[85] in bcs-newton
Re: handwriting recognition on Newton and other PDAs
daemon@ATHENA.MIT.EDU (Mitchell L. Silverman)
Mon Mar 29 18:35:48 1993
Date: Mon, 29 Mar 93 18:37:14 GMT
From: mitch@cfraix.cfr.usf.edu (Mitchell L. Silverman)
To: bcs-newton@world.std.com
Cc: <dsdavai@cfraix.cfr.usf.edu>, Lewis@cfraix.cfr.usf.edu,
On Monday, 29 Mar 1993 16:51:38 -0500 (EST), Sam Hunting
<shunting@world.std.com> wrote:
> On Sun, 28 Mar 1993, Randy Parker wrote:
>
> > All I indicated is that, based on four years of experience in the
> > pen-based computing industry including consulting to HWX engine
> > developers, blink-of-an-eye *CURSIVE* recognition was not a current
> > reality, regardless of microprocessor.
> >
> > Al Willis has since clarified his statement to cover only printed
> > characters, and I can happily say that for *PRINTED* characters that is
> > entirely true, especially on the ARM chip.
> >
> > There is enough misinformation going around about PDAs that we don't need
> > to add religious crusades to the debate. Let's try to stick to the
> > issues. The worst thing any of us who care about this industry can do is
> > to convince buyers that this thing is going to read "chicken scratch"
> > perfectly in 1993.
> >
> > -- Randy
I just joined the list, so forgive me for entering a debate
already in progress. Do I take the foregoing to mean that the
Newton probably *will* recognize *printed* handwriting pretty
well, and pretty quickly? How quickly do you all think it
recognize printed text?
If (as I suspect from the foregoing) the Newton will recognize
printed text quickly, good. I haven't used cursive for
anything save my signature (which I'd expect the Newton to see
as a graphic anyway, since *I* can't read it 8-) ) since the
special ed teachers in middle school gave up on trying to get
me to use it.
One question, though: I print in block caps almost entirely,
with some capital/lower-case distinctions indicated by size,
and others just omitted, detectable from context. Will the
Newton be able to figure that
> ...LATE, SO I'LL SEE YOU SATURDAY. ON THE OTHER HAND, SUSAN....
ought properly to be recognized as
> ...late, so I'll see you Saturday. On the other hand, Susan....
This sort of thing seems pretty simple, even from my limited
background in computing and linguistics--a decent-sized
spell-checking dictionary (e.g. Coach Pro, which can store its
dictionary in RAM) which I would think necessary anyway, with
some facility for indicating proper nouns (e.g. "Saturday")
that need to be capitalized, and exception-processing so that
unknown proper nouns (e.g. "Susan") may either be added (a PDA
might be expected to recognize and capitalize one's
associates' names automatically) or some "guessing" algorithm
developed from an analysis of lots of pre-existing English
text.
I realize that this may be off the topic. It also may well
still be an Apple trade secret, so guesses and discussion are
welcome. I want to buy a Newton, but as a poor college
student, I need to be sure that it will do what *I* need.
If I can take notes on it in a sloppy block-caps style, then
transfer those notes to my Mac and feed 'em to Mark
Zimmerman's FreeText stack for indexing or Eastgate Systems'
Storyspace as a hypertext web, I will be a very happy camper.
Also: I presume that Sam Hunting <shunting@world.std.com>
Randy Parker <rparker@world.std.com>, and Jim Rinaldo
<bcs_jim@MIT.EDU>, all of whom were addressees of the letter
to which I am responding, are on the bcs-newton mailing list?
If not, can someone let me know?
Thanks in advance for any information you can give me.
Sincerely,
Mitchell L. Silverman | As Voltaire never said, "I
Student, New College of | disagree with what you say,
the University of South Florida | but will defend to the death
Reply-to: mitch@cfraix.cfr.usf.edu | your right to say it."
