On Sun, 15 Jan 2006, Wouter van Ooijen wrote:

>> I am not an analyst but my opinion is that anything that does not 
>> scale easily to 16-32x its current size, is not second courced and 
>> widely available, will be a niche within a year or a few.
>
> So that would be everything?

Basically everyting you cannot use, does not exist, no ? E.g. a nice 
chip that requires a minimum 10,000 order for the makers to notice you, 
has an 'inexpensive' development environment that is about 
$5000/seat/year and requires a stack of NDAs to be signed for access to 
their IP does not look like it's going to be very popular, does it ?

> AFAIK no ARM chip is realy second-sourced in the sense that you can buy
> the same chip from a second manufacturer.

No but they are all very similar and porting code is very easy (using 
the C language as common platform).

> One of my favourite books from 30y ago was a book about 
> microcontrollers from Adam Osbourne (the one that later made the 
> Osbourne I). It contained a prediction about the future of 
> microntrollers, which boiled down to: There will be a split between 
> the higher part of the marked, where nice-architecture chips will 
> dominate, and a lower part where even the weirdest architecture can 
> survive if the price is right. I think this still holds, and there are 
> a few more important things like inertia and availability track 
> record.
>
>> With the new ethernet chips, the older architectures will have some 
>> air to breathe, but not for long. The second question a client would 
>> ask after having squeezed an interactive web server into a PIC would 
>> likely be 'does it do DHCP and can we manage it via SNMP' ? Or 'can I 
>> connect it to a DSL modem directly' ? ARM and MIPS can do that sort 
>> of thing now.
>
> That will indeed be the questions for the higher segment of the market.
> For the lower segment the questions will be 'what does it cost at
> 100k/y' and 'can I be sure that I can buy that chip 10y from now'.

Well if you are right then it's a good thing that there is an open 
source C compiler for the MCS51 core because it's going to be around for 
a while .... btw I write a lot on this mailing list but my last 4 
projects at least were MCS51 core powered ... and the MCS51 already has 
a few TCP/IP stacks available ...

Peter
-- 
http://www.piclist.com PIC/SX FAQ & list archive
View/change your membership options at
http://mailman.mit.edu/mailman/listinfo/piclist