Post a Comment
Just out of curiosity how do developers of embedded linux products choose the kernel version? In one picture the version is 2.4.19-rmk7-pxa2 and the latest kernel version is 2.6.19. Do they stay with the version they originally started with or does the 2.4 series offer more for embedded than the 2.6?
A kernel is difficult to be ported to many architectures and professional testing is expensive, and so they go with the version that it's best tested on as many architectures as possible. Remember, the kernel comes from Montavista most of the time, and its on their best interest in distributing the version that has received most testing, because this matters a lot for embedded systems. Most other trolltech-based devices run on that same kernel as well.
Just out of curiosity how do developers of embedded linux products choose the kernel version? In one picture the version is 2.4.19-rmk7-pxa2 and the latest kernel version is 2.6.19. Do they stay with the version they originally started with or does the 2.4 series offer more for embedded than the 2.6?
The answer to your question is "yes".
We do it many ways, depending on the nature of the product. Mostly, we use OSS kernels, tracking them closely until some point where we need stability and then we tend to freeze on the release of the kernel until we get our product out the door.
After that point, it's a matter of tracking, timing, and feature set when we upgrade a product's kernel. Some products are never upgraded. Some are upgraded but only new copies of the product get the newer kernel. Some are upgraded in the field as well.
Some companies do use one of the third party support companies like MontaVista or Wind River, most don't, for various reasons. The companies that do are tied to the release cycle of their vendor, the companies that don't are tied to the whim of the open source community.
The question about a *really* hackable phone is: Will the carriers allow such a thing to connect?
And this one did not quite convince me:
> Today the smartphone operating system is a fragmented, balkanized
> mess -- just the kind of situation Linux and open source software can
> unify and build up.
The question about a *really* hackable phone is: Will the carriers allow such a thing to connect?
No, they won't. They won't cede control over their network to interlopers. And why should they? They own the network, they made the investment in building it, and they alone have the right to decide what kind of traffic flows across it.
"That's not the way that people buy phones."
Maybe where *you* live. Most of the world is a little different.
Where I live (the UK), the majority of people buy the phone at exactly the same time as obtaining their contract with the mobile operator.
The phones are HEAVILY subsidised here in the UK.
Yes you can get sim only options, but at the moment, I do not believe it is all that popular.
[message about buying a phone with the contract]
That is the only way to get a phone that will work with the vendors network.
Do you mean CDMA or GSM Networks?
For GSM Networks you should be able to take just any old GSM telephone (an unlocked one, that is one that is not programatically linked to a specific SIM Card) and put a GSM SIM Card in it, and it should work.
(I have from time to time switched SIMs with a friend, to make urgend calls when my cell was empty)
The SIM Cards store all information relevant to log into the providers network, so it should not be a problem.
I have read that CDMA Network have specific IDs stored in the Cellphone itself, so it has to specifically added to the providers network. Which would make it indeed hard to buy a different phone from a source other than the vendor.
So while it is typical to buy the GSM phone from your vendor, it is more related to having only to pay a very small and heavily subsidized price, than to having no other choice.
Yes you can get sim only options, but at the moment, I do not believe it is all that popular.
True, but people, at least here in Austria, often change the phone they use.
The first one is the one they got from their provider, but there is a huge market for second-hand phones on eBay.
The carriers have no choice: on a GSM network you can use any GSM phone. However, the GSM stack can't be hackable (which most developers don't care about).
Some carriers though, at least in NA, will restrict network access points to certain IMEIs (meaning a phone they sold you or at least a model they recognize), so you could theoretically wind up with a working phone but no access to network services.
Usually doesn't take more than a phone call to have the IMEI added to your account (I've had this happen with my own provider), but you sometimes have to work through until you find someone that will. Very often the canned response is that they won't support phones they haven't been tested and certified.
RE: Shout-out to hippies everywhere
"Today the smartphone operating system is a fragmented, balkanized mess -- just the kind of situation Linux and open source software can unify and build up."
Unless I am mistaken, there is truly only 2 OS for smartphones available and in widespread use. Introducing this one will make a 3rd and the beginning of serious fragmentation.
Not really true.It depends on the authors definition/interpretation of smartphone. Smartphone is a general term used for phone that have pda/pc like functions moreso that a traditional device. From this direction, one could include offerings with linux, symbian, and blackberry in the "smartphone" category... But I still think saying its fragmenting is a overstatement.
Unless I am mistaken, there is truly only 2 OS for smartphones available and in widespread use. Introducing this one will make a 3rd and the beginning of serious fragmentation.
depends on what you mean by 'widespread use.'
Off the top of my head I'm aware of Windows mobile, Symbian, PalmOS, and mPhone in wide use, although people outside of the Chinese market rarely see mPhone.
As far as Linux based smart phone OSes, mPhone is in use in China, and there are at least half a dozen others in developement.
"depends on what you mean by 'widespread use.' "
By widespread use I mean that I can buy a phone to use with a carrier.
"Off the top of my head I'm aware of Windows mobile, Symbian, PalmOS, and mPhone in wide use, although people outside of the Chinese market rarely see mPhone."
I can only go by the market I am in. In California the only smartphones that can be purchased and used with the likes of Sprint or Verizon are Windows Mobile or Palm based. I do know that is not the only market, just what I have seen out here is all. I actually welcome another alternative.
the carriers are waiting for the app that utilizes their network. if you never use the inet from your phone they can't sell you a 1gb data plan. as a palmos dev I would love to develop for qt rather then for the gtk apl bastard/wapoware from access/palmsource...
Browser: Palm680/RC1 Mozilla/4.0 (compatible; MSIE 6.0; Windows 98; PalmSource/Palm-D053; Blazer/4.5) 16;320x320
the carriers are waiting for the app that utilizes their network. if you never use the inet from your phone they can't sell you a 1gb data plan.
My provider in Canada charges $100/200MB, a benefit of being not only the largest GSM network in Canada, but the only one. I shudder, absolutely shudder, to think of what they would charge for 1GB.
We're so far behind much of the rest of the world when it comes to wireless tech that I could cry every time I read about these phones, and think about the potential.
On a semi-related note, any ideas about the OpenMoko effort? The (few) things I saw about it look promising, but I haven't dived into it yet.
I'd love to have a phone where I can fix the minor annoyances myself. For example: my current phone supports importing VCard-files, but only one at a time. Also, when the keyboard is locked, it is not easily possible to have the display light up and read the time. Totally trivial things to fix, but they'd definitely increase my happiness with the phone.



