% fortune -ae paul murphy

Linux: when Yes becomes Nooo

In thinking about what Linux in particular and Unix in general needs to succeed against Microsoft, I've come up with a couple of questions where more talkback contributions could be helpful.

One of those is the concept of the killer problem - modelled on the idea of the killer application (on which more tomorrow), but backwards. What's the basic issue slowing Linux adoption?

My candidate for this is misplaced expertise. Almost everywhere I go these days I run into people who are turned off on Linux because they, or a trusted source such as the nitwit on the next barstool, have had a negative experience with it.

In my own experience there are failed, or failing, Linux installations all over the place with a plurality of personal experiments failing because the graphics subsystems are mis-installed and a majority of business failures happening because the people who set them up already know everything they need to about computing and so implement Linux as a second rate Windows substitute to perpetuate all of the costs and inefficiencies of the technology they know while getting none of the benefits of Linux.

My friend "Liz", for example, is a very bright, hardworking, and extremely knowledgable Windows systems administrator who, at my suggestion, loaded Debian on a spare home PC - some kind of higher end HP box. At first I got positive reports, but after a while it became clear that she never actually used it - just sometimes turned it on, let guilt drive an hour or two of struggle, and then turned it off to work with XP on another machine.

So why? I could never get a straight answer, but when I finally convinced her to let me see the machine the reason was immediately obvious: GNOME looked horrible and ran with the ponderous agility of an oil tanker. What had apparently happened was that the default install hadn't quite got the graphics hardware or the (third party) monitor right, and she had known nothing to suggest that the resulting mess didn't fairly represent a Linux GUI.

And then there's "Larry." His business card says he's the "Manager: Internet Server Operations" and he has a staff of four to look after an e-business site running on several racks of IBM Xeons with Red Hat and a big EMC.

This is a websphere custom implementation that cost his employer more than a few hundred thousand to set up, but completes only a few thousand fairly simple transactions every day with fewer than a hundred thousand daily accesses to the front page.

So here's part of what I can't fugure out: what do those people do all day?

Larry's busy - at least he's usually not available to take my calls. His people must be too, since service requests generally go answered until blessed by a steering committeee - but what do they do all day?

I know I haven't a clue. Do you?


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.