« May 2005 | Main | July 2005 »

June 22, 2005

Desktop Linux: REDUX

Link: jorge @ whiprush.org: Allie's New Radio: Making Desktop Linux Happen: NOT.

Users would disagree, it's not a matter of intelligence, they just want their shit to work. Choice is great when you know enough about something to make a choice, in the meantime, users trust distributors to pick a sane default that lets them get their work done.

Thanks for your comments, Jorge. On this note, we are in total agreement. You're correct, I have not tried Ubuntu, having limited myself to the "corporate" distros from RedHat, SuSE and (years ago) Caldera.

Of course it's clear that smart developers can do the right thing with respect to packaging and installation: witness Windows and OS X. My point is that in the Linux world, they (largely) have not, for reasons that seem to be indemic to the community itself. That is the basis for my argument that Linux will never get it right.

Another commenter echoes this:

I think most of the dissonance about choice is a fundamental misunderstanding of how the Linux software ecosystem works... The open source software ecosystem is a collection of specialized software packages, designed to do one or a few things very well, that work like a set of building blocks. A Linux distribution merely sorts, selects and assembles these pieces for its target audience, with a spit-and-polish touch that also caters directly to its audience... This puts some people off - they want an tightly integrated software package that appears to be a smooth monolithic entity.

That said, I'll see if I can give Ubuntu a try to see if it is, in fact, different. As I said at the outset of my post, as an "enterprise ISV" I am forced to live and breath Windows. I long for better. OS X is better, but, so far, I can't say the same for the Linux desktops I've experienced.

Posted by Allie Rogers at 02:13 PM in Web/Tech | Permalink | TrackBack

June 21, 2005

Making Desktop Linux Happen: NOT

Link: Making Desktop Linux Happen.

Although I make my living, partly, as a result of the ubiquity of a standard Windows desktop in the enterprise, I have no particular love for Microsoft Windows.  It is both the reason for, and the bane of, my existence.

That said, Linux has been spinning it's wheels on the desktop for going on 10 years now.  It's so plainly clear that it is doomed to failure, I really wish the community would move on and focus energies in a more productive direction.  It's not that it's impossible for someone to unseat Microsoft from the corporate desktop.  Rather, it will NEVER be Linux and the Open Source community in it's current incarnation.  NEVER!

Try this.  Take the latest and best Linux distro.  Pick any one of them.  Now install it.  You're presented in a litany of choices for everything from desktop flavor to email client to window manager (whatever the hell that is).  There is this fallacy that choice = good.  WRONG!  BEST = GOOD!  How should I know whether I want or need pine or elm?  Guess what?  The answer is NEITHER.  The answer is: just install the bloody desktop in the BEST configuration.  If, later, I want to add packages, Linux gives me (unfortunately several) ways to do that.  The GNOME/KDE mess is even worse. 

This more than just building a better installation procedure, although that's part of it.  It is fundamental to a culture that seems to want to include everything based on the fallacy of "choice = good".  The desktop doesn't need everything.  It needs one, best thing.  Choice only complicates matters to the point that migration away from Microsoft Windows becomes too complicated.

Who decide's what is best?  Well, ultimately the market does, but in the meantime, the distros should just choose one.  When the market complains that is broken in this way or that, fix it.  That's how Microsoft succeeds.  That's how Apple succeeds.  Take a hint and follow those who have demonstrated how to capture successfully capture desktop market share.

I know open source is equated with a certain level of freedom.  But the corporate world is not about freedom and choice.  It's about standards and getting on with making money without wasting time rebuilding the ultimate email client or desktop.  As a result of all of the choice in the Linux community, it is really hard for ISVs to actually create, build, test, deploy and offer REAL business solutions.  Few have the resources to make it works on each of 5 popular distros, 2 desktops, so many window managers, etc., etc.  Windows is windows, more or less, and it's generally harder to build an app that will NOT work across versions and service packs than the other way around.

So, my advise to Linux is to do what Apple did.  Stop offering "choice".  Start selecting a best-of-breed desktop image and stick with it.  Sure, those who built the bits that are not selected will be upset, but they'll get over it.  Mac OS X succeeds as THE BEST *nix-topping desktop precisely because there is no choice.  There's a, single, great presentation system (light-years ahead of GNOME and KDE despite the fact that is much less mature).  There's a standard mail client, calendar and basic set of productivity tools.  There's built-in remote management and deployment.  There's also built-in popular open source *nix bits, but not 10,000 of them nor a choice of which ones are there and not.  Mail.app is there, but pine is missing.  Get over it!  The point is not to offer choice, it's to offer best.  That's what the corporate world wants.  But given the culture and nature of development in the Linux and Open Source communities, I hold out no hope they'll ever get their head around this simple concept.

Posted by Allie Rogers at 11:14 AM in Web/Tech | Permalink | TrackBack

June 07, 2005

MacDevCenter.com: It's True! Jobs Switches to Intel: A Review of the WWDC 05 Keynote

Link: MacDevCenter.com: It's True! Jobs Switches to Intel: A Review of the WWDC 05 Keynote.

Good summary of the WWDC 05 Keynote.

Posted by Allie Rogers at 05:27 PM in Web/Tech | Permalink

Chips with everything

Link: Chips with everything.

A nice summary of the impact of Apples move from IBM PPC to Intel x86.

Posted by Allie Rogers at 05:18 PM in Web/Tech | Permalink | TrackBack