[A belated response to a thread from some days ago]
There are two main issues here.
The one which most immediately affects non-Linux users who might be
temptable to switch to linux is of course the comparison between
Linux and Windows where superficial criteria of use and performance
are concerned: speed, familiarity, resources like drag-&-drop, use
of "Find", ... This aspect will influence how strongly people feel
tempted in the first place, before they have experience of using Linux.
The one which will really affect users in the long term is getting
their work done.
For that, the software which does what they want, and does it at
least as well and as quickly and easily as the software they already
use, will be the crucial factor. Linux, as such, is "infrastructure".
The applications are not Linux, even if they run on Linux and, indeed,
even if they have been written for or can be compiled for Linux.
People like John Pettigrew, who is 'a self-employed editor, using my
computer heavily for web, MSWord, email and many other things that aren't
"a predictable set"', while their functional needs are determined by
what their work requires, can nonetheless choose what they use to do
the job so long as the job gets done. They may even be in a position
to decide to some extent what their work will require, i.e. tailor
the work they do to what their tools permit.
Others will have their requirements set by the company environment
they work in. In this case, either there are Linux-compatible tools
for the job, or not. If there are, then well and good. If not, then
they have to look elsewhere. This is a common situation, and in practice
many people mix-&-match between what they can do with Linux and what they
have to turn to Windows for.
Personally I was fortunate in my early (1992+) experiences with Linux,
since the nature of the work (scientific, technical, mathematical,...)
lay within areas where Unix itself had already been much used for many
years. As a result there was a lot of suitable software, originally
developed on Unix, to do the job (including troff, now groff, for writing
up such stuff). When not already ported to Linux, I could port it myself.
A mixture of 'octave' (matlab "clone") for numerical computation,
'gnuplot' and 'plotmtv' for graphics, and 'groff' for the writing, just
about met almost all my needs. On occasion I wrote my own C code for
particular jobs. The major exception was a database program: for this,
however, I ran dBase-IV under 'dosemu' and it worked fine. A spreadsheet
would might been handy at times, but I have never liked those things
and so was quite cheerful about not having one (and, if someone sends
you a CSV spreadsheet file you can enjoy using say 'awk' to extract the
data and 'vi' to tweak its format to make an 'octave' data file, and vice
versa to re-construct a CSV file to send back).
But there were other areas where I would have liked to use Linux but
didn't have the resources. CAD, DTP, OCR, ... . As it happened I didn't
actually _need_ to do this things; I just felt I would lke to get into
them. But I had faith that Linux would sooner or later acquire them.
Now, 10 years later, this is becoming reality. But it has been a long
wait, and needed patience. These days, VMWare even allows me to run any
OS within Linux and therefore practically any application.
Fortunately, the pace of uptake of Linux and the demands of users are
generating more and more high-grade software usable on Linux. I'm
confident that, while there are still areas where Linux-compatible
software is inadequate or absent, these gaps will fairly soon be filled
and users will not need so much patience.
And people who are curious about Linux, and wonder both how they will
get on with using it, and whether it can do the jobs they want, are
likely to find it at least as attractive (in most ways) as what they
are already used to.
Best wishes to all,
Ted.
--------------------------------------------------------------------
E-Mail: (Ted Harding)