[Biojava-l] BioJava 1.2 development

[Keith James]

>>>>> "Thomas" == Thomas Down <td2@sanger.ac.uk> writes:

[...]

    Thomas> However, there's also been a lot of new code gone into the
    Thomas> project since 1.20 -- especially some hackathon-related
    Thomas> work.  I've therefore been wondering if we should look at
    Thomas> relaxing the previous policy of `no new features on the
    Thomas> branch'.  Instead, we could perhaps allow code developed
    Thomas> on the trunk to be moved into future 1.2x releases, so
    Thomas> long as:

    Thomas>   - It's self-contained

    Thomas>   - No existing interfaces or contracts are changed

    Thomas>   - It's likely to be widely interesting/useful.

    Thomas> Does this make any sense?  Would it make using `release'
    Thomas> versions of BioJava more, or less, attractive?  I'm rather
    Thomas> concious of the fact that after the 1.0 and 1.1 releases,
    Thomas> the branches went very nearly dead -- which makes me think
    Thomas> that more regular releases, with controlled addition of
    Thomas> new features, could make a lot more sense than the
    Thomas> previous policy.

I think it makes good sense, but I would add one more condition to
your list.

     - It has at enough Javadoc to find out a) what its general
       purpose is and b) who wrote it

I realise that a lot of this went in during a phase of intensive
hacking, but to draw on some Perl wisdom, having no docs means we
can't be as lazy as we need to be, nor as impatient and the authors of
all that beautiful code don't get to exhibit the hubris they deserve
;)

I'm prepared to help, by the way.

Keith

-- 

-= Keith James - kdj@sanger.ac.uk - http://www.sanger.ac.uk/Users/kdj =-
Pathogen Sequencing Unit, Wellcome Trust Sanger Institute, Cambridge, UK