ACSII Gregg Shorthand (AGS)

The following is a great proposal by Andrew for AGS.  Are there any problems with this system?  Is it complete?  I can see how a standardized version of ‘ACSII Gregg Shorthand’ could be a very beneficial thing therefore I am calling on all Gregg gurus to comment on this document.
Thanks,
    Jason
——–
AGS – ASCII Gregg Shorthand
     In order to make a standard ASCII form for Gregg Shorthand, I had to represent each character with a single stroke.  Therefore, th could not be used to represent th, as it would be t and an h dot.  Instead, I use q, an otherwise unused stroke.  To express -ing, an h is placed at the end of the word.

k k
g g
r r
l l
n n
m m
t t
d d
q over th
p p
b b
f f
v v
c ch
j j
s s
w sh
S left s
Q under th
h h
N ng
M nk
a a
e e
o o
u u
eu eu
ou ou
oe oe
ai ai
ia ia
ea ea
A reversed a
E reversed e
< reversed a loop
> reversed u loop
{ a loop
} e loop
* diacritic dot
^ diacritic dash
_ diacritic underline
= capital
& nd
D dn
F dm
G md
J jnt
D dv
Y td
W mn
x x
X xes
ses
O o on its right side
U u on its right side
# raised over th
P us single stroke
+ short space (to express “of the” in Pre-Ann.)
/ intersection
| word-ending (makes the next outline appear underneath the preceding outline)
@ word-beginning (makes the preceding outline appear above the next outline)

This makes a transcription that looks like this:
    17. dS q mOr uev OnU hP & lot lb du Onq
25# oqsWq. q hP nedS pa&h $q D&S sa q rUf leks
evF t ranS. efuuw P tu renU q mOr f Qe }rS uel
duso eful pa& q hP & put On h nU rUf.  qsl nw tS
bh leSt fq tem oq mOr. efurn lh Yu qs uewv tusk
u tu ara tutakka oq mOr en t bks du. ur 
 
The English form is thus:  
     17.  Dear Sir:  The mortgage we have on your house and lot will be due on the
25th of this month.  The house needs painting and the tenants say the roof leaks
every time it rains.  If you wish us to renew the mortgage for three years we will
do so if you will paint the house and put on a new roof.  This will insure its
being leased for the term of the mortgage.  If you are not willing to do this we shall have to ask
you to arrange to take care of the mortgage when it becomes due. 

 from Gregg Speed Studies, 1917; page 58

(AGS – was developed by Andrew Owen)

(by jasonstracner2 for everyone)

23 comments Add yours
  1. Here is an image that shows how the ascii letters map to the shorthand strokes for the current AGS proposal.  If you would like me so send you a copy of the MSWord document so that you and contribute your ideas just let me know.   I was wondering if the diphthongs (eu, ou, oe, ia, …) should not be mapped to single ascii keys.  What do you think?

    Attachment: AGS (ASCII-Gregg-Shorhand).jpg

  2. ___Jason___
    In order to make a standard ASCII form for Gregg Shorthand, I had to represent each character with a single stroke.  Therefore, th could not be used to represent th, as it would be t and an h dot.  Instead, I use q, an otherwise unused stroke.  To express -ing, an h is placed at the end of the word. ——   Since Gregg includes thousands of linked forms, is there any reason why you wouldn't wish to use a DCBS mapping, as you would, say, for Arabic or Hebrew?   Kindest, Stenomouse

  3. ___JSW___ Since Gregg includes thousands of linked forms, is there any reason why you wouldn't wish to use a DCBS mapping, as you would, say, for Arabic or Hebrew? —–   Sorry! Obviously, I meant DBCS, that is "double-byte character system." The advantage of developing a double-byte code page is that you can have all the linked forms, as well as the single-byte forms you've mentioned. Using an Input Method Editor, you could then choose an abbreviated input style such as you've given, or you can choose a more natural, non-abbreviated input style, mapped to the same glyph. In other words, if you have the full-input method selected, you  type "mortgage"; if you have the abbreviated input method selected, you type "mOr." If both of these are mapped to the same glyph — i.e., to the same address on the code page– then you get the same results. A robust IME would also allow you to leave text UNCONVERTED — i.e., you type in "mortgage" and the word remains in standard roman characters.   Now, DEVELOPING such a code page — aye, THAT'S the rub. LOL.   Kindest, Stenomouse

  4. ___Jason___ I was wondering if the diphthongs (eu, ou, oe, ia, …) should not be mapped to single ascii keys.  What do you think? —–    Obviously, they are single-byte characters; each character is located at a single address on your code page. What keyboard strokes you choose to you map them to is a secondary issue.    Stenomouse

  5. When most people say ASCII they mean ANSI single byte characters.  That is what I am shooting for here.  The idea is to make a standard similar to what SAMPA (or XSAMPA or Kirshenbaum) is to IPA.  So if someone sends you some AGS text in an email you can read it or recreate the outlines that they meant (maybe even with a computer program).  Either way you will be sure that you are both using an interchangeable 'official' way of representing shorthand that will work without the need for fancy things like Unicode or a special font.
     
    Don't get me wrong, this IME things sounds super cool.  That is just not what I am shooting for.
     
    Is an IME system easy to create?  I have no experience with that.

  6. As someone who learned the Stenotype (pre-computer transcription era) and as someone who (used to) write Gregg at a high rate of speed, what's the speed potential of all these various mappings?   And the next question is if anyone is thinking of throwing away their pen is what happens when one makes a typo?  How can a mistake be corrected?  (There's a way to indicate an error with the Stenotype.)   Marc  

  7. I did not map the system for performance, unfortunately. No alphabetic type system stands a chance against the stenotype.

    Though, if a program used this mapping as shorthand, then I suppose the backspace key would be appropriate (no need to press * and retype it). 🙂

    —Andw. Owen

  8. Chuck: I can't find that book, and I've been looking for something like it for years. I use my own version of a Gregg for keyboard, and have created VBA code so I don't lose it when the tech department comes and updates my computer and installs a new MSWord version while i'm on vacation (damn them) but I only have about 300 words. Programming it in, and then organizing a way to keep it thru different templates is a problem.   There is a company in the UK, Agilityping, that has developed their own version of a keyboard shorthand and an AutoCorrect install program of 30,000 outlines. It apparently has been developed as a conflict free shorthand in the last 8 or 10 years. They have workshops etc. But it has a number of features that I find difficult — i.e. they use a two letter combination for the suffix -tion. And they substitute Z for virtually every S in the language.

  9. On a related subject …   McGraw-Hill in 1976 published a book titled "Gregg Computer Shorthand for Nonshorthand Writers", and reissued the book in 1987 under the name "Productivity Plus: Computerized Transcription for Word Processing and Keyboarding."  The idea was to use Gregg Shorthand abbreviations for speeding keyboarding with a special computer program.  The abbreviations include words and phrases.  This is sort of Stenotype meets Gregg.  So for example, if you want to write:   "Actually, the supervisor should not have the responsibility for revising specifications."   you would type in the special program   "Aktly, the svisr shdnv the rspl f rvisg spesfs."   Interesting idea that nowadays can be implemented in any word processing package by using the autocorrect feature.    

  10. This is inspiring, Jason. An ideal type abbreviating system for Gregg writers (and a way to turn my Gregg-oriented spelling errors to my advantage!).

    This may not be at all what you had in mind, but if the system is just for type i/o, you might consider shedding the handwriterly artefacts in the code. I don't think a comprehensive one-one mapping of every grapheme is necessary to retain all the Gregg spelling rules—nor even unicode-style rendering down the road. It may not even be an advantage for speed.

    The atomic elements of the Gregg alphabet map completely onto the lowercase ascii alphabet. The mapping is all phonetic with the exception of:

    c [ch]
    z [sh]
    q [th]
    w [ng]
    y [nk]

    a [ay, ah]
    e [ih, eh, ee]
    i [aye]
    o [oh, aw]
    u [uh, oo]

    The rest of the Gregg symbols are either blends of the atomic elements ("nd", "jnt", etc.), methods of applying the phonemes in ways that are more efficient for handwriting (dash under "w" sounds in the middle of words, dot above asperated vowels in the middle of words, dot for an "ah-" beginning, etc.), or punctuation. Punctuation could remain completely conventional, and the handwriting artifacts could be left out entirely, leaving only the blends, which are not strictly necessary to represent Gregg spelling. Nonetheless, they might be mapped to upper-case letters fairly simply:

    M [md, mt]
    N [nd, nt]
    J [jnt]
    V [dv, tv]
    F [tn, dn]
    G [tm, dm]
    W [mn]

    I [ia]
    E [ea]

    These mappings require the shift key modifier, and so represent an extra keypress. The case shift also conflicts with conventional type, and necessitates another "capitalization" mapping. I wonder whether the extra memory load would pay off for them.

    One last effect might be a character such as the period for the space between two disjointed graphemes in a word. Examples:

    p.oprV [postoperative]
    freN.z [friendship]

    Regarding the speed issue, a straight byte reduction ratio might give some idea of the upper limit.

    -Derek

Leave a Reply