El dt 30 de 07 de 2013 a les 17:49 +0200, en/na Lars Aronsson va escriure: > On 07/30/2013 01:20 PM, Francis Tyers wrote: > > El dt 30 de 07 de 2013 a les 12:51 +0200, en/na Lars Aronsson va > > escriure: > >> Do started pairs grow after the initial setup? > > In our experience not much. We've found that typically after the pairs > > are released, they tend to stay static. There are of course some > > exceptions, e.g. Breton-French. > > I made a first attempt at trend statistics, just counting > the occurrences of '<l>' (left item in a dictionary pair) > in the lang-lang.dix translation files, over time through > the SVN history, for some language pairs in /trunk/. > This measurement is not accurate for how useful a > dictionary is, but gives some indication of whether the > size is the same or grows. The sudden initial growth > followed by a flat line is very typical, as you say, > http://wiki.apertium.org/wiki/User:LA2#Diary
Nice graph, and yes, pretty much what I expected. > Exceptions include br-fr, which grew steadily from > late 2008 until the end of 2010 (red curve), > and en-as from early 2010 until mid 2012. > > Perhaps it is natural that the initial setup, which > requires a rare combination of linguistic knowledge > and coding skills is a sudden movement. But to make > the continued growth open to a wider range of > volunteer contributors, perhaps it needs to be made > much easier. That might need a new toolset. This is true. But what that toolset might consist of is quite undefined. It's happened quite a few times that someone tries to make an interface, but fails. I failed in this very same task. > The improvement of existing language pairs (such as > those with less than 10,000 translation entries) could > be as important as setting up new pairs. The mk-en > pair was created in October 2010 with 10,000 entries, > and stayed at that level until it suddenly grew to > 32,000 entries in July 2012. Could we make that > happen to af-nl, ca-it, es-gl, mk-bg, sv-da, and kaz-tat? That would be a great idea. The work on expanding mk-en was done by Filip Petkovski as part of last year's GSOC project. The Kazakh-Tatar pair is being developed by Ilnar Salimzyanov who is planning to bring it to 95% coverage (at the moment it is around 90%). The other pairs would require new contributors I think. Fran ------------------------------------------------------------------------------ Get your SQL database under version control now! Version control is standard for application code, but databases havent caught up. So what steps can you take to put your SQL databases under version control? Why should you start doing it? Read more to find out. http://pubads.g.doubleclick.net/gampad/clk?id=49501711&iu=/4140/ostg.clktrk _______________________________________________ Apertium-stuff mailing list [email protected] https://lists.sourceforge.net/lists/listinfo/apertium-stuff
