[comp.sys.handhelds] 48SX ergonomics and stray comments

conte@crest.csg.uiuc.edu (Tom Conte) (03/17/90)

[When the 28C first came out I posted a long article describing it to
the net.  The reviews I've seen so far have been from people who HP
gave a demo unit to.  I was surprised that some of these reviews didn't
contain some interesting facts and hard criticism, and so I decided to
bang in some additional comments and perhaps start a discussion.]

The ergonomics of the 48SX are somewhat interesting.  Perhaps there
is a limit to the amount of functionality one can design into something
that's supposed to be handheld.  The 28 approached this limit with
the dual-keyboards.  That seems to have satisfied people who used the
calculator mainly on a desktop, but frustrated people who needed to
carry the machine into the field.  Although the case folded in half,
a great deal of flipping had to be done between the two halves to do
anything useful, as the left keyboard wasn't powerful enough.  (I'm
still surprised a third party didn't market a clip to lock the hinge
on the 28-- I tried this several times with a small binder clip and
found it worked amazingly well.)

The 48SX uses a technique that harkens back to the 67-- quadrupally
defined keys.  This gives the keyboard an overall crowded look, which
caused a friend's whife to comment, ``my *God*, what a geeky calculator!''
To add to this, HP has decided to return to moded operation like what
the 41 evolved into in the 41CX.  There are several modes where certain
keys on the calculator perform unlabled functions (e.g., + displays the
cursor coordinates, - hides the menu, etc.)  I was sad to see this.
Language designers call such a design non-orthogonal, and the mode-lessness
of the 28 was a welcome change from the 41CX.  In my first few days
of playing with the calculator there were several situations where I
entered a `mode' and hadn't realized it.  I noticed after a time that
I adopted the `when-in-doubt-hit-ON' habit.  After a few more sessions,
this went away some, but I'm convinced that if I were to stay away from
the machine for a little while I would have to relearn a lot again.
Some of the elegance and simplicty of the 28 has been compromised for
a higher degree of sophistication.  I suppose that's the way of things.

The physical size surprised me.  It is clearly the largest HP I've seen
since the 19C.  I had someone comment on just how big and heavy it is.
(At least it is still smaller than the MCII, a rediculously large
machine, IMHO.) Luckily, HP has made some adjustments for its size.
The back edges are nicely tapered, there are `thumbgrips' on the sides,
and the balance of the instrument seems to compensate for its weight.
(Holding the calculator with the thumb on a thumbgrip seems to work
best it seems, because the bottom half is then accessable to your
other hand) There were some comments that the machine is flimsy, but
I don't agree.  It seems solid, if not a definate improvement over
the 28-style case. The keys seem to be a hybrid between the current
line and the older style slanted HP keys.  They feel of higher quality
than the 42S, and I wonder if some redesign might have happened (or
perhaps this is subjective, considering the cost difference!).
Unfortunately, the legends on the keys themselves are particularly
unattractive. (The font for the legends seems like a bolded sans serif.)
Also disconcerting is that the menu keys don't line up with the menu
line of the display (they are wider-spaced).  I imagine this was
to allow the first row of alphas to be printed beside the keys,
but it feels a little awkward at first.  Also, It's nice that HP
supplies a case for the machine, whereas this was lacking on the
28, since that machine was its own case (but HP sold a case for it
seperately!).  The case looks much like a case I've seen for a Zenith
laptop.  Nice touch.

The display has a wider viewing angle than the 28, but it seems to be
less focused, perhaps due to the stark shadows.  I found the left-shift
and right-shift annuncators impossible to make out, for example.  I
think this was the source of an earlier comment that the constrast was
poorer on this machine.  The blue color is easier on my eyes, however.
A tradeoff, I suppose.

The nitty-gritty operations had some quirks, for example the shift from
an alpha-lock on the 28 to a one-time alpha key threw me off for a while,
but HP provides a system flag change this.  All commands are not accessable
from the menus, also.  For example, STR-> must be entered by hand.  (This
is perhaps because OBJ->, a more general command, is provided.)  Kudos
to HP for making this machine compatible with the 28, a major plus.  I'm
surprised that the DEL key is where it is, I use it very infrequently,
and my 28 habits keep me hitting it as though it were a DROP.  Luckily,
HP decided to again supply a USER mode.  Some day maybe we'll see a
redefinable keyboard with LCD legends that change above the keys as
the modes change.  If anyone ever does this, most of the problems I
mentioned above would disappear.  The engineering of such a machine
is probably a little tricky... ;-)

The positives are numerous, including an outstanding job of supporting
I/O (clap, clap, clap).  I told a friend that Kermit was built into the
thing and he laughed in amazement.  Too bad the connector on the top
isn't standard and we will have to play tricks to improvise something.
(Why don't I just buy the cable?  If the sales literature is correct,
the cable is a 48SX-to-9pin, and that isn't correct anyway for the Amiga
on my desk.)  The EquationWriter makes the symbolic mathematics and
RULES (formerly, FORM on the 28) useful finally.  Unfortunately, the
speed of EquationWriter seemed a little slow for effective interactive
manipulation.  The GRAPH mode is nifty, and I've joked a little more
powerful than X11's `bitmap' program.  PLOTing is vastly improved,
through the inclusion of graph types, and (thank God) autoscaling.
Being able to download a graphics object to a PC is another indication
that the PC/calculator interface was of high priority in the design of
this machine.

I'm very happy with this calculator, and my comments above are because
I know in the past HP has listened to the customers with each calculator
it produces and used this information for the design of the next instrument.
As Bill Wickes pointed out, a large number of the suggestions and gripes
from this very newsgroup have been addressed in the design of the 48SX.
I don't expect HP to turn around saying, `ohmygosh, it isn't perfect,
lets redesign it tomorrow and get it outoutout!'  But I betcha the handheld
they release in 2001 will address the imperfections in the 48SX.

Lastly, a standing ovation to HP/Corvallis: you've fed my calculator
habit happliy for quite some time now with ever-more impressive designs
that all have the same characteristic HP quality to them.  Hope the
48SX sells like hotcakes for you all.

------
Tom Conte      Center for reliable and High-Performance Computing
               University of Illinois, Urbana-Champaign, Illinois

e-mail:	   conte@csg.uiuc.edu   or   t-conte@uiuc.edu