[Pipet Devel] yum, canned locus (was: from Justin)

Humberto Ortiz Zuazaga hortiz at neurobio.upr.clu.edu
Thu Jun 17 13:47:53 EDT 1999


> Humberto Ortiz Zuazaga wrote:
> > 
> > Boulderio is a much simpler solution to the same problem we're looking at

> Do you
> mean simpler in design but less functional, or simpler in design with the same
> function?  If it's the latter, why not take the same approach?

Simpler in design, I guess it has about the same functionality, but the data 
stream is uglier than XML.

Here's an example data record

NAME=AFM036YB2
SEQUENCE=GATCCGAGATACCCCACACACACACACAC....
STS={
        LEFT_START=0
        LEFT_LENGTH=14
        RIGHT_START=140
        RIGHT_LENGTH=20
}
=

When read, this results in a perl object called a Stone, with tag, value 
pairs, sort of like a dict.

If the stone is in $s, then $s-index(NAME) is AFM036YB2, and $s->
index('STS[0]') is another stone contining the STS data.

I've used boulder to do some work on genetic mapping, and found it to be a 
little slow, but I'm not real good at OO perl, and may have been doing 
something dumb. The recent discussion on DOM makes me think it's a problem 
with the document-as-object model, however.

> I think that we need BOTH design-time AND run-time checks.  Are you saying that
> there should be no design-time check or that nearly all of the design should be
> a run-time decision by the locus?

No, you're right. But the design time checks and the run time checks can be 
the same. At design time, failure of a check should be fatal, at run time, the 
locid should try to locate any missing components.

> > But that's the point, everyone is going to be constructing a command line,
> 
> IMHO, we have to expect that bench scientists are no more computer-savvy than
> the general public.  They will expect Loci to operate, on the surface, like any
> other GUI application.  They'll want to choose File->Open, then select a
> sequence or whatever, and then push a button and get pretty pictures (well,
> something useful anyway :-)

So you think users won't want to do stuff like, "lets open this file, do this, 
then select all frobs, oh, wait, now do that instead, and give me the green 
frobs only"? I belive this kind of interactive, incremental analysis is the 
best feature of loci.

Should users only get canned loci, with pushbutton doodads for setting 
options, and not get the entire loci toolkit, including the plumbing 
developers use for fitting loci together?
-- 
Humberto Ortiz Zuazaga
Bioinformatics Specialist
Institute of Neurobiology
hortiz at neurobio.upr.clu.edu






More information about the Pipet-Devel mailing list