| Recent Posts | About the Author | Navigation |
|---|---|---|
|
|
David is an occasional blogger, software engineer, Nintendo fanboy, liberal, news magazine addict, voracious TiVo user, and bibliophile. He was born in St. Louis, grew up in southern Indiana, and returned to St. Louis to attend Washington University. He hasn't managed to escape yet. He's a fan of free wine tastings, too many tv shows to name, and eating out. David makes his living developing web applications used internally by his employer. He doesn't blog about work because he's heard too many stories about that causing workplace troubles. There's more on the about page. |
|
| Recent Comments | ||
|
|
||
| Recent Photos | ||
|
|
||
6 April 2004 - 1:46 pm
Here are two different takes on Linux for the (desktops of the) masses. In the first, Dan Gillmor argues that "Linux on the desktop" is no longer a "non-starter." In the second article, John Gruber argues that the open source development model makes it difficult to design and implement intuitive user interfaces, and that the mindset behind this ("the hard work of development is in building the underlying foundation, and that the easy part is writing a 'GUI wrapper'") will prevent Linux from making serious in-roads on the desktop. I've made the latter argument in this spot before (or at the least linked to those who do) and am still convinced that it holds true. Anyone want to persuade me of the merits of the first article?
Posted by on 6 April 2004 at 1:47 PM
Comments
"It's still too hard to install applications Xandros hasn't put into this distribution system, a common problem in Linux overall."
This is a big deal. A computer that is locked with whatever software (and hardware drivers) it comes with (because of the difficulty in installing new ones) is useless to almost everyone. It shouldn't even be remotely difficult to install a new application, since the average user has become accustomed to the mostly one-click setup files of Windows (and I'm assuming OS X as well). Anything less at this point is unacceptable.
Posted by Chris Hill Festival on 6 April 2004 - 3:30 PM
Note the qualifier in the quote though--this is only a problem for things not in their distribution system.
A linux system with a more mature ports system (Debian with apt or Gentoo with portage) have nearly anything you could want in the distribution system already. I have yet to find something I've wanted that wasn't in portage.
As far as his wireless troubles, I'd say there are 3 possible sources: crappy distribution, really new wireless card, or crappy old wireless card. Any distribution worth its salt supports all of the old Cisco, Lucent/Wavelan, and Orinoco chipsets. Also, in the last 6 months, support for the newer Atheros chipsets (the majority of wireless cards nowadays) has really solidified. What still isn't supported is crappy old cards (many of which aren't even really 802.11b cards--they were designed before the spec was finished), or the bleeding new setups, such as Intel's Centrino.
As far as his complaints about inability to connect to his corporate network remotely, I have no idea wft he could be complaining about. Networking is the main strength of Unix-like systems. Perhaps his company uses some funky VPN that isn't widely established.
On the other hand, he doesn't point out (since he didn't experience) two of the biggest outstanding issues with linux--it is still difficult to get it to play nice with windows on the same machine (rather, windows refuses to play nice--95/98/NT was worked out, but 2000 and XP changed thihgs), and support for new hardware is spotty, like aforementioned Centrino wireless.
Posted by joe on 6 April 2004 - 7:09 PM
"As far as his complaints about inability to connect to his corporate network remotely, I have no idea wft he could be complaining about. Networking is the main strength of Unix-like systems. Perhaps his company uses some funky VPN that isn't widely established."
Joe says that "networking is the main strength of Unix-like systems," but that claim is patently absurd from the interface point of view. Ridiculous, even. Quick, which RC file do you have to change in order to set the IP address at boot? There's no way to answer that unless you know the OS, version, and distribution of your system. Slackware, RedHat, Debian, and FreeBSD all do it very differently, and to change your IP you have to either read lots of man pages or read through the boot scripts themselves. Or you can use the crufty programs to change the IP, but then you still have to find out what the "correct" program out of the 12 or so is the best one to use.
This is the crux of it all: if a user fails to get something working on a computer, the failure could either reside with the user or with the software/hardware. I would argue that the user is *never* at fault. If somebody can't figure out how to get it to work, the interface is buggy, and that's all there is to it.
Posted by Charlie on 7 April 2004 - 12:09 AM