[Arm-netbook] so where does ULP-COM fit into our thinking?

[Simon Kenyon]

On 10/10/12 15:18, Gordan Bobic wrote:
> On 10/10/2012 03:05 PM, green wrote:
>> Simon Kenyon wrote at 2012-10-10 02:43 -0500:
>>> do we really, really think that people will be swapping cards between
>>> enclosures on an hourly basis? i think it is worth looking at the relative
>>> costs of the card and the devices. it would seem likely that people would
>>> have one card per device and only change the card when there was a major
>>> upgrade in the card capabilities.
>>
>> My personal usage would include somehow switching from pocket-size tablet to
>> laptop multiple times per day, whether by moving a EOMA68 card or by simply
>> connecting the display and keyboard by cable.  The advantages as I see it are
>> only a single operating system to maintain and a practically identical
>> software experience regardless of chassis style.  Synchronization across many
>> devices, architectures, and operating systems suddenly becomes a non-issue.
>
> I can see this type of usage leading to a really poor experience as the
> screen resolution changes. All the desktop icons would get re-arranged,
> and different device formats require a fundamentally different user
> interface. Toshiba AC100 was a massive commercial failure purely because
> it is an Android device without a touchscreen, which is pretty much
> unusuable. Put normal Linux on it and use it as a laptop, and it's a
> fantastic machine.
>
> The point being that for a small touchscreen device like a phone or a
> slate you really want a user interface optimized for that experience,
> e.g. Android. For laptop/desktop usage with a keyboard, you want a more
> traditional user interface.
>
> While switching between the two is a nice feature to have for the
> future, at the moment, the OS' user experience provisions simply don't
> exist to make this use-case workable.
>
> Plus, the chances are that a chassis containing a touchscreen, a
> battery, and possibly a keyboard if it's a laptop, is likely to cost
> more than an EOMA module at mass-production prices, so the monetary
> saving on the hot-swapping isn't that great - it needs to be a usability
> and convenience driven advantage, rather than a cost drive one.

i think this "continual swapping" use-case is rather a stretch (to 
bowdlerise what i really think).

having a swappable core is an altogether different proposition.
my current computer has had three motherboards (at least) over its 
lifetime. i was able to swap them because they have the same form 
factor. this is a much more compelling argument for EOMA68.