> On Thu, Jan 7, 2016 at 9:34 AM, dmccunney <dennis.mccun...@gmail.com> wrote:
>> The variant with Ubuntu pre-loaded looks interesting, and would save a
>> fair bit of work.
>> I can imagine the fun getting Fedora up on the Windows variant.  I can
>> also imagine "sloooooow to boot".
>> I doubt the Ubuntu variant will be a speed demon, either, but for
>> things like that, my practice is to boot it and simply leave it up.
>> If you don't need to shutdown and reboot, why do so?
> To clarify: The ComputeStick was very fast to boot from the internal
> storage. It was just booting from the LiveUSB that took forever. I'm
> sure if I'd managed to install Fedora Linux on the ComputeStick,
> Fedora would likely have booted very quickly too.

Ah.  I can see where that might be the case.  I'd guess the LiveUSB
was USB 2, and you would probably want USB 3 to boot from it quickly.

(My desktop is USB 2.  That's not a drawback in practice because what
plugs into USB is mostly things like flash drives used for archival
storage where USB 3 speeds aren't required.)

> We regularly shut down the ComputeStick at the end of the day when we
> were done experimenting. You can leave it up, but the network was
> wireless network only, and our wireless network would "time out"
> connections about every 3 hours. So there wasn't any value in leaving
> it running overnight anyway.

Since it booted quickly from internal storage, my "Just boot it and
leave it up" comment is less relevant.  I do that to avoid rebooting,
but if you have fast enough booting, you may not care.

>> What was the lab's experience with trying to use the Windows model?
> We found it was a nice device that would meet most workloads. We were
> a Google Apps for Education campus, so most of the workload was via a
> web browser, although (IIRC) the Windows ComputeStick came with a
> local copy of Office. You really only noticed a "dip" in performance
> when opening a big spreadsheet (I had a lot of those, as part of our
> project management process). If you were just opening an empty Google
> spreadsheet or an empty Excel spreadsheet or a Word document or a
> Google Docs document, it was fine.

My experience has been that opening really big spreadsheets is always
a resource intensive process.

At a former employer, the facility manager at the office I worked out
of complained his PC was slow.  His machine was basically identical in
configuration with most of the others, and no one else complained.  It
turned out what was really slow was opening a big spreadsheet that
contained the current stats on work in process.  The spreadsheet
resided on an NT server in my computer room.  Another employee would
update it nightly, and email it to him.  He'd open it in the morning.

We used Outlook talking to Exchange Server, and the Exchange Server
was in another office at a remote location.  So when he opened the
spreadsheet, he was opening it across the WAN from the Exchange Server
mail store.  Gee.  No wonder things were slow.  The solution was for
the employees doing the updates to include a pointer to the local copy
in her email instead of actually attaching the file.  The manager's
speed complaints went away.  :-)

> We recommended that the ComputeStick would be a better fit for
> classroom computers and conference rooms, where the workload was more
> predictable (i.e. mostly Google Slides or Powerpoint). The
> ComputeStick could be directly attached to the projector or display
> (also makes the ComputeStick more difficult to steal, as it's either
> locked away or really hard to get at) and use a BT keyboard/mouse. We
> also recommended a USB adapter to provide a wired network connection,
> rather than use wireless.

Compute Sticks might get up and walk off if not closely watched?  :-p

But yes, that experience sounds like what I'd expect, and especially
USB for network connectivity.  If you don't need x86 compatibility,
you can do similar things with ARM based devices.  We are firmly in
the age of "Getcher computer!  Onna stick!  'Ardly moving at all!"
CMOT. Dibbler would be in his element...

Freedos-user mailing list

Reply via email to