Status, states, steps and stages
Charles Matthews
but can one build an industrial strength version?
What I mean is something like this. Unsettled status implies a diagram like
this:
L <- U -> D
meant to imply a system of three states, U which is unsettled (alive or dead
with a move) and resulting live and dead states L and D into which it can be
shifted.
Ko complicates matters enormously. A direct ko is like
L <- U <-> V -> D
with a possible oscillation between internal states U, V. And the initial
state W (whatever) will belong in a diagram L' <- W -> V signifying that the
defending player can erase ko for life with one play, the attacker can play
into the ko state where the defender is to take the ko.
That's not so bad, but not all kos are direct :(. I'm going to have to face
up to the terminology problems in writing about ko for Gobase, so I though I
might as well air the issue now.
One major type is like
L <- X <-> Y <-> Z -> D
so there are three states with possible oscillation. This is 'two-stage' in
the terms introduced in Davies' Life and Death; but Jim has told me in
correspondence that he now doesn't like that, in accord with Bill Spight's
view that 'two-step' is better for this. Therefore I propose to adopt
'two-step' if I have to pin this down. Of course if you're speaking about
status of a group there is a distinction between
L' <- W -> Y
and
L' <- W - > Z
as two-step kos, namely that in the first case the attacker need only ignore
one threat, in the second case it's ignore two threats to win the ko.
The other major type in practice (theory is another matter) is the indirect
ko (yose ko). This really is more appropriately called multi-stage: for
example a typical cascade towards killing a group is via
X <-> X' -> Y <-> Y' -> D
where X' -> Y is an irreversible change (normally filling a liberty) moving
the ko one stage further on.
This seems a bit sketchy so far. Anyone have a more solid picture?
Charles
Douglas Ridgway
> One major type [of ko] is like
>
> L <- X <-> Y <-> Z -> D
>
> so there are three states with possible oscillation. This is 'two-stage' in
> the terms introduced in Davies' Life and Death; but Jim has told me in
> correspondence that he now doesn't like that, in accord with Bill Spight's
> view that 'two-step' is better for this. Therefore I propose to adopt
> 'two-step' if I have to pin this down.
[...]
> The other major type in practice (theory is another matter) is the indirect
> ko (yose ko). This really is more appropriately called multi-stage: for
> example a typical cascade towards killing a group is via
>
> X <-> X' -> Y <-> Y' -> D
>
> where X' -> Y is an irreversible change (normally filling a liberty) moving
> the ko one stage further on.
So what Davies' books called multistep and multistage are now to be
referred to as multistage and multistep, respectively? Isn't this
proposing to swap the meanings of two relatively arbitrary but closely
related technical terms in contradiction to much of the literature?
From my point of view, I have enough trouble keeping them straight as
it is. The concern is that the two terms will have no unambigous
meaning whatsoever, i.e. if I read "multi-step" or "multi-stage" ko,
I'll know that it's more complicated than a regular basic ko, but
beyond that, I won't know anything for sure, and I'll have to figure
it out myself. Perhaps I'm overestimating how consistent the English
literature is already, though. As one citation, here's a decade old
article using "step" like Davies to refer to the second type of ko:
http://groups.google.com/groups?hl=en&lr=&ie=UTF-8&selm=1992Feb27.124938.25552%40rhrk.uni-kl.de
. I've also found Bill Spight referring to the second type as
two-step. Here's another decade old article referring to the first
type as stage: http://groups.google.com/groups?hl=en&lr=&ie=UTF-8&selm=1992Dec17.144429.19902%40irisa.fr.
I haven't found anyone referring to the second type as "stage", Bill
seems to use "n-move approach ko" for this.
Other possible terms are "reversible" and "irreversible", but that
might simply introduce more confusion.
It might be a good idea to include a glossary, whatever you decide to
do.
Douglas Ridgway
Charles Matthews
"Douglas Ridgway" wrote
> So what Davies' books called multistep and multistage are now to be
> referred to as multistage and multistep, respectively?
Yup. Because it looks like a false start.
> It might be a good idea to include a glossary, whatever you decide to
> do.
See p.116 of Teach Yourself Go for a summary of the state of play as I found
it. NB that the Jin Jiang book Fighting Ko already uses two-step
differently.
Charles
Ted S.
<charles.r...@ntlworld.com> wrote in
news:YNBZ8.1544$3x....@newsfep3-gui.server.ntli.net:
> That's not so bad, but not all kos are direct :(. I'm going to have
> to face up to the terminology problems in writing about ko for Gobase,
> so I though I might as well air the issue now.
Why not just write about thickness instead? A series on thickness is
something I'd certainly appreciate. :-)
--
Ted S.: change .spam to .net to reply by e-mail
Homer Simpson: I'm sorry Marge, but sometimes I think we're the worst
family in town.
Marge: Maybe we should move to a larger community.
Charles Matthews
> Why not just write about thickness instead?
I don't think I have anything out of the ordinary to say say about it. I'm
quite careful not to get thickness too often in my own games.
Charles
Ted S.
Well, then, what do you do in the following position:
10 | . . . , . . . . . , . . . . . , . . . | 10 [White 1]
9 | . . . . . . . . . . . . . . . . . . . | 9
8 | . . . . . . . . . . . . . . . . . . . | 8
7 | . . . . . . . . . . . . . . . . . . . | 7
6 | . . 1 . . . . . . . . . . . . . . . . | 6
5 | . . . . . . . . . . . . . . . . . . . | 5
4 | . . . # . . . . . # . . . . . # . . . | 4
3 | . . . . . . . . . . . . . . . . . . . | 3
2 | . . . . . . . . . . . . . . . . . . . | 2
1 | . . . . . . . . . . . . . . . . . . . | 1
---------------------------------------
A B C D E F G H J K L M N O P Q R S T
When White approaches like this, isn't a pincer attack indicated because
it works well with the stone at K4? (No fair telling me you don't play
sanrensei and so this exact position doesn't come up. And no fair telling
me White wouldn't approach sanrensei in this manner either. :-) Should
white respond to a pincer attack with C3, Black will end up with
thickness. At least, if my understanding of thickness is correct, Black
will end up with thickness.
I'm not getting thickness, influence, and heaviness mixed up (again), am
I?
Charles Matthews
> Well, then, what do you do in the following position:
>
> 10 | . . . , . . . . . , . . . . . , . . . | 10 [White 1]
> 9 | . . . . . . . . . . . . . . . . . . . | 9
> 8 | . . . . . . . . . . . . . . . . . . . | 8
> 7 | . . . . . . . . . . . . . . . . . . . | 7
> 6 | . . 1 . . . . . . . . . . . . . . . . | 6
> 5 | . . . . . . . . . . . . . . . . . . . | 5
> 4 | . . . # . . . . . # . . . . . # . . . | 4
> 3 | . . . . . . . . . . . . . . . . . . . | 3
> 2 | . . . . . . . . . . . . . . . . . . . | 2
> 1 | . . . . . . . . . . . . . . . . . . . | 1
> ---------------------------------------
> A B C D E F G H J K L M N O P Q R S T
>
> When White approaches like this, isn't a pincer attack indicated because
> it works well with the stone at K4? (No fair telling me you don't play
> sanrensei and so this exact position doesn't come up. And no fair telling
> me White wouldn't approach sanrensei in this manner either. :-) Should
> white respond to a pincer attack with C3, Black will end up with
> thickness. At least, if my understanding of thickness is correct, Black
> will end up with thickness.
>
> I'm not getting thickness, influence, and heaviness mixed up (again), am
> I?
I think the normal answers are Black C6 or Black F4. It happens that the
Takagawa book "The Power of the Star Point" emphasises the pincer C6.
Doesn't make it a better play. Black F4 happens more often in pro games
now. After which White usually settles on the left side. Then Black has
influence rather than thickness. When Black plays C6 White normally does
invade at C3, rather than fight with a weak group on the edge of Black's
framework (which would favour Black expanding the framework by attacking).
The conventional continuation gives a wall facing east - it can be cut
though, so calling it thickness is a stretch.
Charles
Charles Matthews
> I think the normal answers are Black C6 or Black F4.
I did mean C8 rather than C6.
Charles
Douglas Ridgway
"Charles Matthews" <charles.r...@ntlworld.com> wrote in message news:<xGf_8.4280$Hy1.1...@newsfep1-win.server.ntli.net>...
> "Douglas Ridgway" wrote
> > So what Davies' books called multistep and multistage are now to be
> > referred to as multistage and multistep, respectively?
> Yup. Because it looks like a false start.
Perhaps it is a false start, but keeping the same terms and just
swapping the meanings seems like a bad idea to me. A couple of other
references to consider: the 1996 Mueller-Berlekamp-Spight paper on
thermography [1] uses the terms "iterated ko" and "approach ko" to
refer to what Davies calls multistage and multistep. Landman 1996 [2]
calls them "two-stage" and "approach move ko or multistep ko". The MBS
paper is particularly nice in that it has 70+ example positions with
move graphs.
What would be really helpful is an agreed on, standardized,
systematic, general nomenclature for cyclical positions. Sort of like
IUPAC chemical names, but for kos. Unfortunately, this probably
requires more classification and taxonomy than has been done so far.
[1] M. Müller, E. Berlekamp, and B. Spight. Generalized thermography:
Algorithms, implementation, and application to Go endgames. Technical
Report 96-030, ICSI Berkeley, 1996. Available from
http://web.cs.ualberta.ca/~mmueller/publications.html
[2] H. Landman. Eyespace values in Go. In _Games_of_No_Chance_, MSRI,
1996. Available from
http://www.msri.org/publications/books/Book29/contents.html.
Douglas Ridgway
Simon Goss
>"Charles Matthews" <charles.r...@ntlworld.com> wrote in message
>news:<xGf_8.4280$Hy1.1...@newsfep1-win.server.ntli.net>...
>> "Douglas Ridgway" wrote
>
>> > So what Davies' books called multistep and multistage are now to be
>> > referred to as multistage and multistep, respectively?
>
>> Yup. Because it looks like a false start.
>
>Perhaps it is a false start, but keeping the same terms and just
>swapping the meanings seems like a bad idea to me. A couple of other
>references to consider: the 1996 Mueller-Berlekamp-Spight paper on
>thermography [1] uses the terms "iterated ko" and "approach ko" to refer
>to what Davies calls multistage and multistep. Landman 1996 [2] calls
>them "two-stage" and "approach move ko or multistep ko". The MBS paper
>is particularly nice in that it has 70+ example positions with move
>graphs.
Just swapping the terms around is rare. The usual term from Davies' "N-
step" ko now is N-move approach ko, which IMO is excellent since it says
exactly what it means.
I agree with you that recycling terminology is asking for confusion.
"Iterated" seems very clear to me.
--
Simon
Robert Jasiek
Douglas Ridgway wrote:
> What would be really helpful is an agreed on, standardized,
> systematic, general nomenclature for cyclical positions.
Yes. However, those are researched in game theory, go theory,
go rules theory and will be researched for years or decades. It
is too early to standardize something of that great parts are not
even discovered yet. Just how serious the issue is can be seen
from the a few years old example of my declaration that "ko" is a
bad term because principally every string can disappear during
cycles and because there are other restrictions than those being
related to shapes involving ko. As a consequence, now in go rules
theory I prefer to speak of "restriction rules" rather than "ko
rules".
--
robert jasiek
http://home.snafu.de/jasiek/koclass.html
Robert Jasiek
Charles Matthews wrote:
> This seems a bit sketchy so far. Anyone have a more solid picture?
Status of all stones on the board depends on global game tree
analysis and therefore matters become more complex than your
linear diagrams for local stones suggest. To offer an example
(compare the ICOB2001 booklet), just do not forget the following
status tree describing only one particular stone:
( dead | bent-4-in-the-corner )
Note that under typical rulesets this is unequal to ( dead | dead ) .
--
robert jasiek
Bill Spight
Douglas:
> > > So what Davies' books called multistep and multistage are now to be
> > > referred to as multistage and multistep, respectively?
Charles:
> > Yup. Because it looks like a false start.
Douglas:
> Perhaps it is a false start, but keeping the same terms and just
> swapping the meanings seems like a bad idea to me.
Bill: Me too.
Douglas:
> A couple of other
> references to consider: the 1996 Mueller-Berlekamp-Spight paper on
> thermography [1] uses the terms "iterated ko" and "approach ko" to
> refer to what Davies calls multistage and multistep. Landman 1996 [2]
> calls them "two-stage" and "approach move ko or multistep ko". The MBS
> paper is particularly nice in that it has 70+ example positions with
> move graphs.
>
> What would be really helpful is an agreed on, standardized,
> systematic, general nomenclature for cyclical positions. Sort of like
> IUPAC chemical names, but for kos. Unfortunately, this probably
> requires more classification and taxonomy than has been done so far.
>
It does, but the traditional Japanese terms cover a lot. Unfortunately,
they
are at odds with some earlier English terms. In particular, a yose ko
(which may be simply translated as "approach ko") has been called a
2-step
ko, while a nidan ko (which may be simply translated as "2-step ko") has
been called, I understand, a 2-stage ko. The possibilities for confusion
are apparent. I think that we should stick with the traditional Japanese
categories and their obvious renderings in English. :-)
> [1] M. Müller, E. Berlekamp, and B. Spight. Generalized thermography:
> Algorithms, implementation, and application to Go endgames. Technical
> Report 96-030, ICSI Berkeley, 1996. Available from
> http://web.cs.ualberta.ca/~mmueller/publications.html
>
> [2] H. Landman. Eyespace values in Go. In _Games_of_No_Chance_, MSRI,
> 1996. Available from
> http://www.msri.org/publications/books/Book29/contents.html.
>
At the same time, there are kinds of kos and ko positions that do not,
to my
knowledge, have traditional terms. One example is the rogue ko in
"Mathematical
Go", by Wolfe and Berlekamp. About 5 years ago I wrote up a broader
classification
scheme for a mathematical go dictionary. While it does not cover
superkos or
multiple kos, I am unaware of any single ko that does not fit into it.
:-)
(OC, a ko may fall into more than one category, e. g., a one-move
two-step approach ko.)
Here are the relevant entries from "Mathematical Go Dictionary",
copyright (C) 1997-
2002, by William L. Spight:
<<
Approach ko
Yose ko.
Iterated ko
A position in which taking ko produces the option of taking
ko or prevents the opponent from taking ko. Iterated kos are
distinguished by the number of times ko is taken between winning
positions. The times are called steps: thus, a two-step ko, three-
step ko, etc.
Example: A three-step ko (top side). - O X - X O O - -
- O O X - X O O -
- - - X X - X O -
X X X O -
Mannenko (mahn-nen') (From man, 10,000, + nen, year)
1. A ko which either player may, after taking, convert to a
hotter ko by filling a shared liberty. Example: - X O a X - X
(Top right corner.) Black may convert to a hot ko - X O b X X O
by playing at 'a' or 'b'. White may do likewise - X O O O O O
after taking the ko. If neither player converts, - X X X X X X
the position is seki. - - - - - - -
2. A throw-in ko which is an alternative option to seki or
which prevents seki. Example from Mathew MacFadyen: - X O X - X a
(Top right corner). White may convert to a hot ko by - X O - X O b
playing at 'a'. Black may convert by playing at 'b', - X O O O O O
or make seki at 'a'. (N. B. Black may still make ko - X X X X X X
after a play at 'a' by playing at 'b' and sacrificing - - - - - - -
3 stones.)
Rogue ko
A position in which a player, when he has the option of
winning the ko, also has the option of making a play which converts
the ko to a hotter ko. Rogue kos are typically hyperactive.
Example: (Top right corner.) Black may fill the ko at - X X X a
'a' or play at 'b' and make a hotter ko. - X - O X
SEE yose ko. - X X b O
- - - O O
- - -
[Today I would add that rogue kos, like approach kos, may have a
series of moves to make hotter kos. So we can talk about a 2-move
rogue ko, for instance. -- Bill Spight]
Sente ko
A ko which to take or win is sente. Examples:
1) (Top right corner.) If White takes the ko, he puts - - - X X - X
Black's four stones in atari. - X X O X X O
- - - O O O O
- - - - -
2) (Top side.) If White wins the ko, he - X - X - X - X O - O -
threatens to capture Black's group. It - X O X O X O X O X O -
was also sente for White to take the ko. - - O X X X X O - O O -
- O O O O O X O O - -
- - - - - - X - - -
Ten-thousand-year ko
Mannenko.
Thousand-year ko
Ten-thousand-year ko.
Throw-in ko
A ko created by placing a single stone in atari. Example:
X X O - - O (Top left corner.) White may create a ko by playing
- X O - - O at 'a', Black by playing at 'b'.
X a b O O O
X X O O - -
- X X - -
Yose ko (Yo'seh) (From yoseru, to approach)
A ko in which one player, having taken the ko, cannot win it
immediately, but must reduce the opponent's liberties first.
Yose kos are distinguished by the number of liberties to be
filled to make the position a direct ko; thus, a one-move yose
ko, two-move yose ko, etc. Yose kos are hyperactive. SEE rogue ko.
>>
Best,
Bill Spight
Simon Goss
Useful set of definitions :)
>It does, but the traditional Japanese terms cover a lot. Unfortunately,
>they are at odds with some earlier English terms. In particular, a yose
>ko (which may be simply translated as "approach ko") has been called a
>2-step ko, while a nidan ko (which may be simply translated as "2-step
>ko") has been called, I understand, a 2-stage ko. The possibilities for
>confusion are apparent. I think that we should stick with the
>traditional Japanese categories and their obvious renderings in English.
>:-)
Agreed. The issue is what is the most useful translation of "-dan" in
this context. Do you consider that "-step" is the only acceptable one?
There's quite a high price in terms of confusion in using "step"
anywhere in the ko terminology, because of the history. To my mind,
"step" and "stage" are interchangeable when talking about to-and-fro
processes of this kind.
Best,
--
Simon
Bill Spight
>
> Useful set of definitions :)
>
Thank you. :-)
> >It does, but the traditional Japanese terms cover a lot. Unfortunately,
> >they are at odds with some earlier English terms. In particular, a yose
> >ko (which may be simply translated as "approach ko") has been called a
> >2-step ko, while a nidan ko (which may be simply translated as "2-step
> >ko") has been called, I understand, a 2-stage ko. The possibilities for
> >confusion are apparent. I think that we should stick with the
> >traditional Japanese categories and their obvious renderings in English.
> >:-)
>
> Agreed. The issue is what is the most useful translation of "-dan" in
> this context. Do you consider that "-step" is the only acceptable one?
I strongly prefer it.
> There's quite a high price in terms of confusion in using "step"
> anywhere in the ko terminology, because of the history. To my mind,
> "step" and "stage" are interchangeable when talking about to-and-fro
> processes of this kind.
>
There is a word, "dan", that translates to "stage", but it's different
from the one in nidan ko, etc. Besides, the English "stage", while
indicating progression, seems less reversible to me than "step".
Iterated kos are sometimes fought at different points during the same ko
fight. Returning to stage 1 after stage 2 seems more awkward to me in
English than returning to step 1. But taking a step backwards is normal.
However, in such cases I find myself referring to the "leg" of the ko,
like the leg of a trip. So maybe we could refer to iterated kos by the
number of legs they have. OTOH, "4-leg ko", or worse, "4-legged ko"
sounds funny, too. <shrug>
Best,
Bill
Bill Spight
P. S. While I find myself referring to different legs of iterated kos, I
also find myself referring to different stages of approach kos (which
are irreversible). That's another reason I do not like calling iterated
kos N-stage kos.
Bill
Charles Matthews
<snip>
> There is a word, "dan", that translates to "stage", but it's different
> from the one in nidan ko, etc. Besides, the English "stage", while
> indicating progression, seems less reversible to me than "step".
> Iterated kos are sometimes fought at different points during the same ko
> fight. Returning to stage 1 after stage 2 seems more awkward to me in
> English than returning to step 1. But taking a step backwards is normal.
I'm grateful to have sight of some of Bill's definitions, since up till now
I've been working on second-hand versions.
Approaching it all from a seriously mathematical point of view - we are
given some directed graph G of states and transitions, of an arbitrarily
complicated repetitive position. About the one clear-cut operation we can
do to G is to try to map in the best possible way to a directed graph H in
which repetition (loops of directed edges) can't occur. We can try to do
that by clumping states s and t if there is a directed path from s to t, and
also one from t to s. I suppose there is some standard name for this. I'd
want to call H the graph of stages of G (well, properly speaking, its vertex
set would consist of the stages). In complex examples where you need to do
more than identify those s and t just mentioned this might not be a very
useful notion.
Charles
Douglas Ridgway
> > >It does, but the traditional Japanese terms cover a lot. Unfortunately,
> > >they are at odds with some earlier English terms. In particular, a yose
> > >ko (which may be simply translated as "approach ko") has been called a
> > >2-step ko, while a nidan ko (which may be simply translated as "2-step
> > >ko") has been called, I understand, a 2-stage ko. The possibilities for
> > >confusion are apparent. I think that we should stick with the
> > >traditional Japanese categories and their obvious renderings in English.
> > >:-)
> >
> > Agreed. The issue is what is the most useful translation of "-dan" in
> > this context. Do you consider that "-step" is the only acceptable one?
>
> I strongly prefer it.
>
> > There's quite a high price in terms of confusion in using "step"
> > anywhere in the ko terminology, because of the history. To my mind,
> > "step" and "stage" are interchangeable when talking about to-and-fro
> > processes of this kind.
> >
>
> There is a word, "dan", that translates to "stage", but it's different
> from the one in nidan ko, etc.
And this is why you have a smiley above, right? Because it's not clear
even to English speakers who know some Japanese what the "obvious
rendering" is? In any case, it seems like a bad idea to expect English
readers to have to translate words back into Japanese in order to
understand their meaning. Better to have well-defined English words,
derived from the Japanese if necessary.
> Besides, the English "stage", while
> indicating progression, seems less reversible to me than "step".
> Iterated kos are sometimes fought at different points during the same ko
> fight. Returning to stage 1 after stage 2 seems more awkward to me in
> English than returning to step 1. But taking a step backwards is normal.
To me, also, the plain readings of "step" and "stage" seem
interchangable in this context, with neither seeming to connote more
or less reversibility. I'm a native speaker, so gut feelings count,
but to back this up, I grabbed some dictionary definitions (American
Heritage via dictionary.com, but it shouldn't matter).
step: 5a) One of a series of actions, processes, or measures taken to
achieve a goal. b) A stage in a process: followed every step in the
instructions.
stage: 11a) A level, degree, or period of time in the course of a
process: the toddler stage of child development; the early stages of a
disease. b) A point in the course of an action or series of events:
too early to predict a winner at this stage
Both words have been used to refer to both reversible and irreversible
processes (e.g. step in journey, step in musical interval, disease
stage, smog alert stage are all reversible, step in computer
instruction, stage of child development, stage of rocket are all
irreversible).
We can, of course, create a distinction in technical go usage by
consistently using them differently, as is being done for "counting"
and "scoring". This is at the mercy of usage, however. Unless the
words have been used consistently in the existing literature, however,
an author cannot use them to convey that distinct meaning and be
reasonably certain that the reader will understand. I think that this
is the real question: what words should an author use when writing
about ko in English in 2002.
To me, "step" and "stage" by themselves are inadequate, so I would
always write "approach ko" or "iterated ko" for kos with irreversible
and reversible states respectively. The word "approach" is
established, natural in English, and is apparently also the natural
translation of the Japanese term. "Iterate" is less well established,
and isn't entirely obvious or natural (to me anyway), but at least is
being used consistently. The different sizes of approach ko, as well
as their states, can be called "N-move approach ko" with no fear of
misunderstanding. For an iterated ko, the size could be called
"N-iterated", "N-step iterated", or "N-stage iterated" with only some
chance of confusion. The "iterated" can't be left out, and to me, (at
variance with Bill Spight's excellent dictionary definitions) I'd
prefer "stage" to go with iterated kos, because that's what the usage
is in Davies' widely published and read books. The states of an
iterated ko could be called steps or stages, with my preference again
being to agree with Davies' books rather than Spight's dictionary. As
long as it's clear that it is an iterated ko being talked about, the
chance for confusion seems low.
Douglas Ridgway
Bill Spight
> > There is a word, "dan", that translates to "stage", but it's different
> > from the one in nidan ko, etc.
>
> And this is why you have a smiley above, right? Because it's not clear
> even to English speakers who know some Japanese what the "obvious
> rendering" is?
I have a smiley because I come from the South. We smile a lot. :-)
Actually, I think it is clear which word is being used in such go
expressions as "nidan ko" and "nidan bane". Besides, the character is
different.
> In any case, it seems like a bad idea to expect English
> readers to have to translate words back into Japanese in order to
> understand their meaning.
I agree. I did not intend to suggest that.
> Better to have well-defined English words,
> derived from the Japanese if necessary.
I think that it is a good idea, in general, to use go terms that
translate easily back and forth between languages. That makes for mutual
understanding. :-)
Best regards,
Bill
Simon Goss
>There is a word, "dan", that translates to "stage", but it's different
>from the one in nidan ko, etc. Besides, the English "stage", while
>indicating progression, seems less reversible to me than "step".
>Iterated kos are sometimes fought at different points during the same ko
>fight. Returning to stage 1 after stage 2 seems more awkward to me in
>English than returning to step 1. But taking a step backwards is normal.
>
>However, in such cases I find myself referring to the "leg" of the ko,
>like the leg of a trip. So maybe we could refer to iterated kos by the
>number of legs they have. OTOH, "4-leg ko", or worse, "4-legged ko"
>sounds funny, too. <shrug>
<and later>
I see. Thanks.
Well, both these kinds of ko have more than one ... call it what you
will ... step, stage, phase, leg, state, ... The big difference is the
reversibility or irreversibility, isn't it? Perhaps there isn't a single
English word that conveys to all of us the same sense about that. I
think now that it may be risking too much to describe any ko simply as
an N-<whatsit> ko, whatever we use to substitute the <whatsit>.
If we always take care to call a ko an N-<whatsit> <type> ko, then all
the semantics can go into <type> and the choice of <whatsit> isn't
really critical. Pro tem, I think I'm going to use "N-step iterative ko"
for that type, and just wait and see whether some standard usage emerges
some time in the future.
Best,
--
Simon
Charles Matthews
<snip>
> If we always take care to call a ko an N-<whatsit> <type> ko, then all
> the semantics can go into <type> and the choice of <whatsit> isn't
> really critical.
I quite like this direction of argument. I think it would be good to have a
systematic terminology, a jargon or working terminology in which <whatsit>
might flag something, and also a common terminology. This would be like the
organic chemistry systematics mentioned before in the thread, which is
cumbersome but accurate in its full majesty, is typically somewhat
abbreviated in technical discussion, and doesn't prevent one from just
saying "benzine".
At the moment I'm being influenced by the idea that from a fundamentalist
point of view there are 'levels' - an underlying partial order of them. On
a level one (the game) can visit various 'rooms' which we can euphoniously
call 'legs' in the ordinary case that we can get back out the door we came
in.
Charles
Matti Siivola
[...]
> To me, "step" and "stage" by themselves are inadequate, so I would
> always write "approach ko" or "iterated ko" for kos with irreversible
> and reversible states respectively. The word "approach" is
> established, natural in English, and is apparently also the natural
> translation of the Japanese term. "Iterate" is less well established,
> and isn't entirely obvious or natural (to me anyway), but at least is
> being used consistently. The different sizes of approach ko, as well
> as their states, can be called "N-move approach ko" with no fear of
> misunderstanding. For an iterated ko, the size could be called
> "N-iterated", "N-step iterated", or "N-stage iterated" with only some
> chance of confusion. The "iterated" can't be left out, and to me, (at
> variance with Bill Spight's excellent dictionary definitions) I'd
> prefer "stage" to go with iterated kos, because that's what the usage
> is in Davies' widely published and read books. The states of an
> iterated ko could be called steps or stages, with my preference again
> being to agree with Davies' books rather than Spight's dictionary. As
> long as it's clear that it is an iterated ko being talked about, the
> chance for confusion seems low.
I agree with you and would also use "N-stage iterated". It is better to
add
new word than swith the meaning of old ones.
Matti Siivola