Steno 101, Lesson One
Mirabai Knight
not. If you do, I'll stop linking to it here. But on the off chance
some of you don't subscribe to it, I wanted to let you know that
Lesson One of the Steno 101 series has been posted. I'd love to hear
any feedback you might have for it. Is it too much to learn at a time?
Confusingly laid out? Are the illustrations helpful? Comments and
criticism are very welcome, and I'll be sure to incorporate them into
future Steno 101 lessons.
Steno 101, Lesson One: http://bit.ly/b8uAoj
William Knight
I've been meaning to take the first lesson, been super busy, but I
will take it tonight and let you know of my impressions.
Steno, here I come!
mhack
Steno! I believe that Steno or something quite like it will probably
replace QWERTY in our ultra-networked-mobile-computing future.
I don't know if anyone else following the lessons has any prior
experience, but I come to it almost completely clueless, with only
memories of gazing dumbfounded at Mirabai's lightning capabilities. So
I will probably ask a lot of dumb questions until I've had more time
to think about it and 'get' it.
One of the first questions I have is this: I understand that briefs
are defined by individual stenographers, but is the 'long' form
completely unambiguous? It's not clear to me from the lessons how a
word can be unambiguously translated. For example, 'HAR/MON/KA':
pretty much every human would know that it translates to 'HARMONICA',
but how would a computer know it, other than being given a pre-defined
dictionary? I had to go over to the Wikipedia article about steno to
read that different schools define their own dictionary that
determines unambiguous translation, but the concept is still a little
unclear to me. If you received a steno transcript from a stenographer
from another school and didn't know that school's dictionary, would
you be able to translate the steno text and always get the exact same
result as the original stenographer? This is assuming that 'briefs'
are not used and the text is always in the 'long' form, which I know
is a hypothetical example.
Mirabai Knight, CCP
> I've read (quickly) through lessons 0 and 1 and am inspired to learn
> Steno! I believe that Steno or something quite like it will probably
> replace QWERTY in our ultra-networked-mobile-computing future.
This is my fondest dream, though I don't think it's going to happen
without a lot of work and a lot of luck.
> One of the first questions I have is this: I understand that briefs
> are defined by individual stenographers, but is the 'long' form
> completely unambiguous? It's not clear to me from the lessons how a
> word can be unambiguously translated. For example, 'HAR/MON/KA':
> pretty much every human would know that it translates to 'HARMONICA',
> but how would a computer know it, other than being given a pre-defined
> dictionary?
Well, in more sophisticated steno software there are phonetic fuzzy
logic algorithms that can sort of guess at a missing or mis-stroked
word (if you accidentally wrote "HAR/PON/KA", for instance, it would
probably that you were trying to write "HAR/MON/KA" and translate the
strokes accordingly), but basic steno programs like Plover require a
dictionary for translation. If you want to look through the dictionary
that comes with Plover, go here:
http://github.com/stenoknight/Plover/raw/master/eclipseDict.py
Entries in a dictionary are by no means unambiguous, and frequently a
stenographer will assign multiple groups of strokes to the same
English word or phrase. For instance, you can see that, in the Plover
dictionary, the word "Canada" is written:
KA/TPHA/TKA (pronounced "ka/na/da")
KAPB/TKA (pronounced "kan/da")
KAPBD/TKA (pronounced "kand/da"
So that the stenographer doesn't have to remember which specific
stroke corresponds to "Canada"; they can sound it out phonetically in
whatever way occurs to them at the moment, and the word is likely to
translate correctly. This is why I call steno a phonetic-mnemonic
writing system, because many entries in a dictionary (including all
briefs, by definition) are simply memory hooks, rather than strictly
rules-following.
> If you received a steno transcript from a stenographer
> from another school and didn't know that school's dictionary, would
> you be able to translate the steno text and always get the exact same
> result as the original stenographer?
No. Each stenographer personalizes their writing style to a greater or
lesser degree. It's often possible for one stenographer to read
another's steno notes and more or less figure out what is being
written, based on context and the general rules of a particular steno
theory, but even so there's considerable room for error, and if a
stenographer who didn't know every detail of my own particular system
-- even someone who studied the same steno theory -- tried writing on
software loaded with my dictionary, there would be a great many
mistranslated strokes. This is why each stenographer builds their own
dictionary, and why the process of building a dictionary and of
learning how to write in steno are inextricably interwoven.
William Knight
creation by each stenographer and grows with them over time, that is a
pretty cool thing about steno. But I want to be very clear about the
starting points. I looked at the plover dictionary in the python file
you gave and it looks pretty comprehensive, I'm estimating more than
100,000 words.
What is the source of that dictionary? Is it a standard reference
dictionary associated with a particular school? Presumably the school
that you went to? Does it have any of your customizations in it?
Mirabai Knight, CCP
34,000-entry dictionary, which I was given by my steno school. Then,
over the course of two years, I added about 18,000 entries to that
base dictionary while doing offline captioning of television shows at
my day job, then I bought commercial steno software, which came with a
free copy of the StenEd dictionary (about 100,000 words). StenEd's
theory is a close relative of the New York Career Institute theory
that I was taught, so with a few global changes (mostly having to do
with prefix/suffix handling) I was able to merge it quite successfully
into my personal dictionary. There were significant overlaps, so it
only wound up increasing the dictionary by around 43,000 words. I've
also added around 20,000 more during the last three years of providing
academic CART. So 77,000 entries were from two premade dictionaries
and 40,000 are my own additions. I'm sure the ratio will be different
with each stenographer, but starting with a broadly applicable base
dictionary of written-out theory-compatible entries and then refining
it over time with individual briefs and entries specific to one's own
daily material is a generally good way to go, and I think the Plover
dictionary works well as a base for other people to build on.
--
Mirabai Knight, CCP
917 576 4989
m...@stenoknight.com
http://stenoknight.com
Tony Wright
wrote:
> > If you received a steno transcript from a stenographer
> > from another school and didn't know that school's dictionary, would
> > you be able to translate the steno text and always get the exact same
> > result as the original stenographer?
>
> No. Each stenographer personalizes their writing style to a greater or
> lesser degree. It's often possible for one stenographer to read
> another's steno notes and more or less figure out what is being
> written, based on context and the general rules of a particular steno
> theory, but even so there's considerable room for error,
about was one of the early goals of machine stenography that actually
didn't pan out. I was looking at a very early textbook of machine
stenography from some time around 1915, I think, and in the preface,
the author mentioned one big selling point of the stenotype as being
that now, unlike with handwritten shorthand, one stenographer would be
able to read another stenographer's notes easily. I also read this
claim made in a trade school magazine of around the same time period.
It's easy to imagine why people would have thought this would be
feasible in the early days. They probably assumed there would be
basically one theory and everyone would use standard briefs and
phrases.
--Tony
Mirabai Knight
relevant now than it was back in 1915, because we don't have to worry
about transcribing steno notes manually anymore; there's no real
reason for one stenographer to show anyone else their untranslated
notes for anything but pedagogical purposes. Because the computer can
translate everything instantly into English, that can be the common
medium of exchange rather than the shorthand output itself. I think
this is actually an important reason why steno is in danger of dying
out. Before computerized word processors, lots of people learned
shorthand, both pen and machine, in high schools, business schools, or
home study courses. It was used not only by secretaries and court
reporters, but also newspaper reporters, novelists, and anyone else
who needed to take down words quickly and efficiently. Later, they'd
either transcribe their steno notes themselves using typewriters or
give them to an underling or scopist who was familiar with their style
of shorthand and could do the typing for them. The first phase was
much quicker than longhand note-taking, and the transcription process
took about the same amount of time as it did to type up longhand
writing.
Then in the '80s, people started using word processors, which could
produce infinite copies of printed text with far greater ease than
typewriters. Rather than depending on handwritten first drafts
(whether in longhand or shorthand), people began writing most things
out on the computer in the first place. Shorthand now required an
extra step that word processing didn't, so even though it was slower
to type things on a computer than to take them down in pen or machine
shorthand, the overall process was much quicker and more efficient if
everything stayed on the computer throughout. Shorthand schools
started shutting down, and soon the only people studying pen shorthand
were hobbyists, and the only people studying machine shorthand were
prospective court reporters. Computerized stenography with instant
translation came along hard on the heels of the personal computer, but
its price tag was astronomical, and never came down, even after
computers themselves became affordable to ordinary people. That was
the point at which shorthand stopped being a useful method you could
pick up casually and use for both personal and professional purposes,
and started to become only cost-effective as a focused moneymaking
pursuit, which, over the course of a few generations, has come to
threaten its own future. I don't think steno can survive without a
pool of hobbyists and dabblers, and no one is going to want to dabble
if they have to transcribe their notes manually like back in the old
days. That's why I think Plover is so important, and why I'm spending
so much time and money getting it off the ground. Once people take it
up and start learning steno again, I'm hoping that it will become
somewhat self-sustaining, because it's really not very hard to learn
and it has tremendous advantages over qwerty input methods. It just
needs a push to get going again, in a newly accessible computerized
context.
William Knight
mhack
Python 3.1 (you are pretty cutting edge!). But a question: can I run
plover with just a computer keyboard, or do I need to get a Sidewinder
or Gemini machine? I read the Readme.txt, but it didn't answer this
question for me. Maybe soon you can build a FAQ for stupid questions
like this!
Mirabai Knight, CCP
my big brother William, the one with the mohawk and the home-brew
wearable rig, whom you may know from Part 3 of What Is Steno Good For:
(http://plover.stenoknight.com/2010/06/mobile-and-wearable-computing.html)
Anyway, yeah, I think I might have to start working on a FAQ for
Plover. There are various bits of information scattered about the
blog, the github, and now this group, and it could do with some
consolidation. I'm putting it on my list.
Anyway, to catch up whoever doesn't know, Plover comes in essentially
two flavors: The serial version and the qwerty version. The serial
version works with steno machines running the Gemini PR protocol
(pretty much anything you can buy from
http://revolutiongrand.com/machines, though not their very earliest
discontinued models). They cost anywhere from $1,000 to $3,000, which
is way out of the budget of your ordinary steno dabbler. Fortunately,
the qwerty version of Plover, which works almost identically (it has
an intermittent freezing bug that the serial version lacks; I'm hoping
to hammer it out soon), can use any antighosting qwerty gaming
keyboard, the cheapest and most effective of which is the SideWinder
X4, which costs about $45 plus shipping from online discount
electronics shops.
(http://www.microsoft.com/hardware/gaming/productdetails.aspx?pid=105).
You can see a picture of the two machines side by side here:
http://stenoknight.com/plover/RGsidewinder.jpg
The steno layout on the SideWinder was made by cutting up little
sticky-backed squares of leather and using them to help bring the
ordinarily misaligned qwerty keys back into true.
Any other burning questions? I'll try to put together a proper FAQ in
the next week or so.
William Knight
featured in your Steno and wearable-computing saga. If I can dig up
some old pictures of my rig, I'll send them to you.
Ben Tarkeshian
> It's interesting to note that this consistency that William was asking
> about was one of the early goals of machine stenography that actually
> didn't pan out. I was looking at a very early textbook of machine
> stenography from some time around 1915, I think, and in the preface,
> the author mentioned one big selling point of the stenotype as being
> that now, unlike with handwritten shorthand, one stenographer would be
> able to read another stenographer's notes easily. I also read this
> claim made in a trade school magazine of around the same time period.
> It's easy to imagine why people would have thought this would be
> feasible in the early days. They probably assumed there would be
> basically one theory and everyone would use standard briefs and
> phrases.
>
> --Tony
I'm sure it was not for lack of trying :)
Please post any more things like that you have -- I love to compare
how things turned out
versus "what was supposed to happen."
<below is mostly off-topic, but quite parallel>
From a 1964 IBM book "Computers and Their Uses".
"The development of automatic programming techniques is also being
spurred by the fact that faster and better machines are continually
being placed on the market. To take advantage of these technological
advances, an equipment user would of course have to replace his
present machine with a more advanced computer. But in so doing he may
render all of his existing programs obsolete."
"One possible solution to this problem is a universal programming
language" ...
"A compiler would be written for each different machine. Every
compiler would be able to accept source programs in the universal
language, and would have the ability to convert this information into
a sequence of instructions for its particular computer."
Seems like the same idea for computer languages, which we all know how
well that turned out :)
""There seems to be little disagreement among commentators upon the
contemporary scene that we are rapidly approaching a state of affairs
in which almost all aspects of life will be engulfed by giant
industrial, governmental, and educational organizations. As Thurman
Arnold long ago emphasized, the idea that our economy is composed of a
large number of small, independent enterprises is a fictitious myth.
Even conservative spokesmen concur on the fact that ambitious and
intelligent young men must now increasingly seek careers in big
companies rather than in the formation of their own ventures. They key
problem which faces the modern world is the preservation and extension
of democracy, freedom, and opportunity within the large organizations
which control the lives of members of our society."
Why I am glad Plover exists!
The book was written by an IBMer too, which is very interesting :)
I am not making any judgements here :)
http://www.ibmandtheholocaust.com/northamerican.php
http://www.wired.com/politics/law/news/2001/02/41753
"The use of data processing techniques to substitute for teachers has
been severely criticized on the grounds that it would dehumanize an
important part of life and would fail to supply the warm, creative
atmosphere which is necessary for true learning. One can merely hope
that such equipment, if successful, would be used only for the duller
parts of the educational process, and would never be employed in an
attempt to supplant the inspirational and empathic aspects of the
teacher-student relation"
Seems fairly spot on :)
Nothing new, it just seems the best way to predict the future is to
look into the past...
Somewhere else (doesn't seem to be in there -- maybe Knuth or
somewhere else?)
I read hexadecimal was more accurately called sexadecimal, but that
was too racy at the time :)
I guess that debate is lost to history :)
Hope this isn't too far off base, there are just things you learn in
school and
things you pick up on your own later when you do your own research :)
Tony Wright
> Please post any more things like that you have -- I love to compare
> how things turned out
> versus "what was supposed to happen."
machine stenography and I often wonder why certain decisions were made
early on that survive to this day.
--Tony