I'm just playing with jndi. We have recurrent serialization issues with alphabets, especialy those that aren't 'WellKnown' instances. One way out of this is to pass alphabet identities arround using some naming-and-directory service. Of course, the usual tricks with making RMI think that unique coppies of alphabets exist in multiple VMs should still all work (and then we should /realy/ be propogating novel alphabets using jndi using the same mechanism .. ugg).
Has anybody burned their fingers badly with jndi, or does it work out of the box? I'm just working through the sun tutorial at the moment. We will probably want to be naming friendly for a wide range of BioJava resources (sequences, and dbs to name but two), and jndi seems to have a clean abstraction between the api and the implementation. All comments welcome Matthew _______________________________________________ Biojava-l mailing list - [EMAIL PROTECTED] http://biojava.org/mailman/listinfo/biojava-l
