Access vs. Media

Fri Apr 18 14:53:43 PDT 2003

Mark,

I've long been a fan of using handhelds in classrooms. The experiments
done by the Concord Consortium (if I'm remembering that name right)
and the recent ACM article by Elliot Soloway have convinced me that
this area has a lot of promise.

I agree that the current Palm and Visor's are not powerful
enough to run Squeak acceptably well. I've used Squeak on a 33 MHz
68K machine, and it's just too slow to do interesting multimedia work.
It doesn't even run Morphic well.

I've done quite a bit of work with more powerful PDA's, including several
of the Sharp Zaurus models, the Compaq Itsy, and the iPaq, and you can
do some interesting things with these machines. The key to reasonable
performance seems to be a RISC processor of over 100 MHz.

Yet even on the 206 MHz iPaq, there are some surprising performance
"gotcha's". The SA1100 chip they use has fairly small caches, and there
is no second-level cache as there is on a desktop or laptop. The result
is that, while tiny benchmarks run at at perhaps 1/3 the speed of
a 500 MHz G4 Powerbook, real programs run at 1/10 the speed of
the G4 or less. Furthermore, there is no floating point hardware, so
floating point code such as 3D graphics is likely to be MUCH slower than
that. There is no 3D graphics accelerator, either. Finally, if you need access
to a lot of media (sound and graphics), you'll need to use an external
storage card of some sort, and *all* of these cards are much slower
than the hard disk in your laptop. That goes for the IBM microdrive,
as well as for Compact Flash, MM, and SD cards. The need for external
storage would also increase the per-student cost of the device. Finally,
there is the issue of screen size. A 320x240 screen is just too small for
programming, in Squeak or any other language. Programming is hard
enough even when you have enough screen real estate...

Do all of these performance problems mean handhelds are not a
promising platform for educational applications? Not at all! The
very portability of handhelds makes it possible to take them places
you wouldn't want to take a laptop, such as out to a nearby stream
to measure water temperature gradients or onto a swing to measure
acceleration. And their small size makes them appealing to younger
kids--say third and fourth grade. (Although, in my experience,
ANY kind of computer is appealing to kids!) Still, to do a good job on
this platform would require a lot of careful thought about the UI
and packaging to make the best use of the screen real estate and
other limited resources. That very challenge makes handhelds more
interesting and exciting to researchers like you and Elliot (and me!),
because there are new discoveries to be made.

On the other hand, if you want to use all the educational Squeak
tools that you've already got, such as the Audio Explorer, you might
find it better to go with low-cost laptops. For example, Alan has been
giving all his Squeak demos, including some very impressive 3D demos,
using an Apple iBook. This machine is really only double the cost of
an iPaq with accessories, and it has an 1024x768 screen, graphic
accelerator, large and fast disk, and all the performance you'd
expect from a desktop computer. Plus it has Ethernet (for sharing)
and USB and Firewire for importing/exporting media from cameras
and video recorders.

One argument for handheld's is that they are cheap enough for every
student to have one. This was certainly true several years ago, when
the average laptop cost $3000 versus the $300 for a Palm Pilot. But in
the last several years, laptops have dropped in price by a factor of
two while high-end palm computers such as the iPaq cost at least
double the $300 for a Palm Pilot. So now the difference in price is
only a factor of 2.5 or less. (School systems can get iBooks for well
under the list price.) I think this trend is like to continue a bit longer,
because handhelds are still growing in functionality and are using
the latest (and most expensive) technologies, wherease laptops are
not changing in functionality and use more mature technologies that
are getting cheaper.

In conclusion, I'd say that handhelds ARE interesting, but NOT as
a cheaper, smaller Squeak machine. If you decide to explore educational
applications of handhelds, be prepared to re-build most of your
exisiting tools and UI's. Handhelds should be approached as a new
and different thing, not just as an incremental evolution of the laptop.

	-- John

P.S. I should add that Squeak is a fabulous vehicle for handheld
development. I do all serious programming on my laptop, then
slap the image file on a Compact Flash card and pop it into the
iPaq and things usually just work, modulo performance differences.

At 4:59 PM -0400 10/7/01, Mark Guzdial wrote:
>A few of the academic educational technology research groups in the 
>U.S. have been shifting their focus to Palms from desktop or even 
>WinCE devices.  Their argument is interesting: It's about access of 
>two kinds.
>- ACCESS BY STUDENTS: Many people believe that we're never going to 
>see much impact of computers at the level of 10 kids per computer (at 
>best!) that we have today.  Palms are cheap enough that one can 
>outfit a whole class with them using current budgets.
>- ACCESS TO STUDENTS: What's more, students react to them differently 
>than desktop computers.  My former Ph.D. advisor, Elliot Soloway, 
>says that kids in urban school districts don't even talk about the 
>Palms like computers -- instead, it's more like a Walkman or an MP3 
>player.  It's a media device that happens to have more interactivity 
>to it.
>
>I'm making an argument that a cheap, handheld device is a grand idea, 
>but the Palm is too cheap -- we simply don't want to give up all the 
>forms of media that we have even if we step up to an iPaq or WinCE 
>class device.
>
>I have some questions for y'all about this:
>TECHNICALLY:
>- I've been claiming that all forms of Squeak media (e.g., 3-D, 
>Alice, MPEG, Flash, text-to-speech) run on WinCE and iPaq devices, 
>but I realized that I haven't actually seen those all run on a 
>handheld device.  Can anyone verify that these Squeak media run on 
>the handheld devices?
>- Is it still the case that Squeak can't be ported to Palms?
>
>EDUCATIONALLY:
>- This group cares about these kinds of issues.  What do you think? 
>I'm arguing that more diverse media is worth an decrement in 
>accessibility that's found in the Palm.  Do you buy that?  Or is it 
>just a matter of time before the Palm's cost-performance ratio shifts 
>and we can get the media we want without surrendering the Palm's 
>cost, ubiquity, and accessibility?
>
>Thanks for advice and comments,
>Mark