Multi-stroke briefs

[Pragma Nolint]

Hi all,

It seems some multi-stroke briefs in the plover dictionary are nothing but the component chords stuck together. For example:

"E": "he"
"S": "is"
"E/S": "he is"

I was wondering what the reason behind this is? It works with plover, but in Fly it's looking at each word generated, so "E" then "S" is interpreted as "E" then "E/S" incorrectly giving "he he is". I got around this by compressing the words in lessons into briefs where possible--so any instance of "he is" will show the chord to type "E/S" rather than "E S"--but this decision has caused problems elsewhere in the code. It also presents the user with briefs which is not desirable. Could I perhaps safely remove these entries (from the Fly dictionaries only)?

Thanks,
Pragma

[Mirabai Knight]

Yeah, sorry about that. Feel free to delete those. They're an artifact
of my proprietary software's "Translation Magic" algorithm, which is
occasionally helpful when it's able to interpret a misstroked word
correctly by converting it to English spelling using a phonetics
table, using a spellcheck-style algorithm to pick the nearest valid
English word, and then outputting it without any confirmation from me.
The trouble is that sometimes it gets this horribly wrong (like the
time when I accidentally stroked "no depreciation" TPHO/KE/PRAOEGS
instead of TPHO/TKE/PRAOEGS , and Translation Magic rendered it
"nookie appreciation"), so in cases where, for example, it's
consistently turned my "is" stroke into a plural suffix, I've just
manually defined the phrase so Translation Magic wouldn't mess with
it. But none of that has anything to do with Plover, so please delete
at will. Sorry about the inconvenience!

Yours,
M.

--
Mirabai Knight, CCP, RPR, CBC
StenoKnight CART Services
917 576 4989
m...@stenoknight.com
http://stenoknight.com

[Pragma Nolint]

Thanks for the explanation Mirabai. That's great news that I can get rid of them; it'll simplify things a lot.

Cheers,
Pragma

[Josh Lifton]

I think the default Plover dictionary has its origins in one of Mirabai's
dictionaries, so hopefully she can chime in here.

Josh

[Josh Lifton]

This would be a great patch submission! Anyone?

Cheers,
Josh

[Margaret Synnott]

I'm having a go at this. Am I right in thinking that any entry
consisting of two strokes separated by a '/' which results in two
words would be an example of this?

Mags

(I haven't actually installed Plover yet, so I'm flying blind to some extent)

[Mirabai Knight]

I don't think that is true, unfortunately. For instance, today I had
"Foie gras" (TPOEU/TKPWRA) come up in a class I was CARTing.

[Margaret Synnott]

I realized that soon after I started editing.

I haven't deleted any entries with capitalization, or any of the latin phrases. And there are a few that seem to have something slightly more complicated going on, which I also left alone.

eg "UF/*ERPBD": "you've earned",

(I don't know what the '*' does here).

"AU/PAOER": "you appear", seems to leave out the "a" sound at the beginning of "appear"

I'm currently trying to figure out if "auxiliary verb" is actually made from two other entries.

Mags

[Mirabai Knight]

> eg "UF/*ERPBD": "you've earned",
>
> (I don't know what the '*' does here).

Um, I don't know either. My guess is that this was an ad-hoc
definition to eliminate a Translation Magic error, but I'm not
completely sure.

> "AU/PAOER": "you appear", seems to leave out the "a" sound at the beginning
> of "appear"

This is a misstroke defined to eliminate a common stacking error. It
should be three strokes -- U/A/PAOER. But my fingers or my machine or
a combination of the two have a tendency to strike the second key
before the first one's been completely released, resulting the stacked
strokes AU/PAOER. This is just a kludge to work around that.
Definitely not canonical.

> I'm currently trying to figure out if "auxiliary verb" is actually made from
> two other entries.

Could you say more? I'm not sure I follow.

> Mags

[Ian Dawson]

Is this an argument for having a "standard" theory, then we could use
the clean default dictionary that comes with that theory, rather than
Mirabai's personal one, with all the quirks of a working reporter's
dictionary.

It would also have the advantage of having readily available theory
book. Mirabai, I think your base theory NCRS (?) is pretty proprietry
to your former school, isn't it?

For example StenEd is a fairly "classic" standard theory, and books are
fairly easily come by in second hand online book stores.

On the downside, I suppose we run into copyright issues if we include
standard dictionaries, and other potential legal issues.

Maybe someone should invent one :) I doubt I'm clever enough to do so,
unfortunately.

Do we have a list on site of all the currently available theories which
are supported by dictionaries, and what their theory book is called so
they can be searched for? If not, I guess in volunteering to start such
a list off :)

Ian
From: Mirabai Knight
Sent: 08/02/2012 17:06
To: plove...@googlegroups.com
Subject: Re: Multi-stroke briefs

[Margaret Synnott]

> > I'm currently trying to figure out if "auxiliary verb" is actually made from
> > two other entries.
>
> Could you say more? I'm not sure I follow

"AUBGS/EUL/KWRAER/SRERB": "auxiliary verb"
"AUBG/SEUL/KWRAEUR": "auxiliary"

Auxiliary appears in various forms, but I don't think it's there as
"AUBGS/EUL/KWRAER"

Mags

[Mirabai Knight]

> "AUBGS/EUL/KWRAER/SRERB": "auxiliary verb"
> "AUBG/SEUL/KWRAEUR": "auxiliary"
>
> Auxiliary appears in various forms, but I don't think it's there as
> "AUBGS/EUL/KWRAER"
>
> Mags

Interesting. Yeah, that's not too surprising. Again, my guess is that
AUBGS/EUL/KWRAER/SRERB was misparsed by Translation Magic and I had to
define it explicitly as a corrective. Feel free to scratch it. But I
have a hunch that there's an awful lot of these. Is it simpler to go
through the dictionary and root these out, or to figure out a way to
make Fly not break when it encounters them?

[Matt Thomas]

Lists can be copyrighted, but only the reproduction of the entire list. For instance as long as you reproduced the effort, you could publish the same alphabetical list without copyright issues.

Sent in Transit (mst)

[Margaret Synnott]

I think I'm about 10% through the dictionary, and I'm not sure I've caught all the ones so far. Overall though, I think it's best to have the dictionary that's packaged with Plover be as clean as possible. It's also a very instructive exercise. I think when I finally get around to trying Plover out, I'll have some useful "mental hooks" to help me learn.

Mags

[Josh Lifton]

I completely agree that the packaged dictionary should be as clean as
possible. I've been through it a couple of times fixing/culling various
things and, like you, found it very instructive. I look forward to the
patch!

Thanks,
Josh

[Josh Lifton]

Ideally, we'd weed these out of the dictionary.

[Mauricio]

Is there a way that we can maintain different official versions of the
Plover dictionary for people to load as they need? i.e.practice
dictionary, expanded dictionary, et cetera.

Also, Mags, do you need a hand working through the dictionary? I might
have some time this weekend to take a look at it. We can start at
opposite ends and work our way towards the middle.

Mau

[Josh Lifton]

Support for multiple dictionaries is something I'd like to add after
just-in-time dictionary editing. Now is a good time to air out ideas for
what this feature should and shouldn't include.

Josh

[Margaret Synnott]

That would be great, Mau. It's quite big! :) I've started at the
start, and am about 9% through. I'll meet you somewhere around
"metaphysical"? :)

Mags

[Mauricio]

Hi all

Quick status update on my end: I started today, about an album and
three songs ago. Sitting just over two thousand words in from the
bottom at "quod", I'm finding surprisingly few of these problem multi-
stroke briefs. However, I have come across quite a number of stacking
corrections and boundary error solutions, such as "quote around"
instead of "quota round". Would those present a problem? Should they
be removed as well?

Mau

On Feb 10, 1:56 am, Margaret Synnott <margaret.synn...@gmail.com>
wrote:

[Mauricio]

Another quick question from me...
What is the meaning of the FLUSH definition TPHR-RB and THR-RB? Is it
a command stroke for Plover?

Mau

[Mirabai Knight]

On Sun, Feb 12, 2012 at 2:00 PM, Mauricio <mau.bus...@gmail.com> wrote:
> Another quick question from me...
> What is the meaning of the FLUSH definition TPHR-RB and THR-RB? Is it
> a command stroke for Plover?

Great question. That stroke is actually a big reason why Plover
exists, though fortunately it's not a stroke that we need to implement
in Plover, and it can safely be taken out of the Plover dictionary.
Basically, all proprietary steno software uses a time-based buffer,
which means that the strokes aren't released to the OS until a
specified amount of time has passed. You can change the length of this
buffer manually, but if you set it any shorter than 1.5 seconds or so,
you're at risk of seeing split strokes -- like KAT/HROG being
translated as "cat log" instead of "catalogue". So I have to keep my
stroke timer set to 1.5 seconds, but because my CART clients don't
like having to wait 1.5 seconds to see each stroke displayed, I have
to manually flush the buffer by invoking the {FLUSH} definition every
time there's the slightest pause in the flow of speech. According to
my steno software's dictionary editor, I've used the {FLUSH} stroke
about 100,000 times. That's 100,000 wasted keystrokes. It's maddening.
One of the biggest advantages Plover has over all the $4,000 software
out there is that it uses a length-based buffer rather than a
time-based one, so the translations are released to the OS
immediately, instead of after an arbitrary delay. It's also the reason
Plover can actually interact with the OS in a meaningful way, beyond
just text entry. 1.5 seconds may not seem like much, but imagine
having to wait 1.5 seconds after delivering any command, from Enter to
Backspace to Alt-Tab (most proprietary software can't even send
Alt-Tab as a command, but never mind...); it's horrendously
frustrating and completely unworkable. The time-based buffer has kept
all proprietary steno software subfunctional for decades. Plover's the
only steno program I've ever seen that doesn't have this problem. The
difference is nothing short of staggering.

[Carol Berk]

I don't know anything about Plover, but I worked on the first CAT system for court reporting, which was commanded in DOS. An asterik told the computer to connect The stroke to the word that followed. A*\bout = about. Sorry that the iPad types funny, but you get the idea. Might be useful in today's world, too. Who knows?
And if it's multi stroke why do they call it a brief????? We used short forms, like burp for burden of proof. The older I get the less I know.
Sent from my iPad

[Mirabai Knight]

On Sun, Feb 12, 2012 at 10:04 PM, Carol Berk <cbe...@gmail.com> wrote:

I don't know anything about Plover, but I worked on the first CAT system for court reporting, which was commanded in DOS.  An asterik told the  computer to connect The stroke to the word that followed. A*\bout = about.  Sorry that the iPad types funny, but you get the idea.  Might be useful in today's world, too.  Who knows?

I strongly advise that stenographers don't use the asterisk for prefixes, because fingerspelling is incredibly important both for defining new words that don't appear in the dictionary and for spelling words out letter by letter when there isn't time to define them. The quickest and least error-prone method of fingerspelling is to use the asterisk with the left hand alphabet for lowercase letters and the asterisk plus one other letter on the right hand (I use asterisk P, because the index and middle fingers are stronger and more reliable than the ring or pinky fingers) for uppercase letters.

And if it's multi stroke why do they call it a brief?????  We used short forms, like burp for burden of proof.  The older I get the less I know.
Sent from my iPad

brief, abbreviation, short form, arbitrary -- all terms for the same thing: a non-phonetic steno outline for a longer word or phrase. BURP for burden of proof might be useful in a courtroom, but it's not a good idea to use it in a situation where someone might actually say the word "burp". I've personally written "burp" 102 times since I first started steno. Yes, I used to work in television. 

[Mauricio]

Very interesting... Thank you very much for your explanation.

I just wanted to check in with my progress. I'm sitting at line
106,930 of 124,312 working backwards. My edited file shows a line
total 124,076 now. I went ahead and removed all the flush commands,
which were quite a few! Since they're all over the place, I recommend
that Mags does the same in her file.

Mau


On Feb 12, 9:19 pm, Mirabai Knight <askelad...@gmail.com> wrote:

[Margaret Synnott]

I started back at the beginning of my file to catch all the {FLUSH}
entries. Before I knew it I'd deleted them all, but as long as Mau and
I submit diff files that shouldn't be a problem, right?

Another couple of quick questions:

What's the story with entries that are multiple words, but no '/', just a '*'?
"SKPHR*EUFP": "literature and",

Also, I wasn't sure what to do with these:
"SOPL/ABG": "some academ",
"TPHRO*ER/-T": "error in the",

Mags

[Mauricio]

Those longer briefs that don't have a "/" are just longer single
stroke briefs that represent multiple words. I've been leaving the
longer briefs alone as I go through my parts. I might never use them,
but I can imagine that someone who is used to them might miss them if
they're gone. Actually, two that I intend to start using right away
are "thank you very much" "TH*UFP" and it's partner "thank you so
much" "TH-UFP".

The asterisk it just used as a kind of shift key to make a different
combination. It is often used to type the less common word that would
also fit with the definition, for instance, R-EPL = "recommend" and
R*EPL = "realm". In this case, recommend is a brief in itself, like
the ones I mentioned above, so it doesn't really fit with that
definition, but that's the best example I could recall right now.

The last two things you listed are probably custom entries a) for
words that would never follow each other like "some ABG" or "some ack"
and therefore the best thing to do with that combination is to guess
that the person is trying to write "some academ" (or it could be a
typo), and b) another shortcut brief that experienced stenographers
often use. It looks like it is stacking "error" and "in" in the first
stroke. This would normally be undefined, but in that order with "the"
following, it interprets as "error in the". I would leave those
entries alone as well.

Is this correct, Mirabai?

Mau

On Feb 14, 3:08 am, Margaret Synnott <margaret.synn...@gmail.com>
wrote:

[Margaret Synnott]

I've attached a diff file (from-file: original, to-file: edited file).
When I first started I thought Mirabai would be able to have a quick
look over my list to check my work. It turns out I've deleted over
2000 lines, just in my half.

Mirabai, could you look over the list below? They're the lines I
wasn't confident about deleting, but I wasn't sure I should leave in
either.

Mags

"-T/HO*US": "the House" <------
capital H
"A/PHO*EUPB/SA*EUD": "amino acid",
"PHA*LT/THAS": "abnormalities that"
"URPBD/T-RBGS/PWUT": "understand it{,} but",
"U/SPHAOEPB": "you mean so",
"U/TRUPB": "you run it" <-----
TRUPB is not in the dict as "run it"
"U/TPHOERBGS": "you know{,}"
"U/TPWURPB": "you burn it",
"-T/TA*EUFP": "the taxpayer", <-----
taxpayer appears in several phrases like this, but not on its own
"-T/TA*EUP/-G": "the taxpaying",
"-T/TA*EUP": "the taxpayer",
"-T/STA*EUP": "the taxpayer",
"-T/SKWRA*ER": "the area is",
"SPHAEUPBLG": "major is",
"STAT/SKWRUS/SKWOE": "status quo is",
"HOU/R/STPHU": "how are you{?}"
"KPHEBG/TPEURBS/SEU": "economic efficiency",
"TKE/TKPHRAOEUPB": "decline in",
"TKU/HREF": "did you leave",
"TKOU/-F": "do you have",
"KREBGT/T/-FPLT/-D": "corrected it{.}",
"KOE/HOFRT": "cohort of",
"KOULD/AO*EF": "could've",
"KUT/TPHAOS": "cuts into"
"K/TPHEUBGS/TP": "can fix this if",
"2K3W50D/SOEU/PWAOEFPB": "good soybean",
<----------- numbers
"PWROE/PWOPB/SKWRO/PWAEGS": "pro bono basis",
"STPEBG/TPHOT": "if she cannot",
<----------- lots of entries similar to this
"STPEBG/-PBT": "if she can't",
"STHEUPBG/T-RB": "thing is --",
"SKPEUPB/KRAOES": "and increase",
"SKPEUBG/TPHOT": "and I cannot",
"SKRAOEUPB/TEUFPLT": "a scientist{.}",
"PHAEUPB/THRAPB": "main reason that",
"PHOR/TPWHEUPBG": "more thing with",
"PHEUD/PHEUD": "mid", <-------- should
this just be "PHEUD" ?

[Mirabai Knight]

I think you can safely delete all of these. They're all either slop misstroke definitions or stacking corrections, so they don't need to be in the default dictionary. Thanks so much for all your hard work! It's pretty damn impressive.

--

M.