Cool David! Have you got any stats about the relative performance of the raw and packed implementations? The issue with AlphabetIndex and ambiguities is my fault. I wrote the imlementations not to index ambiguities. What do you use an indexer for? I'm happy for you to commit away. Thomas? Others?
David Huen wrote: > I've written a packed version of a DNA symbol list (4 bits per symbol, 8 > times cheaper than the current). Surprisingly, the performance seems OK - > I expected it to be abysmal. Its constructor can take and return data in > a byte array so data can be serialised from/to a database. > > The drawback is that it uses its own AlphabetIndex because I could not > make the ones in the package take ambiguity tokens for whatever reason. > Most likely I am missing some thing here. My main concern is that if I > were to use this class in a persistent implementation and the Alphabet > Index for this class needed to be changed later, it would scramble > these objects so the necessity to get it right now. > > Any suggestions as to solutions before I consider committing it to > development trunk? > > Regards, > David Huen > > > _______________________________________________ > Biojava-l mailing list - [EMAIL PROTECTED] > http://biojava.org/mailman/listinfo/biojava-l > > _______________________________________________ Biojava-l mailing list - [EMAIL PROTECTED] http://biojava.org/mailman/listinfo/biojava-l
