In the particular case of Sequencher software,
let me comment on a few of the reasons Tim Cutts suggested for a Unix version.
>> Value for money is one. ... With one computer and
>> software licence, we provide sequence analysis facilities for over
>> 1,500 users, which is pretty inexpensive per user
Sequencher is a commercial product that is sold on the basis
that each active process of the program at any one time has to
be a licensed process. You pay for 5 licenses and 5 users can run
the program at once. This is the same for many Unix commercial
products. So there is equal value here between Mac and Unix system.
At IU we have Sequencher on a Mac file server where any of the over
1,500 Mac computers on campus can use it at any time, subject to
the limit on the number of licenses we own.
>> Their machine does not have to sit there being unuseable for anything
>> else because its chewing sequence databases
Sequencher works well as a multi-tasking program, I believe. There
would indeed be value in having Sequencher work on a remote, central
computer though, at the user's choice. It is possible to do this
with Mac systems, but I haven't seen much of it. Clearly this is
a major value of Unix.
One programming paradigm that I'm employing
now in my software is to provide client-server (TCP/IP) functions that are
transparent to the user, so one can use sequence editor on a desktop
and do the crunching on a remote computer, without much concern
to the user about which computer is doing what. This is the BOP protocol
that now appears in SeqPup software, and which Howard Cash is welcome
to add into Sequencher. I think Sequencher could be a much better
product if it added client-server functions to run analyses on
remote systems.
>>The real challenge is making a text based interface that is still
>>friendly.
For complex software, this is a challenge that can no longer be
met. It takes a lot of programming effort to build software that
biologists will find useful, easy to learn and use in a continuing way.
There are too many GUI elements that scientists require now from good
software, that cannot be mimicked in text interfaces. It doesn't
matter now whether your platform is Xwindows, Mac, or MSWindows,
don't expect complex software to fit into a text-mode interface.
My original question regarding Sequencher (Mac), and
Phrap/Phred/Consed and Staden packages (Unix) was aimed at finding
the best program, regardless of platform. There is value in central (Unix)
systems that are multi user, better at multi tasking, and often
higher performance (at at higher cost of course). There is value
in desktop systems (Mac and Wintel) that are managable and usable by
biologists w/o special training, that work in an easy to use manner
and do all the important things a biologist needs.
- Don
--
-- d.gilbert--biocomputing--indiana u--bloomington--gilbertd at bio.indiana.edu
