13 Feb 2002 10:04:16 +0000
>>>>> "Jolyon" == Jolyon Holdstock <firstname.lastname@example.org> writes:
Jolyon> Hi, Should SequencePoster still be in the API (at
Jolyon> biojava.org) as SequencePanel seems identical? Cheers,
I guess that SequencePoster could go as it seems to have the same
functionality as SequencePanel, just a slightly different
implementation. In fact, the possibility of TranslatedSequencePanel
being the the default SequencePanel was raised a while back.
What we have now are:
SequencePoster - older implementation
SequencePanel - main implementation containing a drawing workaround
which allows high pixel coordinates to be drawn accurately by
translating them towards zero, but which gives unexpected results when
making a BufferedImage from the panel. Nice and easy to use.
TranslatedSequencePanel - supplementary implementation which uses a
different drawing workaround which avoids use of high pixel
coordinates altogether and doesn't suffer the BufferedImage offset
problem, but which is not a drop-in replacement (if you drop it in it
will work, but the high pixel coordinates problem remains - you need
to use its extra translation methods to pan the view to high
sequence/low pixel coords). However, performance is much (6-8x)
improved and it generally feels more snappy.
I think SequencePoster should go. Not sure about the others - I think
there's room for both if the differences are clearly documented.
-= Keith James - email@example.com - http://www.sanger.ac.uk/Users/kdj =-
Pathogen Sequencing Unit, Wellcome Trust Sanger Institute, Cambridge, UK