32nm?! Jesus Christ, I thought 45nm was going some! And I don't like the idea of an X86 tablet... bye bye always on convenience

All mobile devices have short life cycles if you haven't noticed

arguably the maemo community has benefited from a (mostly) common platform over three generations of hardware which has kept developers interested and consumers happy with new software.

now we have maemo 5 along with the Omap3 platform, promising a huge increase in performance and an OS that can take advantage of it......................... so we get to the release of the n900 sometime later this year.

how many developers are going to be really interested in long-term development for this platform if they know that by the end of 2010 nokia and intel will be planning the release of x86 based internet devices.

on this schedule there will be no follow on Omap3/4 based device running Harmattan. Harmattan will be running on x86 which makes it unlikely that n900/n910 owners will get the OS upgrade they might expect going on past practice.

maybe this explains why the n900 is going to be a smart-phone rather than an internet-tablet, because the target audience wants a device not a platform, i.e mobile phone users.

For most software developers I expect moving to intel will have no downsides. In fact most people will probably like it because if you decide to still compile on your desktop, cross compilation will no longer be required and testing will be easier now that qemu is no longer involved.

Of course with the n900 maybe a lot of people decide to put their root file system on mmc and compile on the device itself.

Who will notice are the kernel hackers and the people who program the DSP, both of which I assume are a very small portion of the community.

No, I'm pretty sure it doesn't mean that.

It would seem rash of Nokia to can the Maemo line early in the hopes that a future platform will be able to take up the slack, and the end of 2010 is a year and a half away, so I wouldn't consider it all that short of a lifecycle: a year+ of sales before the x86s launch (if they're on time), followed by some months of sales overlap and then some support after sales pretty much stop. Sounds like the N800, and only a bit shorter than the N810... While phasing out OMAPs eventually is certainly a possibility, I've seen nothing to indicate that it will happen, and it would be *****ic to stop them right away; more likely, the Intel processor tablets will start bigger, with bigger batteries, and still have short runtimes, while the OMAPs would stick around in some smart-phone-sized tablets and/or Linux smartphones, at least for a while.

Also, thanks to Intel's involvement, this is likely to pick up some Moblin taint flavor, not just be Maemo 6 from Nokia. Remember the tablets have run through 3 hardware versions and 4 major OS revisions, and the 4th hardware, 5th OS is finally starting to move from geek-toy to mainstream -- even with lessons learned and some major components reused, expecting a brand-new platform to match that on the first release is practically begging all the new mainstream customers brought in with Maemo 5 to go elsewhere when it's not as polished. Better to keep Maemo 5/6 devices around for a while until the kinks are worked out (courtesy of early adopters, AKA the pay-to-beta-test club!), then slide in the second round of Mobmo devices as a smooth replacement.

Trying to read a quick cancellation or "no Harmattan" from the info known seems to ascribe even greater depths of *****icity to Nokia than they've actually shown, which is saying something.

(Also, even if Nokia never touched x86, they'd probably be launching OMAP4 devices then -- on one hand, it's not like the OMAP2 -> OMAP3 switch is being great for OMAP3 tablets, as there's no Nokia backport of Maemo5, so presumably the OMAP3 -> OMAP4 switch would involve a similar break; on the other hand, there is a community port (Mer) of Maemo5 to Nokia OMAP3 tablets, and other ARM and x86 platforms, so even if Harmattan were to target x86, why in the world wouldn't it get ported to old devices?! To the extent that this shift is the same as one of the platform shifts we've already seen, it's no big deal, and to the extent it's different, it'll proceed slowly, because Nokia can't afford to make a big change abruptly.)

That's not really accurate. AMD got ahead during early 64-bit days, but Intel came back and beat them with superior chips. I am sure their marketing practices didn't help, but Core 2 Duos wiped the floor and painted the walls with the corresponding AMD chips at the time.

Considering that Intel's top of the line chips are consistently the best performing on the market, except for a few notable gaps such as early 64-bit and a period of first Athlon's dominance a while back, even if Intel don't try their best, they sure do try enough to genuinely be better than anyone else.

Don't get me wrong - I like AMD as much as the next guy. I am about to build a computer with an AMD chip because they're cheaper than Intel for the power I need (little) and the features I need (VT). But Intel has an absolutely immense amount of brainpower and a culture that keeps them bouncing back when behind and stops them from becoming too complacent when ahead.

cheers for the info guys.

OMAP4 is a much smaller transition than any of the previous ones. The accessories are the same as for OMAP3, as is the microarch. The only major change is more cores.

http://www.intomobile.com/2009/06/24...nd-compal.html

Huh ? One batch Atom powered, and the other ARM ?...

Wake me up at Christmas when/if it all makes sense :-)