Two of my values are:
User interface development is a fundamental part of biocomputing
and bioinformatics sciences.
User interface development is appropriate biocomputing research
material for government funding.
Dave Kristofferson wrote:
>1) User interface software is *not* typically thought of as research
>in molecular biology.
I think we are not dealing with the field of molecular biology,
but with the field(s) of biocomputing & bioinformatics, when
discussing the work of NCBI. User interface software is an
important area of research in biocomputing & bioinformatics.
The current standard in computing software is a graphic user
interface. Software with just a command line interface (e.g.,
the unix BLAST or FastA code), is not useable by the majority of
biologists at Indiana University who need to do sequence searching.
In various fields of scientific computing, NSF has made a number
of awards for user interface work. See below for examples of NSF
funded user-interface development.
and Dave K. wrote:
> 2) General user interface toolkit development is NOT a fundamental
> requirement of doing DNA sequence database work.
I disagree. A good user interface software library, which
can be used on the common computing platforms (mac, mswindows,
xwindows), is a basic building block for getting the data to
users. The scope of user interface work, just as the scope of
their database work (should they include the patent sequences
in their distribution?), should be open to discussion or oversite,
but the need for user interface development is fundamental
to providing usable sequence databases to the public.
A simple but useable human interface to the BLAST or FastA
command-line software is provided by my "Genbank Search"
Hypercard stack working with the e-mail Fasta & BLAST services
of GOS, or by Mike Cherry's HyperBLAST working with NCBI's BLAST
server. From what I saw a while back of Jonathan Kans' VIBRANT
graphic user interface (gui) toolkit, it is primarily a basis
for the Entrez software, to give it a platform-independent gui
so that Entrez and related database access or research
software (e.g., BLAST, MACAW) from NCBI can reach the average
Indiana University biologist (I won't speak for other
universities, perhaps there are some who prefer raw command line
software).
A CDROM full of gene sequences without software to access it is
a step backward for sequence database access. If you are going
to write software today that you expect others to use,
especially computer illiterate biologists, you must supply a
good user interface with that software. If NCBI left all the
user-interface writing to market place forces, much of their
work on organizing the data would go for naught. Software is an
essential part of today's databases -- most especially the large
ones. Software needs to be tailored to databases, needs to be
updated as the databases change, needs to be available to read
the databases. NCBI needs to control the production of software
to go with its data.
-- Don
P.S., I very much appreciate Dave Kristofferson's willingness to be
an advocate for the biotech software industry, even if it is here an
unpopular position, to help shed light on this controversy over what
areas NCBI should work on.
I also applaud Intelligenetics for being a pioneer in user interface
development in biocomputing. They had x-window sequence software
available very early on, and the GeneWorks package [Manual Glynias,
primary author] I would rate as innovative and leading-edge both in
user-interface and integration of algorithms.
Here is a sample of NSF awards for user-interface research that
I picked up from the NSF Gopher:
Title : SGER Grant: Initial Development of a User Interface
for Scientific Information Retrieval
Award Number: 9019391
---
Title : Advanced User Interface for Network Access to Multiple
Database for Biomedical Information Applications
Award Number: 8916178
---
Title : Scientific User Interfaces for Parallel Computing
Award Number: 8814937
---
Title : User Interface Designers Aide
Award Number: 8813179
---
--
Don Gilbert gilbert at bio.indiana.edu
biocomputing office, biology dept., indiana univ., bloomington, in 47405
