Editing MessagePad (section)

====Handwriting recognition====
{{Main|Newton OS}}
In initial versions (Newton OS 1.x) the handwriting recognition gave extremely mixed results for users and was sometimes inaccurate. The original handwriting recognition engine was called Calligrapher, and was licensed from a Russian company called [[:ru:Параграф (компания)|Paragraph International]]. Calligrapher's design was quite sophisticated; it attempted to learn the user's natural handwriting, using a database of known words to make guesses as to what the user was writing, and could interpret writing anywhere on the screen, whether hand-printed, in [[cursive]], or a mix of the two. By contrast, [[Palm Pilot|Palm Pilot's]] Graffiti had a less sophisticated design than Calligrapher, but was sometimes found to be more accurate and precise due to its reliance on a fixed, predefined stroke alphabet. The stroke alphabet used letter shapes which resembled standard handwriting, but which were modified to be both simple and very easy to differentiate.<ref>{{cite web |last=Miller |first=Jack |date=September 19, 1994 |title=Graffiti Arrives! -- Press Release enclosed: Palm Computing Delivers Major Advance in Handwriting RecognitionTechnology. |url=http://www.grot.com/zoomer/zoomer-list/zoomer-list-3Q1994/0048.html |url-status=live |archive-url=https://web.archive.org/web/20160627002502/http://www.grot.com/zoomer/zoomer-list/zoomer-list-3Q1994/0048.html |archive-date=June 27, 2016 |access-date=April 10, 2008 |website=3Q1994 Zoomer-List Archives, Zoomer mailing list archives, Zoomer PDA WWW Information Service |publisher=Palm Computing, Inc. |place=LOS ALTOS, Calif.}}</ref> [[Palm, Inc.|Palm Computing]] also released two versions of Graffiti for Newton devices.  The Newton version sometimes performed better and could also show strokes as they were being written as input was done on the display itself, rather than on a silkscreen area.

For editing text, Newton had a very intuitive system for handwritten editing, such as scratching out words to be deleted, circling text to be selected, or using written [[caret]]s to mark inserts.<ref>{{Cite web |date=1995 |others=PDF file created by Kady Ingham on July 16, 2001 using Acrobat Distiller 5.00 for Macintosh |title=Newton: Apple MessagePad Handbook |url=https://download.info.apple.com/Apple_Support_Area/Manuals/newton/0307258ANEWTONMP.PDF |website= |publisher=Apple Computer, Inc. |publication-place=Cupertino, CA |id=Part No. 030-7258 (listed in https://web.archive.org/web/20080406215949/http://docs.info.apple.com/article.html?artnum=50092)}}</ref>

Later releases of the [[Newton OS|Newton operating system]] retained the original recognizer for compatibility, but added a hand-printed-text-only (not [[cursive]]) recognizer, called "[[Rosetta (Newton)|Rosetta]]", which was developed by Apple, included in version 2.0 of the [[Newton OS|Newton operating system]], and refined in Newton 2.1. Rosetta is generally considered a significant improvement and many reviewers, testers, and most users consider the Newton 2.1 handwriting recognition software better than any of the alternatives even 10 years after it was introduced.<ref>HWR accuracy:

* See comments in Wired's Apple Newton Just Won't Drop: {{Cite magazine |last=Kahney |first=Leander |date=August 29, 2002 |title=Apple's Newton Just Won't Drop |magazine=WIRED |url=https://www.wired.com/2002/08/apples-newton-just-wont-drop-2/ |url-status=live |archive-url=https://web.archive.org/web/20070305034333/http://www.wired.com/news/mac/0,2125,54580,00.html |archive-date=March 5, 2007}}
* See text under "Handwriting Recognition" in [http://www.pencomputing.com/developer/newton_20.html Pen Computing's First Look at Newton OS 2.0]: {{Cite magazine |last=Huang |first=Vernon |date=December 1995 |title=First Look: Newton OS 2.0 |url=http://www.pencomputing.com/developer/newton_20.html |url-status=live |magazine=Pen Computing Magazine |issue=7 |archive-url=https://web.archive.org/web/20220330084907/http://www.pencomputing.com/developer/newton_20.html |archive-date=March 30, 2022}}
* See "Opportunity Squandered" in Pen Computing's Why did Apple kill the Newton?: {{Cite magazine |last=MacNeill |first=David |date=June 1998 |title=Why did Apple kill the Newton? |url=http://www.pencomputing.com/frames/newton_obituary.html |url-status=live |magazine=Pen Computing Magazine |archive-url=https://web.archive.org/web/20061212155656/http://pencomputing.com/frames/newton_obituary.html |archive-date=December 12, 2006}}
* See comments under "Software" in MacTech's MessagePad 2000 review: {{Cite magazine |last=Ringel |first=Edward |date=1997 |title=Newton? Get Serious! |url=http://preserve.mactech.com/articles/mactech/Vol.13/13.04/NewtonGetSerious/index.html |url-status=live |department=Tools Of The Trade |journal=MacTech Magazine |publisher=Xplain Corporation |volume=13 |issue=4 |archive-url=https://web.archive.org/web/20061020113709/http://www.mactech.com/articles/mactech/Vol.13/13.04/NewtonGetSerious/index.html |archive-date=October 20, 2006}}
* Comments by Pen Computing's editor: {{Cite magazine |last=Blickenstorfer |first=Conrad H. |date=June 2000 |title=From the editor: Commentary by Pen Computing Magazine's editor-in-chief |url=http://www.pencomputing.com/editor/editor_34.html |url-status=live |magazine=Pen Computing Magazine |issue=34 |archive-url=https://web.archive.org/web/20061016131338/http://pencomputing.com/editor/editor_34.html |archive-date=October 16, 2006}}
* See user testing results discussed in part 6 of this A.I. Magazine article on Newton HWR: {{Cite journal |last1=Yaeger |first1=Larry S. |last2=Webb |first2=Brandyn J. |last3=Lyon |first3=Richard F. |date=1998-03-15 |title=Combining Neural Networks and Context-Driven Search for Online, Printed Handwriting Recognition in the NEWTON |url=https://ojs.aaai.org/aimagazine/index.php/aimagazine/article/view/1355 |journal=AI Magazine |language=en |publisher=American Association for Artificial Intelligence |volume=19 |issue=1 |page=86 |doi=10.1609/aimag.v19i1.1355 |issn=2371-9621 |id=0738-4602-1998}}
* MessagePad 2000 review at Small Dog Electronics: {{Cite web |last=Mayer |first=Don |date=July 1, 1997 |title=Kibbles &Bytes; #29: Don's Review of the Newton Message Pad 2000 |url=http://www.smalldog.com/newsarchive/kibbles_display.php?id=29 |url-status=dead |archive-url=https://web.archive.org/web/20080416113926/http://www.smalldog.com/newsarchive/kibbles_display.php?id=29 |archive-date=April 16, 2008 |access-date= |website=Kibbles&Bytes; weekly E-Mailing |publisher=Small Dog Electronics |publication-place=Warren, Vermont, USA}}
* See comments under "Note-taking" in MessagePad 2000 review at "The History and Macintosh Society": {{Cite web |last=Klingsporn |first=Geoffrey |date=May 1997 |title=The Postgraduate Newton: a month in academia with Apple's new handheld computer |url=http://www.h-net.msu.edu/~mac/mp2000.html |url-status=dead |archive-url=https://web.archive.org/web/20060920125747/http://www.h-net.msu.edu/~mac/mp2000.html |archive-date=September 20, 2006 |access-date= |website=H-Mac Discussion Network, H-Net: Humanities and Social Sciences Online |publisher=The History and Macintosh Society (H-MAC), Michigan State University}}
* What's Right With The Newton: HWR: {{Cite web |last=Wittmann |first=Michael C. |title=What's Right With The Newton |url=http://perlnet.umephy.maine.edu/bio/wittmann/pooter/newtuse/hwr.html |url-status=dead |archive-url=https://archive.today/20060908164609/http://perlnet.umephy.maine.edu/bio/wittmann/pooter/newtuse/hwr.html |archive-date=September 8, 2006 |website=Michael C. Wittmann: Personal pages: Random Interests I: Computing |publisher=Physics Education Research Laboratory, Department of Physics and Astronomy, University of Maine}}
* Assessment by Apple developer: {{Cite web |last=Yaeger |first=Larry |date=February 13, 2006 |title=Apple-Newton Handwriting Recognition |url=http://shinyverse.org/larryy/ANHR.html |url-status=live |archive-url=https://web.archive.org/web/20120717074702/http://www.beanblossom.in.us/larryy/ANHR.html |archive-date=July 17, 2012 |website=ShinyVerse |publisher=Indiana University}}
* [http://www-writing.berkeley.edu/TESL-EJ/ej07/mr1.html Prof. Warden's MessagePad 130 review]: {{Cite journal |last=Warden |first=Clyde A. |date=January 1997 |title=MessagePad 130 |url=https://tesl-ej.org/wordpress/issues/volume2/ej07/ej07m1/ |url-status=live |department=Hardware Review |journal=The Electronic Journal for English as a Second Language (TESL-EJ) |publisher=University of California, Berkeley |publication-place=Berkeley, CA |volume=2 |issue=3 |eissn=1072-4303 |archive-url=https://web.archive.org/web/20061231035454/http://www-writing.berkeley.edu/tesl-ej/ej07/mr1.html |archive-date=December 31, 2006}}</ref>  Recognition and computation of handwritten horizontal and vertical formulas such as "1 + 2 =" was also under development but never released.<ref>{{Cite web |last=Sharp |first=Maurice |date=November 6, 2006 |title=[NTLK] Random historical thoughts... |url=https://www.newtontalk.net/archive/newtontalk/2006-November/0127.html |website=The NewtonTalk Mailing List |publisher=NewtonTalk & Friends |postscript=.  Also archived at https://marc.info/?l=newtontalk&m=116284333623124.}}</ref> However, users wrote similar programs which could evaluate mathematical formulas using the Newton OS Intelligent Assistant, a unique part of every Newton device.

The handwriting recognition and parts of the user interface for the Newton are best understood in the context of the broad history of [[pen computing]], which is quite extensive.<ref>{{cite web |last=Ward |first=Jean Renard |date=2020 |orig-date=1992 |title=Notes on the (relatively unknown) History of Pen-based Computing |url=http://users.erols.com/rwservices/pens/penhist.html |url-status=live |archive-url=https://web.archive.org/web/20170418162718/http://users.erols.com/rwservices/pens/penhist.html |archive-date=April 18, 2017 |access-date=March 21, 2010 |website=Annotated Bibliography in Touch/Pen Computing and Handwriting Recognition by Jean Renard Ward |quote=This is a short outline of a presentation on the history of "pen computing" I gave for the Pen Computing group of the Boston Computer Society, March 11, 1992. Unfortunately, I did not keep the full set of slides, which had pictures of the products and systems. However, Dan Bricklin (CTO at Slate Corporation at the time) has posted a video of the same presentation on youtube, which does show all the pictures and slides: http://www.youtube.com/watch?v=4xnqKdWMa_8.}}</ref>

A vital feature of the Newton handwriting recognition system is the [[modeless]] [[Error detection and correction|error correction]]. That is, correction done ''[[in situ]]'' without using a separate window or widget, using a minimum of gestures. If a word is recognized improperly, the user could double-tap the word and a list of alternatives would pop up in a menu under the stylus. Most of the time, the correct word will be in the list. If not, a button at the bottom of the list allows the user to edit individual characters in that word. Other pen gestures could do such things as transpose letters (also ''[[in situ]]''). The correction popup also allowed the user to revert to the original, un-recognized letter shapes - this would be useful in note-taking scenarios if there was insufficient time to make corrections immediately. To conserve memory and storage space, alternative recognition hypotheses would not be saved indefinitely. If the user returned to a note a week later, for example, they would only see the best match. Error correction in many current handwriting systems provides such functionality but adds more steps to the process, greatly increasing the interruption to a user's workflow that a given correction requires.{{Citation needed|date=February 2008}}