[Biojava-l] FramedFeature files

[Mark Schreiber]

On Mon, 10 Sep 2001, Thomas Down wrote:

> One query, though, which I think ought to be discussed
> a little bit before this goes in for final: FramedFeature
> extends StrandedFeature, but it also includes forward/reverse
> strand information in the ReadingFrame, a duplication which
> worries me a little.  You recognize this problem in the
> template class (which doesn't have a strand field).  However,
> this introduces another (probably very minor) issue:  when
> realizing new features, the idea is that implemetations will
> search for the most specific feature implementation which matches
> the supplied template.  So if a sequence implementation has
> a StrandedFeature implemetation but not a FramedFeature implementation,
> right now you'll end up losing the strand information too, which
> seems kind-of unfortunate to me.
> 
> Do you have a specific reason why you don't want to handle
> strand and frame as separate properties, in much the same way
> GFF does?
> 

Irk ... this is getting thorny, from memory I didn't want people to be
able to specify a strand that would conflict with the frame, to me the
idea of frame also specifies strand but i guess it doesn't have to. I
liked the idea of getting the needed info in one go. Perhaps two method
calls will be needed or I could make a Frame class as a singleton to hold
the two int feilds, these could be specified as static constants in much
the same way as they are specified as ints at the moment.

Mark

> 
> (By the way, would you like a CVS read/write account so
> you can check this in yourself?  If so, just drop me
> a suggested user name and an SSH public key)
> 
>    Thomas.
> 

~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
Mark Schreiber			Ph: 64 3 4797875
Rm 218				email mark_s@sanger.otago.ac.nz
Department of Biochemistry	email m.schreiber@clear.net.nz
University of Otago		
PO Box 56
Dunedin
New Zealand
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~