Date: Mon, 14 Jul 1997 11:01:11 +0200 (MET DST) From: Arno Hollosi Subject: Re: Many Faces of Go and FF[4] support > MF silently discards and ignores, with no warning message: > VW, WS, BS, CI, DG, AP, LT, NW, NB. Ok. > MF converts all the following to english comments on input: > CH, GB, GW, TE, BM, DM, IT, DO, UC, HO, SI Ok. > The only double MF writes is TE[2], to indicate the correct move in a problem. > BTW, shouldn't IT and DO have property value Double, rather than none? No. A 'very doubtful' move is likely to be a bad move. And a 'very interesting move' is likely to be a good or very good move. > MF preserves, ignores, and prints a warning (unrecognized property) for: > AR*, BL*, CA*, DD*, EL, EX, FG*, ID, IP*, IY*, KO*, LN*, OB*, OM, OP, OV, OW*, > PM*, SE*, SU*, TC, WL* > * are FF[4] properties not supported by MF. SU, SE, IY and IP are game-specific properties of 'Lines of Action'. A Go program doesn't have to understand these. Why don't you support WL & BL? Some of users of Primiview (my application for Amiga) said they like to see how much time was left -- adds some exitement to the game :-) And supporting BL & WL is easy. > MF converts the following marks to triangles: > RG, SC, TB, TW Do you really mean 'convert' or just 'display as triangles'? I hope you mean the latter as TB & TW shouldn't be mixed. > MF converts L to LB > MF converts M to MA Fine :) > MF handles AB and AW at any node, even mixed with moves. It shows a > warning if moves and setup are mixed in nodes. It can only handle AE > at leaf nodes. What's a leaf node? The last node of a tree? Is MFGO able to read the following: >> (;FF[1];B[dd](;W[cc])(;AE[dd]B[de];W[cc])) << Because many old files do exactly that. People tried to show variations as siblings with viewers that displayed them as children. Mixing has been said to be bad style but it occured frequently. E.g. until 2 months ago, the Go Teaching Ladder had lots of files which mixed setup and move properties in a node. > If an input file has a move in the root node, MF adds a new node after the > root and moves the root move there. Ok. > MF can skip junk in front of the inital (; (like e-mail headers), and can handle > files starting with ( instead of (;. What kind of check do you use? (I guess this is interesting for us all :-) Are you able to read a file with missing ';' and e-mail header in front? In SGFC I look if there are three or more '[', ']' before the next '(' or ')' occurs. > MF expands compressed point lists on input, and will never write a compressed > point list. Ok. > On output, MF always uses [tt] for pass, and writes > GM[1]FF[4]VW[]AP[Many Faces of Go:10.0] at the start of the file. It always includes > HA, SZ, and ST properties in the root node. MF writes 2 upper case properties only, > but can read old style mixed upper and lower case properties. Good :) /Arno