[Biojava-dev] bioJava, jsci

[David Huen]

On Tuesday 07 Oct 2003 9:24 pm, Silvère Martin-Michiellot wrote:
> Hi,
>
> We are a group of people building the JSci package at
> jsci.sourceforge.net. We are looking to extend/implement biology
> features. We need support for biologists to help us extend the features
> of the API as we are not experts by ourselves.
>
> Thus far we have implemented a basic package with aminoacids, DNA, bases,
> mRNA, proteins. We would like to receive suggestions, ideas about
> possible extensions, API review, etc.
> It seems for us we have to work in the same direction as you do to
> minimalize coding effort.
> We are also thinking about implementing XML support but from
> http://www.pasteur.fr/cgi-bin/biology/bnb_s.pl?english=1&query=xml it
> appears we are a bit struck to what is cool, working, standard and what
> is not (see for example the comparison from
> http://www.cse.ucsc.edu/~douglas/proximl/proximl.html).
> We are finally wondering how to organise our packages to help users who
> would like to use BioJava along with JSci.
>
> Let us know.

I am replying on my own behalf as a BJ developer and nothing I say should be 
read as representative of the BJ development community.

BioJava is still a relatively young project and the developer resources we 
have at hand comes solely from volunteers and tends to be focused on the 
immediate needs of those people.  There is a general model for much of the 
code and development is primarily based on extending and modifying taht 
model.  Because of the limited resources, we tend to remain fairly focused 
on developing methods of representing and computing on biology, 
particularly in the representation of DNA sequence.  Our volunteers 
generally have day jobs that keep them busy and have enough difficulties 
finding enough time to code the needful before even considering possible 
collaborative ventures.

I have attempted to look at your site and if I surmise it correctly, you 
have within JSci various math algorithms and an implementation of MATHML.  
I do not have time to look into it much further than that.    Perhaps you 
could expound on JSci's objective more extensively.  You mention that you 
have also implemented a molecular biology package: I don't know what your 
APIs are but if they are fairly different from ours (and ours can be 
initially surprising!  Sequences are not a sequence of letters...) it will 
be unlikely that we will want to break our interfaces to support good 
interoperability with what JSci APIs.  Our code is very tightly bound to a 
specific model and changes in it will break functionality greatly.  The 
major area we attempt to maintain some degree of interoperability with are 
the other OBF projects (BioPerl, OBDA, etc).  perhaps you might like to 
point out areas that you think might be mutually explored at this stage - 
they don't seem immediately apparent to me at this stage.

Regards,
David Huen

-- 

David Huen, Ph.D.              Email: smh1008 at cus.cam.ac.uk
Dept. of Genetics              Fax  : +44 1223 333992
University of Cambridge        Phone: +44 1223 333982/766748
Cambridge, CB2 3UH
U.K.