posted by Barry Smith on Tue 23rd Dec 2003 20:02 UTC
IconSo far this series has checked out Lindows 4.0, Libranet 2.8.1 and MEPIS 2003.10. Meanwhile both and Xandros have been busy little beavers and spat out new versions of their software, each on the same day. Since I am trying to find the best bang for my buck in a Debian based commercial distro, and since I am already a registered user of LindowsOS, I felt compelled to download a free copy of Lindows 4.5 to have a second look at this thing.

Hardware note: Main system = P3 1 gig, 384 meg RAM, 56K USR internal faxmodem, onboard i810 video disabled in BIOS, 32 meg Radeon 7200 PCI video card. I took the fan out and worked on it a little, but it is still too noisy. Secondary system = P3 450, 128 meg RAM, 4 meg ATI Rage agp, Lucent winmodem.

I am still on dialup, so I was forced to humbly beseech my main broadband equipped contact again. They came through for me once more and a week later I had a copy in my hand. Lindows users who either can't afford or don't have broadband available are pretty well screwed when it comes to downloading their ISOs. I have read that the issue with being able to resume a download had been fixed. I am afraid it still doesn't work for me. I am getting quite tired of this issue.

In this series I have been looking at things that most reviewers either seem to ignore or skim past. I want to know how a distro *feels* in daily use. I want to know if you are forced to spend more time working on the distro than actually producing some productive work. I want to know how convenient things are. I want to know if it is worth the trouble.

But in the interest of balancing things out, I have also decided to utterly ignore most of the things that a standard review homes in on. For example, I just don't care whether a particular distro uses kernel 2.4 or kernel 2.4.2 or kernel I only care about whether it boots up and recognizes my hardware.

Eventually you reach a point of diminishing returns. After three years of jumping from one Linux distro to another, the main thing that strikes me about all these many distros is their absolute similarity to each other. It takes a little bit of time to catch onto this, but sooner or later anyone will see it.

Sure, at first glance they *look* different. But most of the distinguishing points are superficial and cosmetic. One of the nice things about Linux is the fact that appearances can be adjusted. When I used Lycoris I liked the appearance of their desktop. But I could easily take the same wallpaper and icons and make my Debian system look identical. The same thing holds true for RedHat's Bluecurve, or Knoppix's eyeball wallpaper, and any other theme or style. If you want Windows desktop similarity all you have to do is make a folder named 'My Documents' and plop it on your desktop. If you want a 'My System' folder just make one and drop in some shortcuts to your disks, printers, control panel and anything else that suits you. You can even make your own icons from scratch if you want, which is kinda fun.

But there isn't much reason to prefer one distro over another in terms of appearance, basic functions and applications. Most of the Linux distros that I have tried seem to spend their time chasing each other and keeping up with the Jones's. Ultimately they all use (more or less) the same kernel, and roughly equivalent versions of XFree86, KDE, Gnome, Gaim, Tuxracer, PPP, Mozilla, etc., etc. et-endless-cetera.

Same basic kernel. Same basic desktops, with only nitpicking differences between different versions. Each new DE version release sparks a renewal of the holy war to make sure that *their* distro is the first to incorporate the latest (and therefore greatest) version of KDE or Gnome or whatever. To me it is all rather mysterious and puzzling. I hear hardcore geeks ranting about the differences between KDE 3.0 and 3.1. No doubt there are some differences, but I have not noticed any of them in my daily use. Nor do I really and truly care.

Judging from the evidence of my own eyes, I have concluded that Linux distro developers regard their end users as either being capable of writing their own device drivers in assembly language, or being too stupid to live. One or the other. Either you have to know the undocumented tricks of the trade that only years of experience will bring, or you need to be blocked away from anything that might "burn da widdy baby's handsie". Neither extreme is an accurate assessment of the average end user. So I decided to write some reviews for what I consider the average end user.

For convenience I mentally subdivide Linux distros into three groups. RPM-based, Debian-based, and Slack-type. I have chosen Debian for my needs because RPM is a pain to use and I am not smart enough to cope with Slackware.

What else am I actually looking for?

1) Stability (please don't try to tell me Linux is uncrashable.)
2) Package selection (free) and how well packages are integrated into the coherent whole
3) Flexibility in updating (free or cheap)
4) A paranoid obsession with security
5) Anal retentive focus on hardware recognition
6) Good community forums, full of spirited debate
7) Prompt technical support (free)
8) Documentation (I would say "good" documentation but I have yet to see a distro that provides it.)
9) Lots of pleasant GUI tools to get things done with (Yes, I can cope with the command line. I have been using the command line in first DOS, and now Linux, for twenty years. But why should I have to do it in this day and age? I could use a drain snake to clean out my own sewer line too, but I prefer to let a plumber do it.)

No distro offers all of this. But some offer more than others.

One thing that doesn't really concern me is whether or not a Linux distro gives away their product for free. Perhaps I have been coarsened and hardened by my years of proprietary software serfdom, but I don't have a problem in principle if people want to sell their product instead of giving it away. As long as a company complies with the legal requirements under which the GPL portion of the software was first released, I do not object if they want to get paid for their work on the non-GPL parts. That alone marks me as a non-purist I suppose, but there it is. Besides, I could get a bootleg copy of any distro on the planet if I went looking for it.

On the other hand, if a release is fully compatible with the free Debian repositories (like Libranet and MEPIS) then you have the security of knowing that you will never be left high and dry because the developer went out of business. That sort of thing must be considered when you are calculating which basket to put all your eggs into.

For instance: Lycoris, Mandrake, and RedHat all use RPM-type package management. Yet RPMs from Lycoris will not work on Mandrake, and RPMs from Mandrake will not (consistently) work in RedHat, and RPMs from RedHat cannot be depended on to work in Lycoris. That is not acceptable to me. Part of my objection to the Microsoft monopoly was the way they have used proprietary file formats and lock-in to make it difficult/impossible to use anything outside the Microsoft realm. I did NOT finally throw off the Yoke Of Microsoft so that I could go straight to a proprietary Linux variant and buckle myself into a different type of harness. I am willing to settle down with one particular distro that suits me best, but I am not going to give up my option to access other sources. Thus, another reason for me to select a Debian-based distro.

Lindows is fully compatible with the Debian repositories of course, but the Click-N-Run (CNR) warehouse is not. You can use either one, but don't try to use both apt-get and CNR on the same system. (*CRUNCH* goes the dependencies.....) I don't know yet if Xandros 2.0 is fully compatible with the Debian repositories. I have heard people claim both sides of the question. It will be one of the first things I check when my copy gets here. At least with Lindows, I know that if I choose to do so I can abandon the CNR altogether and switch over to the standard Debian repositories at any time. This is very important to me.

Lindows makes lofty claims about version 4.5 being vastly superior in terms of hardware recognition. Hardware recognition was one of my main problems with 4.0. So lets see if things are really improved that much.

Table of contents
  1. "Introduction & Explanation"
  2. "Main Review - Lindows rides again"
e p (0)    60 Comment(s)

Technology White Papers

See More