Kevin Coonan, MD wrote:
<soapbox subject="Operating Systems and the Tyranny of the Majority">
I am a long time lurker, first time troll baiter...
The clinical informatics folks, esp. those in CIO positions and commercial
system vendors needs to wake up and realize what a nightmare Windows is.
There remains a naiveté in plush offices about the alternatives (both OSS
and proprietary alternatives to Win32).
The Linux community needs to make some better cases and fix some of the
end-user annoyances and get these people using GNOME or KDE for their
day-to-day work.
one basic parameter in all of this is the cost of the OS license versus
the cost of the hardware. You can get a Dell sub-$1000AUD (=$750 USD)
desktop computer, where Windows XP professional is about $300AUD of the
cost. It won't be long before more than half the price is the OS
license. And what future does MS have in store for us in terms of
ongoing licensing costs? Whereas what would a Mac mini cost over 5
years? What would a home-built desktop running linux cost over 5 years?
I suspect Toshiba's latest laptop running linux would be very
competitive over 5 years.
While end users (particularly Laptop users) may need to rely on Win32/OS X
for some time (waiting for drivers....waiting....waiting), esp. if Novel,
WordPerfect, IBM/Lotus don't start churning out some M$ Office alternatives
(I have tried OpenOffice and rapidly returned to the Borg's latest release).
CrossoverOffice, WINE, etc. are a crutch. Currently important crutches, but
crutches none the less. A mutually agreeable file system (other than FAT!)
would also be a huge benefit for those of us who are in dual-boot limbo for
the foreseeable future.
I agree with most of this.
Gates, et al, will remain the 227kg gorilla, so seamless file exchange is
going to be an ongoing requirement. Painful, but true.
If you code applications for compatibility using a standard language you can
deliver to any of the above (with the exception of threads in C++ and the
lack of a STL-like thread library).
well actually, a standard meta-model is much more important. Languages
should be seen a surface syntaxes on a decent OO/functional meta-model.
MS have actually recognised this with dotNet - it lets you use multiple
languages, and binds them to the one common language runtime. Java in my
view, apart from being a weak language (but getting better finally with
1.5...now it feels a bit like ANSI C instead of K&R C;-) has made a
strategically wrong turn. I think the future is the CLR way, not the JVM
way. (Not that MS didn't make hundreds of errors in the CLR design, and
they still haven't got generics, nevertheless....)
- thomas beale
