> Now that the IT guys have seen it in use, they're a lot warmer to the idea
> of deploying more LGX in the office. Especially on days when the windoze
> boxes are causing them nothing but grief.
> Has anyone else gone this route?
At my last job, I persuaded our CIO to let me re-allocate hardware that was
at the end of its 2-3 year lifecycle for NT servers and install GNU/Linux
to meet needs for projects where we didn't have immediate budget or where
other teams were simply taking to long to deploy a solution.
Often times, we could roll out a free software / open sourced solution
"prototype" which would go into production, and remain in production,
faster than other solution development teams ( e.g., Lotus, Windows, 4gl's
) could handle the analysis phases of such projects. Some of the teams
simply couldn't feasibly develop a solution due to limitations inherent to
the proprietary legacy software we had deployed (e.g., we'd pay an external
vendor to develop an application which we needed to change but for which we
only had rights to the binary image, not the source code).
We ended up with a half dozen or so systems that way running things like
Apache, MySQL, Samba, etc. Once they proved their usefulness we started
getting money allocated specifically for hardware to run GNU/Linux.
We were able to increase ROI for hardware by increasing the hardware
lifecycle and meanwhile get management to recognize the strengths of
GNU/Linux versus proprietary software which increased its deployment as a
All of this was with no license cost for developer seats or production
The main place where we really killed other teams were in productivity and
cost of developer seat licenses / deployment licenses since our
productivity was whole orders of magnitude higher and since our licensing
cost us next to nothing even though we unnecessarily but intentionally paid
for every copy of Linux software that we deployed. Another main advantage
that GNU/Linux had was as middleware to get 2 disparate proprietary
applications that needed to fit together but didn't to "talk to each
Today, I'm working in an environment where GNU/Linux and Solaris are the
norm in server-space and we are pushing Linux/X11 out as a client platform
or as a server-based thin client solution for web kiosks.
We spend a fraction on software/hardware/staff than we did in my last
environment which favored windows/mainframe as a primary client/server
The upshot of this is that GNU/Linux will make you look like a hero in
environments where they are used to the high costs and low productivity
generally associated with proprietary software development systems such as
Windows, Lotus, or 4gl's.
When you can tell an IT director with a straight face that you are going to
save them $25,000 - $50,000 in licensing cost, and another $25,000 or more
in labor, and that you've already got the application prototype ready to
roll, they have a very hard time *NOT* listening to you.