Getting business data right takes effort but the problems and results are fascinating. The time and effort needed to get data straight is a stout argument for sticking to a common POSIX or Unix core so that things can be reliably found in the same places across various systems (or at least a common way of asking the system if and where resources are located!). Large business data models and often the applications built to utilize them seem to rev far slower than the current pack of central subsystems and kernel APIs do and this is a problem — just contemplate all the COBOL still running out there! Business customers need the data and applications that they drive to work and don’t really care what OS it runs on, so long as their data doesn’t disappear and business keeps costing less to perform.

My customers are raving mad about the crap Microsoft pulled with Windows 8 and have quite often taken refuge to my Linux systems instead. Great. But now nearly every company paying for Linux development (RH, in particular) is racing to mirror Microsoft’s mistakes! This isn’t helping anyone, not even the people funding the efforts.

The market play can go to whoever provides stability on the business platform — and that’s true for the OS layer, the virtualization layer, the data interchange layer, the remote filesystem layer, the programming API and even to a large extent the GUI layer. If Mac weren’t so darn expensive or if IBM were to suddenly reactivate the part of its hive-mind that used to know all about consumer OS and interface design then they would clean house readily. Instead we’re left with whatever the (mostly) FOSS community can come up with… and instead of seeing that community concentrate on filling the “great consumer OS” hole we’re utterly distracted by the idea that everything has to be a mobile OS or ever more ridiculous reinventions of the Web.

With a few million in funding I suppose I could support a reduced server-end “data appliance” distro in-house and force a drone army of code monkeys to keep packaging applications across distros X, Y, and Z regardless how ridiculous they each become. That would be a case of “worse is better” and probably make really good money — too bad I don’t have the funding. Har har!

*** Me too! I worked in firmware development for many years. Came there from a background of electronics. We kicked sand on business programmers at the beach. The only things that interested us were efficient bit-saving code running right next to the hardware.

Then I sort of retired from that world and started my own business. Business problems became much more interesting when they started to eat my life! So over the past 5 years I have written probably a hundred thousand lines of Perl.

On my server, I run Slackware – it is indeed very unix-like, and I revel in knowing where stuff is, and in its straightforward architecture that lets me get things done without too much “magic”.

OTOH, Slackware does not have a good package manager, and for desktops I run Ubuntu. Because I do not typically hack on my desktops, and prefer that they just work. Lately though Ubuntu has gone to this really horrible “Unity” interface, and I have to manually regress it to gnome every time I install one. Might be time to look for a different desktop distribution.
– Jerry Kaidor

As for “following”. I’m not sure what you mean. If you mean “follow” in the “web-based social networking site” sense then you’ll be disappointed to know that I don’t have a Twitter, Facebook, Google+ or whatever else account — life is too fragile, fleeting and precious for me to spend my time bothering with artificial friends, SEO boosts or manufactured reality. If you mean “follow” in the sense that I should take what contrastrebellion.com has to say, then you’ll be happy to know that I did give it (brief) look, and agree completely — which is why when I find some time I’ll give the CSS a look.

An interesting BSD list thread titled “Lennart Poettering: BSD Isn’t Relevant Anymore” surfaced as a response to the Slashdot story last summer.
There is a post here that I think we could all benefit from reading. I know you’re a big fan of SELinux, but he makes a good point on the maintenance model being in question.