> I would caution against not correcting Le Quang Liem >> Le, Quang Liem,
Maybe it *is* best to go along with Twic conventions. But Franz's 2015 spelling.ssp does indeed make this correction anyway. BTW, i have patched his 2015 spelling and it is now in our subversion as spelling.ssp. Also in subversion, is updated/correct handling of FIDE's player rating data (contained in Franz's ratings.ssp) which is used by ScidvsPC's maintenance->AddPlayerRatings feature. This hadn't been working *accurately* for data newer than 2012, when FIDE started releasing ratings more often than bimonthly. S.A. On Sat, Jun 13, 2015 at 1:20 AM, Matthew Larson < matthewwendelllar...@yahoo.com> wrote: > I would caution against not correcting Le Quang Liem >> Le, Quang Liem, > even if it objectively correct. It seems as though, even if incorrect, > Ding, Liren, for example, is the convention in the chess world. For > example, TWIC pgns use this style. For this reason, if spellcheck is going > to consider Le Quang Liem correct, then it would be important to change Le, > Quang Liem >> Le Quang Liem, at least to make delete twin games work > correctly. I’m not sure how one would determine if a name should be written > in the Chinese style or the western style. Checking if federation is CHN or > VIE (or maybe other countries) wouldn’t work, as there are people with > western-style names in China and people with Chinese-style names in other > countries. It also wouldn’t work to check if the last name is among the > 10,000 most common Chinese surnames, as there are people like the american > player Joshua Mu, where Mu, Joshua is better than Mu Joshua. I think it is > better to leave this as it is. > Thanks, > Matthew Larson > > > > On Jun 12, 2015, at 2:34 AM, Steve A <stevena...@gmail.com> wrote: > > >>> I am interested in manually improving the spelling.ssp file. I created > my reference database by importing every pgn I could find into one base, > and then using spellcheck and delete twin games to get rid of duplicates. I > have noticed that spellcheck does not correct some East Asian names (e.g. > Le Quang Liem does not go to Le, Quang Liem) and hyphenated or names with > spaces in them (e.g. DeFirmian, Nick does not got to De Firmian, Nick). > Although I wouldn’t be able to find everything like this, I could correct > the spelling.ssp file when I run into something like this (by finding a > duplicate game not deleted because the names are different). > > Yah - improving our spelling correction is on the todo list. But i'm not > sure if the spelling file is too bad. Franz has just released a new version > of it (which i patch a little for release with ScidvsPC). > https://sourceforge.net/projects/scid/files/Player%20Data/Latest%20data/ > > Re not adding a comma to Asian names, i found this on wikipedia > http://en.wikipedia.org/wiki/Chinese_name > "According to the Chicago Manual of Style, Chinese names are indexed by > the family name with *no inversion and no comma*" > I am not familiar if (for eg) Vietnamese names follow a similiar > convention. > > Do you/anyone have any other issues with spelling.ssp ? > > We *do* need a list of improvements to be done to improve our spell > checker (I probably would leave chinese names without a comma). Things that > come to mind are names with a comma but no space, and name capitalisation. > > > I think I have discovered what seems to be a more important issue. It > seems as though ScidvsMac is not able to correct more than 2000 names at a > time. So, when I run spellcheck player names on my reference database, it > detects 260,000 corrections, but when I hit Make corrections it only > corrects 2000 times. > > Yes - this need investigating/addressing. I was not aware of it. > > > The spellcheck feature works beautifully, except that it doesn't remove > (wh) and (bl) at the end of the player name, which could be fixed easily by > adding a %Suffix " (wh)" "" line. > > Sounds reasonable, though i have never seen these used myself. > > Steven > > >
------------------------------------------------------------------------------
_______________________________________________ Scidvspc-users mailing list Scidvspc-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/scidvspc-users
