BioWidget meeting

Georg Fuellen fuellen@dali.Mathematik.Uni-Bielefeld.DE
Mon, 5 May 1997 08:58:51 +0000 (GMT)


SteveC wrote,
> > I recieved a message from Alan Robinson, the organizer of the EBI 
> > BioWidget meeting, asking if I would be willing to give a to give a talk 
> > on Bioperl. I told him yes and that it might be tag-team style talk. I'm 
> > perfectly willing to split the talk with any of you who will be there. If 
> > I'm the only speaker, I could touch on the Bioperl work you all have done 
> > in addition to talking about my own work.

SteveB wrote,
> Good.  He told me a couple days ago that he'd penciled in 30 minutes which
> could be broken into 2 x 15 minutes.
> 
> > We need to submit a title ASAP and an abstract by 18 May. Since 
> > this is going to be a community effort, how does this sound for a title:
> > "Bioperl: Object-Oriented Perl Modules for Bioinformatics". It's general 

Title is ok!!

> > since the content of the talk is still not nailed down. I'll try to come 
> > up with an initial outline/abstract soon. If any of you have specific 
> > suggestions for the plan of the talk, i.e., what should definitely be 
> > included/excluded, let me know.  
> 
> Title sounds good.  The contents will depend on who else can make it --
> parts (3 and 4) below will be longer or shorter depending on who can
> speak.  I would propose that the following main topics be covered: 
> 
> 1) What's Perl, why's it good for biology.  Why not all Java?
>    Georg has written a lot of good stuff about this.  We probably do 

See http://www.techfak.uni-bielefeld.de/bcd/Perl/Bio/Docs/whatsup.html
Regarding Java, I've appended some wild thoughts at the end, since this
is not in any way my area of expertise.

>    want to explictly mention advantages of Perl over Java for some tasks.
>    
> 2) What's the Bioperl project trying to do?
>    A) First, core modules which can be a standard currency for which
>       everyone can write their functions to operate on.  This means that
>       different people's functions can be linked together without always
>       writing conversion routines.  They also allow for complex yet
>       maintainable programs
>    B) Later, lots of functions to operate on these modules
> 
> 3) What's been written, explicitly and what's planned
>    A) Bio::PreSeq   
>       Discuss basic functions/interface.  Modification to be included 
>       in Bio::Seq.  Example of usage to do something really cool 
>       very easily.  Bio::SeqBig (?) - superset of Bio::Seq to support
>       all sorts of NCBI annotation, etc.
>    B) Bio::UnixAln
>       Discuss basic functions/interface.  Example of usage to do something
>       really cool really easily
>    C) Bio::Struct
>       ...still on the drawing board.  Volunteers wanted!
> 
> 4) SteveC's own code and perspectives.

best wishes,
georg

=====
Re Java, I hope that it will enable us one day to have a GUI-like 
interface by way of the WWW-browser, but this opens a whole can of worms.
Last year I compiled the list of ``Java drawbacks'' that you can find below,
but I'd really suggest to say that Bioperl can win big from an interface
to Biojava stuff, and vice-versa this may be true well. I dunno whether 
a generic object interface could help; last I heard some study saying
that simple concepts like Javabeans may outcompete COBRA really soon.

A prelimininary, biased, outdated, and ill-informed list of Java ``problems''...
  <li> Java is less high-level, more restrictive, less efficient
  <li> Java is strongly typed, but lacks templates
  <li> Java links "too late", and collects garbage "too late"
  <li> Java applet security problems are ``inherent'' in the design of the
       language (but Perl is "catching up".. ;-)  (((Hint at Penguin)))
  <li> Java is not "Free Software"
  <li> Java is too much "cutting edge"
  <li> "Inferno"/"Limbo" is coming up anyway :-)
  <li> Java is nice for designing user interfaces, while Perl
       acts behind the scenes...