address
* Open the navigation app, to drive to this address

On most systems, you would have two apps open : SMS and movie/browser (or 3 for the second example : Mail/Browser/Navigation).

But... user are multitasking a lot, as anything can happen during this task flow. You were cooking while doing this, and have to go back to check the recipe ? Someone calls you ? In the movie app the director name reminds you of a friend you had to mail a file ?

Mix some simple tasks for some time, and now you have several apps opened. The broswer app is shared between two task flows (the movie, and the recipe), the mail app two (you were writting a mail, but some other mails were received in the mean time and you read them), ....

At the end, you have some task flows, and a multitasking view that doesn't represent it anymore.

What I think would be better, would be that app were hidden behind "views". Then the multitasking page would now list the task flows, and not anymore the apps used.
A task flow would be a chained list of "views" used to get to the intented result : the task being complete.

That way, you could add a new task flow, interrupting the current one, when something happens, and then be able to leave it there for some time, while switching from one flow to another.

I am not sure if this is well explained, and how to do this to be usable (that's were someone like Jaakko would help a lot!). Ask me if you want to rephrase some part, or add a drawing.

Did I not make it clear enough that it was a fiction?

That was IMHO the mistake. You do not make money on components, you make them on complete solutions. Even a phone is not a complete solution, but it is closer than just a bunch of screws. Or an OS. Does Google make money on Android?