this is really a big one (tread jack but sue me).

the SSU is a step up from the firmware "bombs" of earlier updates, but its still (imo) highly heavy handed.

and on top of that, nokia (again imo) is abusing the deb system with its single package system wide update "solution".

but then i guess if one want to be cynical one should have seen the writing on the wall back with the 770, as it didnt ship with a xterm, or for that matter a root account the user could access unless he had enabled r&d mode.

and i guess its the same pattern over at android or now webos. the linux part is just there to make rapid gains in hardware support and development tools, rather then starting from scratch yet again.

its not really there to embrace the open source ideals...

and yes, the problem is very much that we are in a limbo, a very big limbo.

thing is that while anything from the linux kernel to a main release of debian, open source works in a transparent, evolutionary kind of way.

if big changes are needed to be done in the kernel, people turn them into side projects that are allowed to mature alongside the existing system until it its feature parity and can be slotted in with no loss.

same with debian. things go in at the testing end, gets bumped to unstable, and finally reach stable.

but at any point, you as the user can accept to go with the new, rather then the old, and nobody will slap you silly for it.

but with the ssu you cant. its either all or nothing. you cant tell the app manager that you dont want the kernel updated as you have rolled your own. no you have to take the kernel or the ssu package wont go in, and therefor none of the other packages will go in either.

and its not really clear what the ssu will touch on, until its rolled out.

and lets take a pet monkey of mine, the wireless connection dialog.

during chinook you could put the thing into offline mode, hit refresh on the mail client and get a automatic connection to a saved wifi spot if you selected ok in a dialog popup.

with diablo that got broken. and it has stayed broken over some 3+ ssu (i have lost count). and now that nokia has focused on fremantle, and plan to go for a rewrite of that subsystem, the bug is either in limbo or "fixed in fremantle".

and as the diablo dialog is not open source, people are left with no chance of a fix, especially as they dont know if fremantle will be new hardware only, or not.

the only real saviour here is mer, but it seems hellbent on doing the maxium change rather the minimum change (ripping out and replacing nokia's closed bits).

all in all, both users, devs and anyone else is left like this:

to me, nokia needs to decide if they want to get into the open source boat, or stay on the closed source pier. right now they seem to hope they can evolve some very long legs by keeping one in each, with the boat sailing...

the question for me is this, how easy will it be to get existing software onto fremantle/maemo5?

that is, how big a change is there?

Well even if the SSU overwrites your particular altered kernel, I'd expect you'll want to port your alterations to the latest version anyway; Nokia presumably don't just randomly release kernels, they release them as there are improvements in there.

This then brings us onto the question of whether such patches would be accepted into the Maemo patchset so they are rolled out along with the SSUs.

Here's my concrete and specific:

(1) Reassure us (in concrete terms, not just with soothing words) that the death of the "Internet Tablet" label doesn't mean that the device will lose it's strength as the current best pocketable internet tablet. For example, knowing that the screen resolution, screen size and included browser are no worse than the current ones would be enough.

(2) Reassure us (in concrete terms, not just with soothing words) that the "touch UI without a notion of 'focus'" won't degrade the usability of the device with existing Linux software that needs a focus. After all, it's a Linux device and we need existing Linux software to run well.

(3) Tell us as something substantial about the form factor. Something like this would be enough: "There will be a device with a keyboard. There might also be a tablet, and might also be a device with a form factor not seen before. They will all be pocketable, although larger than a cellphone." And dammit, if there's not going to be a D-pad, just tell us will you, and let us get on with coping with it. Obviously all this stuff has already been decided. Can't you just let a junior staffer leak the concept video? It seemed to work well for the 5800.

That's the problem, not the solution. It's too vague, and expects intelligent developers to blindly trust a benevolent corporation. The real world doesn't work like that.

When you say "Should we stress more all this stuff that is known and is coming?" that's deceptive because you seem to mean "Should we stress the stuff that's officially announced?". If you started to stress all the stuff that Nokia knows and is coming, of course that would be something entirely different and quite wonderful.


Regards,
Roger

Rather than that, is there any chance of at least some of these patches being accepted upstream, in the mainline kernel?

getting offtopic, but not all kernel patches are compatible with every device used.
I think we would have to be very careful (as careful if not moreso than Nokia themselves) merging community updates into the SSU kernel automatically.

ie. i certainly wouldn't be happy if the high speed memory patch were automatically included and it stopped me from using my mmc.

If there are specific individual patches which can be confirmed stable on all (practical) devices and which don't cause problems and have support code around to be utilized then we should not be trying to pull it into community kernel and instead should make sure we make our case properly and unanimously for global inclusion

Linux kernels have had the ability to turn features on or off at boot time for quite some time.

It shouldn't be very hard to build an Advanced Configuration page where features can be toggled on and off, with warning notes where appropriate. I haven't used the High Speed patch exactly because it is such a PITA with Nokia's "Seamless Software Update" bundles...

Of course, an *open* platform would also feature multiboot-friendly kernel and boot system with failsafe option etc.

Maybe in Mer...

You mean like this one:

https://bugs.maemo.org/show_bug.cgi?id=969

A simple bug, a very simple and robust solution - bug filed two years ago, but applied to every kernel from OS2005 till now. Nokia's answer - fixed in Fremantle.

I nominate eiffel (Roger) as the Nokia "internet tablet" owner's community representative for "Cool Head of Reason." You continuously try and force all parties to remain honest and reasonable. Many thanks Roger. Keep up the excellent words.

Seconded. All in favor?