[OT] "Pre-announcement" of Python-based "computing appliance" project.

[Richard Hanson]

While this post is primarily off-topic, I'm using Python to test
ideas, and hopefully, to develop a prototype for the project. I've
noticed that this group is frequented by some of the finer minds in
compsci; this post relates to HCIs and possible improvements to such.
But, be all that as it may, if this project can be shot down -- this
is where I'd prefer it to be done. :-) 

(I've been very impressed with the civil adult behavior here. Polite
*and* humorous -- my kind of people! In any event, I apologize if
anyone is offended by my posting of this, here.)

I had hoped to not announce until I had a minimal prototype ready.
However, my quickly disappearing time resources are being "devoured by
locusts"[1], some time-resources spent just trying to keep online
access, what with my current stable of failing Windows-based (alas)
computers and the current trend towards complexity rather than
simplicity in the world. 

(Python is the epitome of complexity-controlling languages; I think it
appropriate that this project which aims to markedly simplify the
complexity of the user interface [no lofty aspirations, here :-) ], be
done in Python.)

---

My project, "code-named" Pathfinder, is planned as a computing
appliance with "all batteries included" which answers 95%, say, of the
typical user's computing needs. It would be convenient for both the
"power user" and Grandma. Such arcana as harddrives, directories,
files, OS vs. apps distinctions, etc. would be hidden from user view.
(Some of these items wouldn't even need to exist under-the-covers.)
Something like Pathfinder could ultimately be, perhaps, the Model-T of
the not-yet-here Computer Age (with lots of help designing and
implementing such, of course).

(If you think I'm deluded -- you may well be right <wink> -- but read
on. HCI was my main area of interest back in my active [i.e., under
employment] programming days. Laziness, and now "mouser's arm," have
been, and still are, my main sources of inspiration. :-) )

---

One of the goals of Pathfinder is to have an integrated OS and system
on flash ROM, say, totally separated from user data (user data being
automatically and redundantly backed up transparently to the user)
with the system supplying the more commonly needed functionality in an
integrated manner. For example, Pathfinder would not allow apps to be
added by the user -- no need, as virtually all batteries are provided.
Authorized upgrades of the system ROM would still be allowed by the
spec. 

Another part of Pathfinder is a perhaps better HCI -- no need to steer
a mouse all around the screen just to arrive at some point which could
have been immediately gone to. And no need for sixteen ways to do
everything. Instead, a touchscreen-and-stylus, and an array of
dynamically labeled buttons in a panel on one side of the
browser-detail pane, would be the primary *one way* (other than the
keyboard) to interact with the system. (Detail views in the adjacent
browser-detail pane *may* need to have additional controls depending
on the object type: AV, text, PIM, etc.) 

The dynamically labeled buttons would either evoke some action as a
"leaf," or evoke a new "sub-menu" of newly labeled buttons (ala the
"softmenu" keys of the HP28 and HP48 series calculators). 

The browser-pane view, using only the dynamic button-panel and a few
screen-taps, could be "narrowed-down" from the all-objects, global
view to a very select, subset view through filter, sort, and attribute
selection. In this way, the desired "needle in an increasingly more
humongous haystack" could be found with only a few screen-taps of the
button-panel. Once an object (user or user-configuration data), or a
set of such, is in view, a few more button-panel taps could add (or
modify) the object(s) attributes as desired to "add value" for future,
further ease-of-use. 

Basically, the current windows-icons-mouse-pointer interface would be
entirely replaced by the "button panel" and touchscreen. (A mouse
could substitute for the touchscreen in the interim as the project is
being prototyped on conventional platforms.) No more menubars,
taskbars, Start Menus, endless series of dialog boxes, steering around
cascading menus, etc. With only a few dynamically labeled buttons,
say, properly nested in a sensible but easily changeable tree,
"zillions" of leaves of tasks, filters, orders, attributes, etc. could
be reached with only a few screen-taps. 

The project would be open-source, of course. (Not sure how any
eventual hardware could be open-sourced...?) I have more to say about
the specs and philosophy of the project if anyone is interested. (I've
had this project on the back burner throughout a seven-year-long
family reunion book undertaking; only now can I apply my available
resources without *that* distraction, however.) Contact me off-group
if you've a mind to.

Thanks very much if you have read all the way through this... uh...
something or other.


for-posterity-in-an-ever-increasingly-complex-world'ly y'rs,
Richard Hanson
_______________________________________________
[1] "Borrowed" from Jerry Pournelle's frequent usage.

Note: I could edit this forever, but I'm punching SEND, now -- ypoes
and such are thrown in for free. ;-)

-- 
sick<PERI0D>old<P0INT>fart<PIE-DEC0-SYMB0L>newsguy<MARK>com