On Sun, Apr 21, 2013 at 10:18:04AM +0200, Alexander Wagner wrote: > Sounds like it should be changed only if one is going for some si5. > > Just out of shear curiosity: is 16.7e06 indeed a "real world" limit? > I mean, are there larger databases of /sensible/ games? AFAIK even > cb online is only at something about 6-7e06 and they already cover > quite a lot of lower end games.
Actually I found the limit exists while mangling with my databases and trying to merge them to a single database. So yes, this is a very real world problem. Of course, there are a lot of duplicates in those databases (I collected them from different locations in the web) but to sort out duplicates I still need to be able to merge them. There are a lot of sensible games in those databases a friend of mine, who is using CB megabase, is missing, so CB is far from being complete (i.e. it's not an ideal point of orientation for size) However, with the spreading of automatic notations boards, more and more games played will be published on the web. Sooner or later - I guess sooner than most expect - the limit will be a problem even after dupecheck. For now I end up in splitting my databases by eco-codes, merging the splits for each eco-section, doing the duplicate check on those and merge the results again to a single database. 5 times more work than without the limit, i.e. I consider this as a workaround not a solution. Therefore I would vote for SI5 and using 32bit integer instead 24bit :) ------------------------------------------------------------------------------ Precog is a next-generation analytics platform capable of advanced analytics on semi-structured data. The platform includes APIs for building apps and a phenomenal toolset for data science. Developers can use our toolset for easy data analysis & visualization. Get a free account! http://www2.precog.com/precogplatform/slashdotnewsletter _______________________________________________ Scid-users mailing list Scid-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/scid-users
