|
@Marc Benford:
I would suggest you adopt your solutions #2 and #4 as a pair; except also go a small bit further...
____________________________________
Context Dependence Rots: Consider XSAN
All chess notations represent trade-offs, benefits and drawbacks.
The SAN used in regular chess is good in some ways, but SAN's drawbacks are substantial and not merely minor. During tournaments, maybe 30% of amateur players have an error in the scoresheet they submit to the Organizer. Often the error can be corrected, but in 20% of these cases error cannot be corrected.
SAN's proneness to errors extends into replay while reading moves from a book. Just one example is the reader moving the wrong color pawn when replaying the move notation 'b4'.
SAN is error prone because SAN suffers from its need for Context for accurate interpretation. Neither you nor I know whether 'Nxe4' is a White versus Black knight: to know we must have further contextual information, like whose turn it is, or the Origin square (instead of just the Destination square).
SAN really failed me when I started to replay gamescores from chess960 games (FischerRandom, chess9LX, or the dubious moniker "Free Style", whatever name). Turns out our habits for quick interpretation or translation of chess moves are marinated in the context of the one traditional start setup position.
Therefore, I feel that chess deserves to have the option of a notation which is "Context Independent", to the extreme of being fully replayable in reverse, from the end position to the starting setup position.
An enhanced LAN notation feels like a good candidate initially. 'b3-b4' is definitely a White move, and we know it was the pawn's second move.
'Nf6xe4' is more informative than 'Nxe4'. 'Nf6xe4' is probably a Black knight, but this notation still lacks certainty. Plus we have no idea which type of piece was captured.
'13... Nf6xBe4' tells us the knight's color was Black (the two extra dots ".."), and that the captured piece was a white bishop ("xB"). However, when you replay a whole game this way, you realize that you miss having the Destination square given early in the notated move. Yet 'NxBe4-f6', or 'Nxf6-e4xB', are both awkward. So,...
Let us try '<' an alternative to the '-' for captures, and give yourself one minute to get used to it. Further, let us adopt the cleaner ':' instead of yet another lowercase letter 'x' (as David Bronstein did in 'Zurich 1953), and tag a '/' onto the end:
[A] 'N:e4<f6/B' - "Knight takes on square e4, coming from square f6, captures a white bishop".
Now the more important Destination square occurs before the Origin square.
This notation comes close to being the best Context Independent notation that is still human-friendly (unless someone can post a better design?).
I like how the four sorta-separate sections of the notation are well-delimited for the human eye to parse:
N e4 f6 B ( : < / )
But [A] fails to tell us the color of the knight or bishop. Relying on a verbose move number plus triple-dots is a context dependency, in addition to being verbose in many situations.
So, let's try a solution which uses Uppercase versus lowercase letters more strategically: all white pieces use Uppercase letters, while all black pieces use lowercase letters.
This casing strategy is slightly helped by our previous replacement of the lowercase 'x' by ':'
[B] 'n:e4<f6/B' - "Black knight takes on e4, from f6, captures white bishop."
Good, now we are there except for one remaining problem: 'b' for a black bishop could confuse the eye with the 'b' column:
'bb2<d4', compared to say the clear 'qb2<d4'.
I have no perfect answer to the 'b'=black_bishop problem, but I can offer the best available solution that I chose: I use 'V' & 'v', instead of 'B' & 'b'. The idea is that, in English, a "vicar" is a "substitute for a bishop" in the Catholic church, in the majority of dictionaries. (Pronounced "VIK-err".) So we tweak 'bb2<d4' to instead be 'vb2<d4'.
[C] 'Re5<1' where we infer the column'e' because no override is present, is better than the unnecessarily redundant 'Re5<e1'.
Seems a shame that chess notation evolved to give sequence numbers to move-pairs (pairs-of-moves) instead of to true 'moves'. All white moves should be odd-numbered, and where occasionally needed, all black moves would be even-numbered.
XSAN:
I call the notation described above 'XSAN' ("eXstended Standard Algebraic Notation").
_______________________
Conclusion:
'Context Independent' notation should be available (XSAN?)
A simple software program can easily validate XSAN, and translate XSAN into SAN, or into LAN etc.
However, much more complicated software is needed to validate SAN, and to translate SAN into any other notation.
XSAN can be a central hub for all chess notations (P-K4 Descriptive, SAN, LAN, Coordinate, etc).
Instead of needing one software to translate SAN into LAN, and another to translate SAN into Coordinate or Descriptive, we would need only a software to translate SAN into XSAN.
For the creative piece types in your chess variants, you could use a two-letter pair for each piece. Later, when you want to use software that can only digest piece names being single letters, the XSAN hub could be valuable.
Bottom line, chess deserves to have a "Context Independent" notation option.
Thanks.
GeneM
.
|
Pages: 1
How should we use algebraic notation in variants? (Read 382 times)
