Author: Robert Hyatt
Date: 20:25:35 10/19/00
Go up one level in this thread
On October 19, 2000 at 12:40:58, Ricardo Gibert wrote: >On October 18, 2000 at 22:35:43, Robert Hyatt wrote: > >>On October 18, 2000 at 16:22:36, Ricardo Gibert wrote: >> >>>On October 18, 2000 at 14:11:03, Jeremiah Penery wrote: >>> >>>>On October 18, 2000 at 14:07:24, Ricardo Gibert wrote: >>>> >>>>>My conclusion from your list is the PGN standard is poorly designed. It lacks >>>>>flexibility and makes outright bad decisions. It is obvious that some of the >>>>>reasons for the variations are that there are *good* reasons to vary. >>>> >>>>I'm not sure what you mean by this. What is a good example of a reason to vary >>>>some of these things? >>> >>>#5 is motivated by the desire to work with the word wrap feature of many >>>editors. Many people do not like to have the move number separated by a line >>>from the move itself since they can be misled into thinking it is a move by >>>Black. Omitting the space is an effective solution for this. It is perhaps a >>>matter of taste, but there is no good reason not to accomodate this alternative >>>form. Also, omitting the space is more compact. >>> >>>An example: >>> >>>1.e4 e5 2.Nf3 Nc6 3.Bc4 Nd4 4.Nxe5 | >>>Qg5 5.Nxf7 Qxg2 6.Rf1 Qxe4 7.Be2 | >>>Nf3# | >>> >>>1. e4 e5 2. Nf3 Nc6 3. Bc4 Nd4 4. | >>>Nxe5 Qg5 5. Nxf7 Qxg2 6. Rf1 Qxe4 | >>>7. Be2 Nf3# | >>> >>>Note, the difference with the move 4.Nxe5. I prefer the first version, which >>>avoids the separation of the move number from the move itself when the word-wrap >>>feature is employed. >>> >>>Bruce Moreland listed some good suggestions in his post. >>> >>>"Inflexibility" is *my* primary gripe. See #13. >> >> >>I totally disagree here. The purpose of a "standard" is _not_ "flexibility". >>It is preciseness. IE the C pseudo-standard is lacking in many places. Is >>a char signed or unsigned? Left up to implementors. Do bit fields start from >>msb or lsb in a word? same answer. Hard to call something a "standard" when >>you can't measure anything against it, because _anything_ will pass it. > >I never suggested "The purpose of the "standard" was "flexibility".' Your >analogy to C makes no sense to me. There is no harm in including e8Q and e8(Q) >in addition to e8=Q for instance. > Why? One way of expressing promotion is far better than N ways. How about C allowing a statement to end with a ";" or a "." or a "!"?? Makes the parsing problem much tougher. SAN is _not_ difficult to produce. There is even some public-domain software to do this, in the Edwards SAN kit. I don't see any need to have to handle e4, e2e4, e2-e4, Pe2-e4 and so forth. Nor any of the other corruptions including ed: for captures. If everything that can be specified in one precise way is specified in one precise way, then interoperability is greatly enhanced. Without having to develop a lex/yacc grammer to parse chess moves. >> >>I despise 1.e4 moves. Why? a parsing issue. I use the output specs and >>have been happy. I try to parse all the idiot ways of saying "castle" but >>that ought not be a problem if everyone just 'read' the standard. > >The "parsing issue" is trivial and therefore should definitely not take a back >seat to the standards usabilitity. I agree. But I see _no_ valid reason to prefer 1.e4 over 1. e4. And since most would agree that 1. e4 is preferable when reading chess moves, I would suggest using that _one_ way of doing things to make everything homogeneous. > >The fact so many depart from the standard in so many says something about how >well it meets the target audience needs. Or it says something about how lazy someone can be. The tags are pretty precisely defined, grammar-wise. Including quote marks. I could do without the quotes, but they are in the spec and I produce them. I would bet that if everybody read the spec, they would be reasonably happy with following it. Most just follow "legend-PGN" which is that PGN spec that exists in someone's mind, with lots of unknowns that are filled in incorrectly. And then that is passed on to others with the expected permutations happening. If there is no ambiguity, most problems go away. If there are not multiple choices for a given item, everything is clearer and easier to handle overall.
This page took 0 seconds to execute
Last modified: Thu, 15 Apr 21 08:11:13 -0700
Current Computer Chess Club Forums at Talkchess. This site by Sean Mintz.