% fortune -ae paul murphy

Unopposable Linux vs immovable ignorance

Last week's discussions on why desktop Linux often fails in big organizations drew, among many knowledgeable comments, this note from "webninja":

Interoperability and integration

I tried Linux at home for a couple months, just to see how real it was as an alternative. The one thing that sunk it for me was not having a common copy/paste buffer between apps. You can't even cut a URL from a doc and paste it in a browser - that's just broken. Best-of-breed in this case is like picking car parts. Try and make those "best" SUV wheels work with the "best" Ferrari engine installed on the "best" Mercedes chassis.... get it? Having all the major apps designed by the same team as the OS might be anti-trust-bait, but it just works.

I know nothing about webninja, but the combination of good grammar with that reference to a best of breed strategy and the misuse of an Apple slogan in defence of Microsoft, suggests to me that this is someone who either is, or listens to, a nineties Windows or data processing guy promoted beyond his abilities.

So suppose he's your boss at work - and most of us have worked for this guy in one form or another - how would you go about getting a Unix proposal past him?

Would you, for example, respond to the expertise on display in the comment by pointing out that the first commercial GUI to do this well was SunView? that Apple had it next, that Microsoft didn't get it for another ten years? or that the problem he complains of has to be either imaginary or created by a completely screwed up install?

I wouldn't: my assumption with someone like this is usually that if the facts mattered he'd know better already - so arguing on the facts now will just reduce your downstream ability to communicate with the guy.

And what that means is that if you want him to approve or support any kind of Unix project - Linux, BSD, or Solaris - then the last thing you want to do is mention any of the traditional reasons for choosing Unix: cost, reliability, or software accessibility. Instead, you get what you want by cynically leveraging his mistaken beliefs about his own role and technical competence against him.

One approach, for example, that I've found to work well is simply to wrap your idea in a user inspired pre-feasibility software search. Pick whatever it is you want to do, have a user raise a requirement that could fit - and stick your hand up as a knowledgeable volunteer ready to review the third party software market to see if there's something out there that will do the job.

With luck he'll view the user as a nuisance to be placated, you as someone whom he can't quite trust, and the project as something that will never get past budget review anyway - so assigning you to the job you just volunteered for looks like a win-win for him.

Sure, he thinks he's putting you down - but what you do is say "thank you" and shut up while going off to use his imprimatur to play with some application or other on whatever Unix variant or software you wanted to spend some time with. Eventually you'll want to write a nice little report and maybe do a few demos - just don't say anything about "platform" costs or performance unless, or until, your user friend kicks the thing over to an IT steering committee and they ask for a budget.

Most such projects never reach that stage, but you can generally keep doing this until one does - and at that point your "webninja" will find himself caught between the rock of having approved the project and the hard place of having to explain why paying a lot more to do it his way makes sense "now" but didn't "then."

Be aware, however, that even the dumbest peter principled manager out there won't let you get away with this twice in a row - so pick something worth doing and plan on either taking your new skills elsewhere or creating your own little post project Unix niche where you are.


Paul Murphy wrote and published The Unix Guide to Defenestration. Murphy is a 25-year veteran of the I.T. consulting industry, specializing in Unix and Unix-related management issues.