How to customize minix for mobile phone?
326 views
Skip to first unread message
Ajay Prabandham
Dec 17, 2010, 1:04:35 PM12/17/10
to min...@googlegroups.com
Dear Guys,
I am posting this not to give you a readymade recipe or secret formula for customizing minix for us,
in a preferably low-cost cell phone running only open-source and free software. The biggest problem is with the Hardware industry. I mailed Dr Tanenbaum the same question, and we was pof the view that HW manufacturers are generally unwilling to share specs and other info (Political, not technical, decisions)
A few years ago, a few Masters and PhD students with support from Faculty at IISc Bangalore, India designed A low cost notebook called "Simputer". One could make an effort to link minix and simputer (i know it seems far fetched, biut stranger things have happened on this planet, haven't they?)
The idea in my mind is to ask Simputer to fabricate a tablet=sized computer (or if they already have one, give us the specs). Or we could start of
f with what they and what we have now. my questions, which i feel need discussion, are:
1. Would Minix3 do well on n embedded platform? Can it be customized to have a really small kernel footprint?
2. There are many desirable qualities of a potential mobile OS. Would Minix's USP as a fault-resilient and fault-tolerant microkernel architecture add considerable value to a cell phone?
3. For the practical spects, what changes need tpo be made to the various OS subsystems? memory management,
filesystems,
interrupt handling,
device framework,
cross-platform development support,
IPC mechanisms,
scheduling policies,
concurrency support (kernel, LWP and user threads)
Security
process synchronization
Why don't we start thread where interested group members could start active discussion...especially for issues in point 3?
Regards,
Ajay
part...@gmail.com
I am posting this not to give you a readymade recipe or secret formula for customizing minix for us,
in a preferably low-cost cell phone running only open-source and free software. The biggest problem is with the Hardware industry. I mailed Dr Tanenbaum the same question, and we was pof the view that HW manufacturers are generally unwilling to share specs and other info (Political, not technical, decisions)
A few years ago, a few Masters and PhD students with support from Faculty at IISc Bangalore, India designed A low cost notebook called "Simputer". One could make an effort to link minix and simputer (i know it seems far fetched, biut stranger things have happened on this planet, haven't they?)
The idea in my mind is to ask Simputer to fabricate a tablet=sized computer (or if they already have one, give us the specs). Or we could start of
f with what they and what we have now. my questions, which i feel need discussion, are:
1. Would Minix3 do well on n embedded platform? Can it be customized to have a really small kernel footprint?
2. There are many desirable qualities of a potential mobile OS. Would Minix's USP as a fault-resilient and fault-tolerant microkernel architecture add considerable value to a cell phone?
3. For the practical spects, what changes need tpo be made to the various OS subsystems? memory management,
filesystems,
interrupt handling,
device framework,
cross-platform development support,
IPC mechanisms,
scheduling policies,
concurrency support (kernel, LWP and user threads)
Security
process synchronization
Why don't we start thread where interested group members could start active discussion...especially for issues in point 3?
Regards,
Ajay
part...@gmail.com
Jon Shadow
Dec 17, 2010, 3:22:36 PM12/17/10
to min...@googlegroups.com
We'd better leave mobile phones to meego or android.
Konstantin Zertsekel
Dec 17, 2010, 4:35:55 PM12/17/10
to min...@googlegroups.com
> We'd better leave mobile phones to meego or android.
Wrong!
In the long run micro-kernels will definitely win the world.
The real obstacles are:
(*) practical usage excellence of micro-kernels (like that of QNX)
(*) business and marketing leverage
Without money making incentive, it is plain impossible to rule the world with open source projects.
As to the smartphone arena, can anybody give a short summary of Operating Systems gatekeeping the smartphone market?
Kosta.
In the long run micro-kernels will definitely win the world.
The real obstacles are:
(*) practical usage excellence of micro-kernels (like that of QNX)
(*) business and marketing leverage
Without money making incentive, it is plain impossible to rule the world with open source projects.
As to the smartphone arena, can anybody give a short summary of Operating Systems gatekeeping the smartphone market?
Kosta.
Reply all
Reply to author
Forward
0 new messages