In article <m0v8ayk-00078qC at server.branch.com>,
Howard Cash <howardc at GENECODES.COM> wrote:
>Is there an assumption here that UNIX-based tools have an inherent value
>that is not reflected in an unstructured polling of newsgroup readers?
>Without challenging that premise, what do people think is the main value
>of having sequence assembly software on a UNIX platform? If Sequencher
>is already such a favorite on the Mac, is there something to be gained by
>going to the trouble to develope a UNIX version?
>>As an engineer, I like UNIX for software development, but I am biased
>towards the Macintosh for delivering ready-to-use analysis tools. Larger
>research groups with existing UNIX-based data-management systems ask us
>about porting to that platform, but is there a compelling reason for us
>to develop Sequencher for Unix? What would the advantage be in the
>average lab, other than the undeniable freedom to work in front of the
>computer of one's choosing?
Value for money is one. I am part of a team running a sequence
analysis service for Cambridge University. With one computer and
software licence, we provide sequence analysis facilities for over
1,500 users, which is pretty inexpensive per user, even for a large
UNIX box, and the end user can then buy a cheaper terminal and still
have high speed searches.
Their machine does not have to sit there being unuseable for anything
else because its chewing sequence databases, therefore allowing them
to get on with other stuff on their Macs and PCs, like writing papers.
Then there is the undeniable advantage UNIX has of stability. Macs
and PCs (with the possible exception of PCs running Windows NT or
UNIX) are not very stable, and I'd rather go home leaving my sequence
crunching work overnight without worrying that the machine might crash.
Plus a UNIX box automatically throws in a large number of other
benefits, such as user file security, electronic mail, convenient
submission of batch jobs, and so on.
Admittedly some of these are available on other platforms
(particularly Windows NT, which has both job sheduling and mail) but
as far as I am aware there is as yet no large sequencing package
specifically designed with Windows NT in mind.
UNIX really loses out on the user interface. I totally fail to
understand this. X-server software is now cheap for PCs and Macs.
UNIX is a very suitable operating system for this kind of work, but
lac of attention to user friendliness by most UNIX software developers
is its main detraction. Admittedly there have been very good
attempts, such as the Motif CDE and OpenLook, but they have never
reached the polish and ease of use of the Mac or even Windows (please
no flame wars on Windows vs. MacOS, this isn't the point).
If you are going to port to UNIX, make the X user interface good, and
people will go for it. Unfortunately this is difficult when Motif is
still a proprietary product, although most of your target market will
be running mainstream OS's like IRIX, Solaris or Digital UNIX, all of
which support Motif these days.
The real challenge is making a text based interface that is still
friendly. This is extremely important in UNIX because nothing causes
a UNIX machine to grind to a halt faster than 30 simultaneous
X-windows sessions!
Just a few thoughts.
Tim.
--
--------------------------------------------------------------------------
Dept. of Biochemistry Tel: +44 1223 333596
Tennis Court Rd.
Cambridge, CB2 1QW, UK
