The problem is that it's not enough to just have "low cost" chipset. What's more important is how we're going to buy it. Neo900 will be produced in very low quantities in comparison with phones from big companies, so prices of such "modern" pieces are going to skyrocket.

We can use those chips you mentioned. What we'll sacrifice by that is full compatibility with Fremantle and... price. If you're ok with paying above 1000 EUR for such smartphone and you can get ~500 more buyers to do that, then it can be produced - why not?

We have to keep in mind that this is community based project without any support from big gamers. We don't have Nokia's cash nor as many paid engineers as they have (or had :P). There's no Google behind us wanting to spread their platform. Price is going to be higher that in competition - that's for sure. But of course we can lower the difference with higher demand.

I still think that freedom GTA04, and now Neo900, can provide is worth higher price. It costs a bit, but it's an unique feature you won't get anywhere else.

a few annotations: This is about hw design. I don't expect every of you to have an EE background, so I will try to explain some facts to you.
Designing and building hardware isn't like the software development you maybe are used to. In software development when your app turns out to segfault under certain conditions, you apply a patch, rebuild, and worst case you deploy an update to all your customers. No matter if you notice that segfault during testing or 6 months after you shipped the product.
For hardware when you find out you created a tiny oopsie in your device, then odds are you maybe have produced and sold bricks worth a million dollar or more.

I came up with this Neo900 idea since there already is a tested and (mostly) working GTA04 which we won't mess with too much to reshape it into N900 housing. And still we need to do thorough lengthy QA tests to make sure we don't sell bricks. For sure we won't go and use a new CPU family or do any similar risky massive design changes.

Last not least the idea been to keep fremantle as OS since it's tested and known to finally work sufficiently bug free (recall early fremantle version like PR1.0, and the funny little bugs we had in it, anybody?). We neither gonna change that for new OS versions that maybe have look and feel of fremantle but a new untested "engine under the hood", nor will we change CPU family which as well would need such "new engine". In the end the major idea behind Neo900 is not about a giant leap ahead in performance but rather about sustainability of what we already have and love so much. If you want cutting edge then you need to go for cutting edge product, I suggest Samsung, Apple, or even maybe Jolla.

While I love the Maemo OS, I've no need for the hardware keyboard any more. What would it take to get a case with a slimmer profile with no keyboard?

1GB RAM is a definite requirement for me. I'd be willing to spend $650 US. Another item of interest for me would be that the hardware specs can run Ubuntu phone OS. With this, it becomes a developer device for that platform as well.

This idea appeals to me. I'm ex-Nokia, maemo/meego/... kernel team, and would be willing to spend some of my free time assisting with kernel and low-level-user-space issues. (If you need a kernel tree maintainer, that was much of my job for the last 2 years in Nokia, for example.) I can probably pull in a few volunteers from the ex kernel team, I'm still in touch with many of them.

To be honest, I'm not a huge fan of much of the n900 userspace, some of it's as buggy as crap, and is painfully bloaty, so I don't think people should be getting too attached to it. But yes, my n900 is my daily device, and has been for about 4 years. It's almost falling to pieces, but i've made sure I've got spares so that I never have to be without one.

But of course, this Fremantle part is about development direction associated with the project. Once you have the device in your hands, you can do whatever you want with it - even install Windows Phone if you happen to work at Microsoft and have access to their sources :P

We want you to be able to keep Fremantle as it is if you want to (in order to achieve that, we have to work on it, of course). But if you don't want Fremantle, then install SHR, QtMoko, Replicant, Mer... or port Ubuntu, Firefox OS, Sailfish, Harmattan... whatever you wish

Having more kernel developers on board would be great, esp those who know the OMAP3 stuff.

For sure fremantle is bloated in some of its parts, but it has the advantage to "just work". I don't know if you use CSSU, but the team has done a pretty descent job cleaning the bugs in PR1.3. So fremantle with CSSU is definitely not the PR1.3 fremantle. I know that, I am using N900 as my only device since Feb 2010.

FatPhil,
this sounds almost too good to be true :-D Of course you're more than welcome to help keep the kernel working. And we like to see you pulling in some of your ex-colleagues as well, no doubt about that. I think you might want to have a chat with Freemangordon and Pali, both active in kernel maintenance (buzzword kernel-power). Maybe join freenode IRC channel #maemo-ssu where both of them guys often hang out? I'm sure you will be given a heartly welcome there.

Again many thanks and I'm looking forward to meet you on IRC.
jOERG

Just work is why the openmoko while fun as a project gadget was useless to me as a phone, I never knew if it would actually ring when called.

The openmoko hardware dream matched with the maemo software effort makes an umbuntu grade geek gadget which we can still give to our mother, girlfriend, even Richard Stallman once configured.

i just love the n900. even though it have the charging problem. but still i am not selling it. i am waiting for your project to get another n900

Ok, what is Fremantle compatible then? Looking at the specs, this will be around droid 2 (CPU 1 GHz OMAP 3620, Droid 2 Global runs an OMAP3640 clocked at 1.2 GHz, Memory 8 GB flash ROM, 512 MB RAM), droid 1 was 3430 and sgx530 with 256 ram, a lot like n900. Since droids 2 can be bought for around 130$ (refurbished around 50$), maybe would make sense to find most compatible old device and build on that (form factor will probably make this impossible, but 100$ vs 700 euro...), with omap3 the only bump is ram/rootfs