CNN article on Linux 25
An anonymous reader sent us a link to another
article on CNN about Linux.
Talks about what is needed- mainly installation help (duh) to acheive
mainstream success for a distribution. Talks about consolidation
of distributions in the next few years. Nothing truly unique
or insightful, but some things that are worth thinking
about if you hadn't before.
CNN Article? No, it's LinuxWorld... (Score:1)
Oh well...
re: (Score:1)
Well I agree with his comment on frosting (??). It's a shame people can only tear articles apart for mistakes relating to GPL or possible FUD. Look to the underlying point of the article - Linux needs better "frosting" (although diehards probably won't care, and quite rightly so) to make it more attractive to the general user. There is still a long way to go. I just switched from Windows to Linux for most things (except games), so the Linux world must be doing something right! In fact, I'm enjoying playing with Linux so much that I'm finding it difficult stopping myself from buying some nice little hardware upgrades!! DOH!
Easier installation? Show me the code. (Score:1)
People continually say that Linux "needs" an easier installation, or simpler configuration, or whatever. Write it. That's the only way it's going to get done. It seems there's a bit of a catch 22 -- the people who know enough about Linux to hack up a distribution (or tools) aren't very interested in coding this sort of thing. Oh well.
I prefer building everything from the ground up, so a distribution is pretty meaningless to me (although when I'm lazy, I go from a Slackware base). All you people saying specific distributions have specific problems in their install, tell it to the creator of the distribution, not to Slashdot. When I see a comment on a bug/feature in a program I hack on mentioned on Slashdot, I try to bring it up to other developers or fix it myself. That requires someone here reading, though -- we're not psychic!
RedHat's installation is all open source, you should have no problems patching in (for example) early starting of gpm, or of a better way to select packages -- do it, and send it in to the developers. Patches rule.
The author... (Score:1)
is Rick Cook!
If you've ever read his books, you know he must have been a UNIX dude from way back, and his comments on Linux are pretty intelligent, too.
Sure, there isn't a whole lot of new stuff here, but the strategy looks sound, like what RedHat as a commercial firm is currently trying to do... (and winning
(besides, his books are funny and fun-to-read
Consolidation = Death (Score:1)
God built it that way.
NO PEACE!!!
{anything more I say on this subject woulld be trolling. It's a joke. Sort of.}
What the author says... (Score:1)
The next part would be to make it easy for vendors to hook their applications into my installation system. That would encourage them to write applications for my version of Linux first.
What the author means by "installation system" is not clear; an InstallShield for Linux, perhaps? RPM and Debian already do an excellent job about installing and removing packages, within the limitations of user interface variety and such (e.g., KDE or GNOME or whatever integration). This in the end would wind down into another UI debate.
Fundamentally, I'm betting that I not only have a superior feature set in my "frosting," but that I can execute better than anyone who tries to copy me. (In the case of offering help, there are ways around the GNU license as well, such as separating the code from the text and copyrighting the text. Not recommended, but it would probably work.)
This is only one of five or six possible value-adds to Linux. In order to make them work, I don't have to violate the GNU license. All I have to do is make the frosting on my cake more attractive -- and sufficiently different -- so people will choose me and it won't be as easy for those who choose someone else.
So all this sounds like "bastardising" Linux in some way, doesn't it? That "frosting" comment reminds me so much of the Halloween documents it gives me shivers. If such a thing were to be made, it better be just as open as Linux. Even better, make it part of the LSB or some other similar standard.
^D
they want the results (Score:1)
>the process, they want the results.
This is an interesting observation considering how much time people will spend trying to figure out and use complex GUI based programs when they could do the same thing in less time, including the learning curve, with a non-GUI based tool. With a solely GUI based tool scripting is extremely difficult, and you have to learn how to do it the way the vendor wants you to (for example, Word Basic). Contrast that with something like the Gimp, which contains script-fu, but also contains bindings for other scripting languages -- making the learning curve for scripting the Gimp much less steep if you already happen to know one of the languages that has bindings.
If people were not interested in process, there wouldn't be classes which teach you how to use office suites because it would be so obvious how to use them. I know people who have a hard time using the tools they have because they can not remember the process (select a menu item, change a few options, hit ok) it takes to do something -- this problem transends the interface even, and is part of the general laziness people have. Unfortuantely, most people are more thought-lazy than time-lazy -- that is, rather than think through something, they'd rather just randomly try things until it works.
Also, if people were only interested in results, they would realize that it's worth their time to learn the process if they know they are going to continually want the same kinds of results on a continual basis.
Making installations easier (Score:1)
1. Easier package installation
Slackware was just beast, making me sit there and wait for each package to be installed for it to ask if I wanted to install the next one. I think there was an option to automatically answer those, but I didn't want an all-or-nothing type installation. Red Hat is a little easier, but again I like to choose my own packages individually, which is still a chore in Red Hat. I'd like some sort of hierarchy where I can disable/enable an entire category, *or* I can go and choose the packages therein individually. Currently Red Hat is either/or.
2. Mouse support.
Other distros may have this, but Red Hat doesn't. The first thing it should do is try to determine the mouse type (it asks later on) and then run gpm so that you can actually use the mouse during the installation. Keyboard navigation can become a task in newt (I'm assuming that's what Red Hat used for their installation program).
3. Better GUI.
Red Hat (as I said) uses newt for the installation program. Functionally that's fine, but for a little more ease of use, if they could use a simple VGA X server (will work on just about any card) instead (maybe having newt as a backup for those few cards that won't work or for non-CD installs) that'd be pretty nice.
4. Better usage profiles for default installations.
Red Hat, for example, has only a couple usage profiles, you either want everything or the bare minimum. It'd be nice to have profiles for say, developers, end users (non-developers), mail server, etc etc. with default packages selected to fit that profile (and then let you further customize which packages are actually installed).
5. Better X-window configuration.
I've never gotten an XFree86 server to work from the Red Hat installation; I've always had to tinker with video card/monitor settings after the install. Some of this is unavoidable without a more robust device information database, but someone without the experience will never get X-window to work by themselves if it doesn't work out of the box, and by the time they get help, they will be deeply discouraged. It goes without saying that Red Hat needs to dump fvwm95 in favor of (at least) a more smooth interface (obviously they are trying to do this by using GNOME and possibly E [although I wouldn't make E the default if I were them because it is by and large unfinished; maybe WindowMaker, Blackbox or a better fvwm2 setup]).
6. A post-installation tutorial.
Of course not many people will just jump right into Linux without at least a book or someone there to show them what to do (esp. if they've never used UNIX). An interactive tutorial for the newbies would be nice, that goes through basic discussions of the CLI, X, root v. user, and where to get more help, etc. This could probably ship as a seperate CD, so that the first time root or a new user logs into xdm or something, they get a prompt to enter in the CD and the tutorial program is launched when they do (or they can choose not to launch it of course).
--
Aaron Gaudio
"The fool finds ignorance all around him.
Distributions... who cares? (Score:1)
--
Aaron Gaudio
"The fool finds ignorance all around him.
No, there is another way (Score:1)
Provides:
Provides: libc
.
.
.
rpm will see these and will consider those things installed. So basically when you go to install a package and you get missing dependancies, you first see if indeed you have those dependancies installed by hand (via cpio or something) and then add them to that file. BTW there is also a tool called rpm2cpio which should translate the rpm files into cpio files which can be installed on a Slackware (or any Unix) system.
--
Aaron Gaudio
"The fool finds ignorance all around him.
Bulletproof install program - agreed (Score:1)
Linux installers have ways to go for other reasons too... at our office we just installed RH5.2 using a SCSI CD-ROM drive, and when it booted after install it couldn't see the CD-ROM drive, due to no SCSI support in the installed kernel. That's pretty bad.
There are some good efforts being made towards simpler Linux: check out www.seul.org.
More pseudoinformation (Score:1)
Other good articles on Linux (see inside) (Score:1)
From the Pioneer Press (St P. Minn) is an article about Linux, including a local HS installation.
http://www.pioneerpress.com/tech/docs/tech1.htm
This article has a good pro/con list for Linux
http://www.pioneerpress.com/tech/docs/tech2.htm
and finally, this one has links on, of all things, how to pronounce Linux
http://www.pioneerpress.com/tech/docs/tech3.htm
Later...
(isn't it about time for another non-news annoying Katz squawk?)
If the GUI is difficult, its not written right. (Score:1)
Scripting within a GUI should be no different than any other method. Its all dependant on the scripting language definition.
If something is harder to do using a GUI then one of the following is true...
1. The GUI is designed wrong.
2. You don't how to use it.
I cannot believe the number of people who claim the command line is easier for new users... get real. In some cases there are things that can be done quickly via the command-line - BY people who have the time to devote to learning it.
If something is quicker by the command line there is no reason an icon cannot be constructed to perform the same task at a click (even bundled with a dialog asking for the variable data).
Computers are not for nerds anymore.
.
Bulletproof install program (Score:1)
Take BeOS for example... it doesn't make any dumb assumptions (like that it will be the only OS on the computer), but the installation is still extremely simple. Even the "first partition" thing is a nobrainer, and the whole installation is done in about 5 minutes.
Slarty
Hey, I still use Lattice C! (Score:1)
I also don't agree with his assessment on the GPL (as many others have stated). Add-ons in most any form can be handed out without source. Just don't expect people to keep buying your product if you don't maintain it.
Making Slack easier (Score:1)
Did you not use menuconfig?
"make menuconfig" is a handy, dandy, menu-driven type of thing. Much easier. Set it and forget it.
--
Selfish install program (Score:1)
I am thinking that most end-user types don't really need to install the OS. What needs to be easier is application installs. My wife installs software on her PC with those wizard based installers, but I don't think she's going anywhere near an RPM. World domination means that non-geeks can use it too.