Michael Dell Sees Future In Linux Desktop

Robert McMillan writes: "Linux Magazine has just published a pretty interesting interview with Michael Dell -- not exactly widely considered to be a Linux booster. But he is keynoting at LinuxWorld in San Jose tomorrow and he does talk about why Dell is now interested in Linux. Interestingly, he also says he sees good things for Linux as a desktop OS."

Comments

Sounds like glazing to me.

[Lonesmurf]

LM: Have you spent much time with Linux developers? For example, have you met Linus Torvalds?

Dell: Yes.

So, Mike, which is it? Did you spend time with some Linux dev guys or did you meet The Man himself?

Hmm.

Rami
--

Open Letter to Michael Dell

[coats]

I'm going to buy a new home system soon. I run Linux both at home and at work, and I will probably buy a moderately "heavy" system -- 512 MB, 800+ MHZ, 40+ GB fast disk, 21-inch monitor, top-end graphics. etc., and I want it to be running Linux.

As a matter of fact, I tried to look at your web-site yesterday. But not only did it not have any obvious Linux-system links, it was configured to work with Javascript-enabled browsers only.

Not only do I regard active content such as Javascript to be a security risk, I have also had too damned many denial-of-service attacks disguised as Javascript web pages!! Too many web designers regard it as an opportunity to "one-up" frame-bombs. For this reason, I categorically refuse to run Java or Javascript.

So either clean up your act or else forget about me -- and my friends who consult me -- as a customer.

Don't dis M$

[Shotgun]

Don't you just love the way he sidestepped trashing MicroSoft. Not that I blame him. He wants to sell computers, not fight a holy OS war. But couldn't you just tell by his tone that what he really wanted to say would go something like:

Q. And what about MicroSoft? How will Linux affect them?

A. I hope that piece of crap OS get's plowed under like last years chaff so that we don't have to support Gate's business model anymore. I'm so sick of supporting his closed source flotsam I could puke. Linux will let us cut our support staff in half, and with the source code we'll be able to push the envelope of what the PC can do. In a few years, you will see us shipping Dell computer that truly inovative and not just Wintel tag-alongs.

How can Wine "overtake" Windows?

[cpeterso]

Direct3D emulation under Wine is a little behind, but will probably eventually overtake Microsoft's implementation too.

How can Wine's Direct3D emulation (or Wine itself, for that matter) "overtake" Windows? It is trying to BE Windows. Can Wine really be a better Windows than Windows?

Re:Building a desktop OS from scratch

[SimonK]

If multi-user isn't useful "on the desktop" why does even Windows 95 (ORS 2.5, at least) have user profiles ? and NT have file permissions ? You seem to be confusing "desktop use" with "home use", but even for home use, multi-user capabilities are useful: my brother and parents have their 98 machine set up with different profiles for different users.

I'd suggest that instead of removing command line and multi-user functionality from the system, what you really want to do is to hide it. NextStep did at least part of this fairly well, by hiding the "normal" Unix directory hierarchy, because no user ever really needed to see it.

Similarly, I'd suggest hiding the existence of different user permissions, by running as little stuff as possible as root, and never allowing root logins, but making all manipulations of the (hidden) system files dependent on a concealed "sudo". This prevents users from accidentally fubaring the system (I assume you don't always login as root, for exactly this reason), while meaning they don't have to know anything about users or permissions beyond entering their name when they start the system.

Re:Building a desktop OS from scratch

[Tet]

under the usual criteria for judging server OSes (scalability, reliability, security) Linux is not a particulary good server OS

I'd have to disagree with you there. Yes, NT fails miserably on all of those, which makes Linux look good. However, that doesn't mean Linux is lacking in those areas. In fact, it does pretty well for all of them. I'm not claiming it's the best in any area, but it's far from the worst. It doesn't scale to the high end as well as Solaris, DG/UX or Irix, for example. Nor does it have the reliability of Tandem or the security of the various MAC-enabled OSes. However, for probably 80% of businesses, it's a suitable server OS for their needs. It has sufficient scalability, reliability and security to get the job done.

Not to sound hypocritial, but.....

[vapour]

"Interestingly, he also says he sees good things for Linux as a desktop OS."

Yet yesterday, 90% of slashdot readers berate AOL for providing their suite of access products for Linux

What's it to be ?

Re:Not to sound hypocritial, but.....

[nosilA]

Yet yesterday, 90% of slashdot readers berate AOL for providing their suite of access products for Linux

There is no such thing as 90% of slashdot readers berating anything - if they did, you'd have to sift through 200,000 or so posts per article... 90% of the vocal comments on that article may have been against the AOL thing, but I'm not even sure that's true. Granted, I browse with highest scores first unless i'm moderating, but most of the posts I saw were saying that this is a *good* thing - or at least not particularly the product of evil.

Even if 90% of the comments on that article were vehamently against it, that's still only a small percentage of /. readers - 365 comments, many of them by the same users, that's only .18% of the "slashdot community" at best... Believe it or not, not everyone here agrees with everyone else, but don't feel the need to post "no, you're wrong" every time someone disagrees with them (obviously I do, or I wouldn't write this post, but that's beside the point)...

So, for the sake of not being entirely offtopic - Personally I think having Linux as a reasonable desktop competitor to M$ would be great, sure, let the idiots use Linux - people will write applications for them, and we'll write applications for us. However, I don't want to see a linux machine on every desk any more than I want to see an iMac or a Windows machine on every desk... competition is the key to survival... I think it's great the Michael Dell sees a viable market, I think it's great the AOL is supporting linux, these businesses wouldn't do anything they didn't think would succeed... So maybe Linux will put a major dent in M$'s market share, and everyone will get better applications, and better prices. Yay free market economy!

-nosilA

Re:Not to sound hypocritial, but.....

[finkployd]

Another thing to remember also is that slashdot is not THE linux community. It may have a A linux community inside it, but as a general rule, the most productive and talented linux hackers are not posting to every slashdot story.

Finkployd

Interesting.

[jw3]

Very interesting, indeed. There are two points I would like to mention, though. Funny that they did not talk about the merits of the OS at all -- only marketing. There are Linux users, so let's support them. So one reason for all of that is the growing user base, which is obvious and not necessarily very intersting per se. However, I take it as a better prophecy for the future (for me as a Linux user) then, for example, the fact that Ralph Nader is using Free BSD.

The other thing is -- by supporting Linux, they rather seem to be competing with SUN & al. then with Microsoft. A lot what he says remains unspoken (e.g. he says "Look at SUN and Microsoft", and then doesn't even mention the latter). Linux warriors may be more obsessed with Microsoft, but one conclusion I draw from that interview is that rather the big commercial Unix corporations can be the real competitor / enemy (however you state that) of Linux in the corporate market.

Best regards,

January

To quote the Jargon File...

[DragonHawk]

Can I use Linux as a verb?

... any noun can be verbed.

A hint for Dell

[dsplat]

Dell needs to put a link to their Linux support page somewhere on their main page. I'd settle for a link at the bottom that says "We also seel systems with Linux preinstalled". I'm looking for a new system anyway, and I am considering buying from Dell, but I need to be able to find the configurations in question.

Linux on the desktop and Dell

[ACK!!]

The latest batch of Dell notebooks look pretty good and have me thinking about them as a real option. However, I wonder how much support this will really translate into. Any *NIX variant demands a good deal of expertise to troubleshoot when something goes wrong. It demands a higher level of Technical Support knowledge than I usually see from most big manufacturing houses.

Will Dell just give you the option and not help you if the OS and the hardware don't play nice or will they back up the kind words with corporate action?

Re:Linux on the desktop and Dell

[gammatron]

Most "big manufacturing houses" (Compaq/DEC, IBM, HP) have their own unix, and therefore, their own unix support group. These groups start out supporting linux until the demand gets big enough to spin linux support into its own group. So your comments may be justified if you're talking about Dell or Gateway, but if you're with Q, IBM, or HP you shouldn't have any problems getting support. Now, they may charge you an arm and a leg once your warranty runs out, but that's another story :)
--

Not to put a downer on things

[Vanders]

It worries me slighhtly that so many companies these days are focusing on Linux as the only alternative. Sure, this is great for Linux, but is it good for computer users in general? Bear with me here...

Out of all the available Operating Systems out there, how many are based on just two standards? There is Win32 (Windows 9x/ME, WinNT), and POSIX (All *nix variants, most BSD's, BeOS, QNX etc). Only two standards? Where is the choice? Where is the inovation?

It seems to me that all anyone is interested in is twisting and squeezing Linux into ever more bizarre and improbable situations (Linux on palm tops, Linux for embedded devices, Linux for games consoles etc.) Is the market stagnating, where no one dares break away from the pack and try something new for a change?

Maybe it's envy on my part, but why can't we have an Open Source project that isn't based around some form of Linux or a POSIX kernel? Is there any room for innovation in OSS these days?

Just had to get that off my chest, sorry.

Re:Not to put a downer on things

[David A. Madore]

Want innovation? Have a look at the Tunes project.

Re:Not to put a downer on things

[deefer]

There _have_ been OS innovations; it's not the software houses that are being slack.

BeOS, Next, all very innovative. Amoeba, Minix, and loads of dusty academic OS's exist.
But nobody really uses them. I've seen seasoned developers weep with joy once they've played with the BeOS API, system admins go green with envy when they see the security & file system of Amoeba. But there is no takeup by industry.
Why is that?
One word : interoperability
Look at all the grief that is required just to get WinXX talking to Linux. Allison, of Samba fame, goes as far as saying "SMB sucks!". Yet we're still tied into the same ways of doing things. It's the profit motive for the S/W companies that actually rewards them for stopping cooperation between OS's (MS Kerberous, anyone?)... So innovations aren't easy to migrate to, it's an all or nothing thing in a lot of cases. Add to that the cost of retraining _all_ staff, and the costs are horrendous.
The only way I can see out of this is to get some Standards police, and have a unilateral legislative agreement that STANDARDS MUST BE ADHERED TO.
Then you get the companies whining that they cannot "innovate" for competitive advantage, and xth Amendment agitators (in the US) saying how the Government is censoring/stifling free expression.
In short, hte companies have us by the short and curlies and they won't let go until they're smacked down _hard_. I don't see this happening.

Strong data typing is for those with weak minds.

Re:Not to put a downer on things

[codemonkey_uk]

Vanders wrote:

why can't we have an Open Source project that isn't based around some form of Linux or a POSIX kernel?

Who is "we"?

This is probably the key to your question, by expecting an imaganary "open source" community to take your project of choice and blast it into the stratosphere, you are creating unreasonable expectations.

You are looking at the headline *NIX Open Source projects, so of course you can't find anything else. Most non *NIX platforms don't have an open source community, they have freeware, and shareware, but not free open source.

Nothing stops you from starting your own open source project. Just don't expect people to support you if you don't toe the party line.

PocketIFS is an open source fractal hacking tool for the PalmOS. It has receaved favorable reviews, one bug report, and not a single line of code from anyone but me. (Perhaps because it Just Works).

Thad

Re:Not to put a downer on things

[FreeUser]

We as in the Open Source community. I'm not saying people should be forced to use YAOS (Yet Another Operating System), but at least the OSS community should have the choice.

The Free Software/Open Source community has plenty of choices. Some people develop for *BSD, many for Linux, a few for BeOS, and a very few brave and masochistic souls for Windows/*.

Most have chosen Linux for a variety of reasons, ranging from the licensing (GNU GPL) of the OS and underlying libraries to the simple fact that, for whatever reason, right or wrong, welike and prefer Linux. This may not reflect your preference, but it does reflect theres, and if you don't like it you are free to start your own project, or fork off any of the free software projects you wish.

The community hasn't ralleyed around Linux because it doesn't have a choice, its ralleyed around Linux because its license (GPL) and quality reflect the desires of the community more than anything else out there.

That having been said, there is a (smaller but just as enthusiastic) community which has formed around *BSD, also largely because they favor the license (FreeBSD) and the quality of the platform more than anything else out there.

BeOS is neat, but proprietary. Bug fixes are arduously slow in coming, precisely because the OS is proprietary and there is limited manpower fixing bugs, creating new features, etc. This makes the platform as a whole less dynamic and less satisfying for me, the end user, when I discover my video captures die after 9 minutes and don't get a fix for six months. It is hard to form a community around a product one cannot contribut bug fixes to, or get bug fixes from in a timely manner. while there is a small, vocal, and enthusiastic developer community or Be despite this, the proprietary nature of the platform (with all its disadvantages) make it singularly unattractive to open source developers regardless of how nice its programming API may be.

Ditto for MacOS. Proprietary OSes are simply unappealing to most open source developers.

Ditto for windows (with the caveat that the APIs are ugly, the system even less stable and much less elegant than any of the aforementioned OSes, etc.)

We have plenty of choice, and most of us have chosen, with FreeBSD and Linux the clear winners. However, the beauty of the free software movement (and open source) is that our decisions in no way force you to conform -- you are free to take the source code and develope it for whatevre program you like. You even have the right to bitch and moan because the rest of us don't share your platform preference. However, the rest of us have the right to ignore you, or even chortle with amusement, should we in turn find your choice to be as silly as you find ours.

It is called freedom, and it is a wonderful thing.

Re:Not to put a downer on things

[SimonK]

Well firstly, "the open source community" is not a very well defined or coherent thing, but it has one thing in common: almost everyone in it has used some variant of Unix at some point. Thus, almost all open source software is either for Unix, or is very Unix-like. NT-Emacs and Cygwin are wonderful things, but their main function is to make NT more like Unix. Similarly the open source projects proposed for MacOS X and BeOS are mainly ports of X-servers or X widget sets.

Historically, this has come about because the tradition of opening source really began with Unix-like or pre-Unix systems on minicomputers. That whole community coverged on Unix, and the poeple who carried open source through the Dark Age of the 1980s were people from Unix backgrounds. Thus Sun used a variant of BSD, Gnu software is all *like* Unix software, and the only open source software for the early 16-bit home computers made them more like Unix machines.

Its not that we all think Unix is wonderful. It has many problems, the biggest one being X, but people are unlikely to abandon it for an alternative just for the hell of it. The Unix Way of doing various things has developed over quite a period of time, and has proved better than, or at least equal to, most other paradigms in OS design (most importantly the VMS and NT Way). A great deal of useful software, especially in the internet server arena, which is becoming of more concern even for small business and home users, is very much Unix-centric, as is seen in the difficulty of porting it to different OS models (Apache or Sendmail on NT don't work very well).

Now I'm going to start telling you what you really think, so please feel free to tell me I'm wrong ....

Most people who are looking for alternative to Unix, including the various Free Software projects that have been set up to persue this goal, are people who do not know a hell of a lot about OS kernel design, but who see the surface features of Unix - a slow, ugly, insufficiently standardised GUI, and a hard to understand CLI - and want a more coherent and simpler user experience.

My point is that there is no fundamental difficulty in creating these things on top of an otherwise Unix-like system. It needs some considerable thought, and it also probably needs a bunch of existing stuff, such as conventional CLIs, and X (at least in its current incarnation), and possibly even the model of IPC to be binned, or hidden from the user.

In a nutshell: not only is there no chance of getting the Open Source community to adopt a different OS, but there's no really good reason to ask it to. Much better to build on what we have. The much more important problem, in my view, is getting open source people to take on the concerns of "ordinary users".

Not everyone is a fan on Dell it would seem...

[summer]

osOpinion has an article by Tom Nadeau which gives a rather different take on Dell. http://www.osopinion.com/Opinions/TomNadeau/TomNad eau50.html

Re:Not everyone is a fan on Dell it would seem...

[skoda]

If you read the article, you'll se the author has little understanding of economics, the stock-market, or life in the real world.

1) This kind of business-city dealings have been going on for years, and written about in business 'zines, discussing whether it's a good idea or not. If the city thought it would be a raw deal, they could have told Dell "No" to its demands.

2) "...preferring to bring in desperate workers from non-computer fields who cannot find any other kind of work. " Those bast**ds! How dare they hire people who would otherwise be unemployed! And then having the audacity to pay them >$20k/yr in a state with a low cost of living! Really, if these jobs are truly the "third-world assembly work" the author claims, does he really believe that "sharp, well-educated people who have computer experience" would want them? No, they go for the high-tech, high-pay first-world information-age jobs. Duh.

3) "Dell is in the assembly business, not the high-tech business...Dell pretends he is running a high-tech company...But Wall Street knows better...Dell's stock has dropped from 55 to 35. This is what happens when people figure out that [Dell is just a manufacturing plant]. " This is just wrong. Investors have not suddenly realized that Dell is *gasp* a manufacturer of PCs. Hello! This company has been doing the same thing for about 20 years now. Their business model and implementation is not just now being understood. The reason the stock price has dropped is because the commodization of PCs (caused by the low RAM, HD, CPU prices) has created a significant sub-$1000 PC market, which has much lower margins than the mid- and high-end systems, which historically provided the real corporate profits. Dell has been trying to figure out, along with *every other* PC seller, how to make money from low-cost, low-profit systems. Everyone has been dinged due to profit problems in the transition from the past several months. Furthermore, there was a sympathy effect with the drop of MS stock (from the DOJ action) than impacted everyone for a while. Furthermore, investors might also think Dell is making bad decisions with this TN plant, or with the Linux moves. But it's not because they suddenly figured out that Dell builds PCs.

3b) This is a facetious argument regardless, because manufacturing is a non-trivial operation. To create a manufacturing plant with a huge capacity, very low error rate, in which every system can be customized, and the components are fragile, is a challenge. You need smart people who know know about manufacturing process, analysis, finances, and technology. I doubt Dell has a bunch of monkeys grabbing parts from bins and throwing them into cases. I would wager that they have a rather sophisticated assembly line, developed by smart, well-paid people.

4) "It's pretty obvious that Dell's slickmeisters have taken these country boys real good." Yep, that smooth-talking, well-heeled Texas boy sure did pull a fast one on those inbred, knuckle-dragging, TN red-necks. Good argument; when things go wrong, accuse the other guy of being smarter than you. Man, I hate smart people! Always using their intelligence to their advantage!

In the final analysis, the author has two real gripes. The first is that rich Michael Dell got a sweet deal from TN, and hasn't yet fulfilled his promises. Well, if the TN people are so stupid as to give mucho benefits to a company with no guaranteed benefit to themselves, I have trouble feeling sorry for them. His other gripe is that the introduction of a large business has negatively impacted the surrounding areas. That's a real issue that cities and states have been and still are grappling with for a while. But it's not Dell's fault. It's Nashville's fault. They invited them in, and now they have to deal with the consequences.

Re:Building a desktop OS from scratch [OT?]

[zzen]

This might be a little old - but I have an intense feeling that the current OS paradigm cannot hold for long. I will make myself even more clear - MacOS does it wrong, Windows does it wrong, Linux and all UNIX does it wrong, too.

OSes are still made by a bunch of programmers for a bunch of programmers. And the computers are now SO pervasive and the economical loss resulting from bad usability of these tools is so big that I believe (hope) this cannot last long.

Tool was the important rediscovered buzzword in the last paragraph. Because tool is all that computer is. It's not a piece of art, it's not a personal friend -- it is a mere tool. Like a hammer or a pencil or anything. Only a bit more versatile.

The main OS shift is probably happening today with PDAs and mobile devices. They are the first widely used single-purpose computer technology base TOOL. And more functionality will migrate from desktop PCs to tools. Bluetooth will help along this line.

I have a dream where you walk up to a computer (at that time rather a terminal), you touch it - and the usage of it is as evident to you as the usage of a hammer. (No, it's not going around and bashing things with it. :)

Michael Linuxes for business...

[digitect]

...not pleasure. [Can I use Linux as a verb?] I quote:

LM: Did you actually take a look at the technology?

Dell: Yeah. I have a little lab next to my office here, and I got a desktop PC and installed Red Hat Linux on it. I played around with it a little bit.

...and...

LM: OK, how much time do you personally spend thinking about or dealing with Linux these days?

Dell: I don't really have a number that comes to mind for you.

I bet he didn't even partition the hard drive himself. Granted Dell is a hardware supplier, I don't need its CEO to be a Linux efficianado, but some expressions of capability at the top sure would convincing.

His comments sound like many others these days: "We're going to support Linux because everybody else is and we don't want to be left out. But trust us, we know what we're doing."

Hardly "a new Linux evangelist" in my book. Our community needs companies that develop business units whose sole function is to support us (Linux), not just have us tacked on to rest of the OS support department. I think we'll see that the most successful companies five years from now will be those that got in to Linux with both feet today, provided dedicated service, knowledgable expertice, and serious committment from individuals that are compassionate of the cause. If anybody can spot an imposter, its going to be us.

Linux hasn't been easy until now, and we sure aren't doing this on a lark. How about showing some interest before you ask for our money? Sorry Michael, I'm not convinced.

Re:Michael Linuxes for business...

[DaveHowe]

Michael Linuxes for business not pleasure
I'm sorry, but I can't see why this is a problem - Dell owns a *very* large computer company; that he retained control when his company GOT to be large is commendable, that he still has a little "play area" right next to his office is almost unheard of.
However, the end result is DELL computers will ship with Redhat and full support. This *is* a business decision, not a geekish one, so how else can you expect him to treat it?

I bet he didn't even partition the hard drive himself. Granted Dell is a hardware supplier, I don't need its CEO to be a Linux efficianado, but some expressions of capability at the top sure would convincing.
Well, he claims to have installed Redhat himself (I suppose he could have just overwritten the existing partitions set up by a more competent geek, but odds are good he started with a "clean" unpartitioned disk and let the Redhat bootable CD do the partitioning) and redhat usually makes some reasonable default decisions. However, again, I don't see how this is relevant - He got a Redhat up and running himself, played with it for a bit, and was willing to give an interview praising Linux and advertising that his company feels it is ready to be shipped pre-installed.

Our community needs companies that develop business units whose sole function is to support us (Linux), not just have us tacked on to rest of the OS support department.
Why? few if any companies have a separate support department for NT/Win9x/OS2/Novell so why should they have a separate one for Linux? one of the *big* points that is always made about Linux is that it doesn't need as much support, and that *that* support is often better found from the community anyhow. However, even if that *was* true, you have obviously failed to read that part of the article carefully, as you missed:

We've put support in place; we're building a dedicated Linux support queue for customers who call us, just as they would if they had Novell or they had Windows.

which is pretty much waving the "get your dedicated linux-os support like this" flag at you...

Linux hasn't been easy until now, and we sure aren't doing this on a lark. How about showing some interest before you ask for our money? Sorry Michael, I'm not convinced.
I am sure MIchael will have troubled sleep knowing you don't want his products.
--

Re:Building a desktop OS from scratch

[Shotgun]

Do you want a desktop or a personal OS?

The purpose of the OS is to control access to parts of the computer by other parts of the computer. When you sit down at the PC, you can consider yourself just another input device. Why shouldn't the OS control input? This is done through multiple logins.

Now a personal computer would be 'personal'. It would only ever have one operator. But shouldn't the OS still protect your personal computer from unauthorized access? Unless you're planning to have the thing embedded in your brain I can only see requiring a login as a benefit.

Get rid of the CLI? The most productive interface available once the learning curve has been overcome? Why?

I'm glad that you have a perfect power supply where you live so that a journalling file system has no advantage, but the unfortunate fact is that the rest of us live in the real world. The journalling FS is there to protect data. I can see how it would be unnecessary to protect your saved-games, but some of us use our destops for real work.

The GUI is an incredibly heavy piece of software that is near impossible to prove correct with todays techonology. Wrapping it into the OS is stupid. What you propose is to take the most important part of the computer and wrap it with instability. Just because I'm the only person using the computer doesn't mean that I don't have jobs running in the background.

So if we design a new OS, we want have problems with licensing problems or feature bloat? Some people decided to re-write Netscape, a mere application. What do people complain about?...licenscing problems and feature bloat (whether they actually exist or not).

After replying to your post, I get the feeling that you are a troll. "Heh everybody, dump Linux and use something that looks exactly like Windows instead!" Do a little research to find out the advantages of the 'UNIX way' before posting that we should dump it.

Support problems, Hope M. Dell reads this.

[pete-classic]

I just quit Dell Server Support about a month ago. About two weeks later, a friend of mine, who was the only
other person in the department competent to do level 2 support for Linux quit.

Dell server support has plenty of Linux "boot camp" graduates, or people who suddenly realized it is "cool,"
but all of the real experience was driven out of the department. I, really, by Michael J. and I suspect my
friend really had to leave because of Gene B.

I hate to post this, because I am a stock holder, and I am losing my ass right now, but Dell is driving good
people away because they don't "fit the mold." I want to see Dell succeed (because when Dell does well I
make money) but I couldn't, in clear conscience, recommend a Dell server with Linux factory installed
because there is no one to support it.

It seems that today Dell as all "alliances" and no substance.

-Peter

Re:Dell is our friend :)

[(void*)]

For two years, he was interested in Linux. That's fantastic. [/SARCASM]

Re:Building a desktop OS from scratch

[Florian]

Apple is building its MacOS X on a similar foundation as Linux (i.e. BSD), so that disproves your claim. We already have advanced multimedia support (think of ALSA and MESA...) and journaling file systems (ReiserFS) in Linux. Please remember that Linux is a kernel. Nobody would stop you to build a distribution that takes away all Unix-inherited complexity. Some ideas:

• Throw multiuser and files access permissions out of the system. have the user automatically login and work as 'root' just as in windows 95/98/me and in macos. sure, this creates a lot of security issues, but they could be tolerated on desktop machines with dialup-only internet access. by far more simple for unsophisticated users.
• Simplify the file system structure. If multiuser functionality has been removed, it is only longer necessary to have /home, /sbin, /usr/sbin and to have both global and user-specific configuration files. Join /bin, /sbin, /usr/bin, /usr/sbin, for example. Rename all system directories so that they are easily understandable in natural language: /usr/bin to /Programs, /etc to /Programs/Settings /lib to /Programs/Data. Create /Fonts, /Sounds, /Pictures, /Documents, and so on.
• Throw out X11 - since no desktop user needs its network functionality - and replace it with a framebuffer-based GUI.
• Create lightweight desktop applications not with configurability, but simplicity in mind. Avoid redundant functionality (i.e. button bars doubling menu entries). Most of these apps are already there, just port them to a common toolkit (fltk, for example).
• Reduce inter-application interfaces to classic Unix pipes, sockets and libraries. Avoid bloat and slowdown through Corba and similar interfaces.
• Adapt and simplify LinuxConf to act as the system configurator
• Standardize on one scripting language in your "distribution" (for example, Python). Avoid that several bloated scripting languages have to reside on the system just because system utilies (package managers etc.) need them.

I agree that the resulting system will have litte in common of what we know and appreciate in GNU/Linux. But it would be the perfect system for people who don't want to replace the complexity and impenetrability of Windows with yet another type of complexity and impenetrability. Hammering nice graphical interfaces on top of that complexity won't help.

Re:Building a desktop OS from scratch

[Andrewkov]

I agree. Don't give up on the excellent foundation provided by Linux, but what about replacing X with a new, complete GUI system?

I'm visualizing a system which is quite independent of anything else - new apps and games would have to be written specifically for it, and installation of apps would be handled from within the GUI via an installation wizard. These new programs would avoid all the headaches associated with installing software on Linux - they should be installed in a directory under the GUI system's directory, and depend on a new, stable set of libraries, to avoid problems with dependencies. Also, all programs would follow a pre-defined look-and-feel, support the same cut-and-paste features, etc. Make it cohesive from the start.

This GUI system would be written from the ground up to be an all-in-one desktop solution. X-Windows has a lot of cool features, like network transparancy, but it is too complex for the average user, and developement for it is also complex.

This would still be a monumental project, as it ignores all existing software, but it would be a lot more feasable than writing a new desktop OS from scratch.

Comments?

Re:Building a desktop OS from scratch

[dabadab]

Well, you should learn more about UNIX history. It was NOT designed for mainframes. The first versions ran on DEC PDP machines and they were the thing most closely matching today's PCs. And yes, it was meant as a desktop OS, serving the needs of the programmers of the Bell Labs.
I DO use Linux as my desktop OS (both at work and at home) and I am very happy with it while Win9x (the supposed REAL desktop OS) keeps constantly annoying me.
My mother (well, she is not exactly a hardcore computer geek) uses Linux as her desktop OS. She is quite happy with it.
And anyway, I just see no reason to NOT to use Linux as a desktop OS. Could you mention just ONE feature that shows that it should not be used for that?
(Remember, in the old DOS days people whined about having to shut down Linux as opposed to the 'just switch it off' method of DOS - and see, what happened in Win9x)

Re:Building a desktop OS from scratch

[Tet]

Everything has its purpose right? Linux's purpose is to be the best server operating system available

Actually no. Linux's sole purpose is to give Linus Torvalds a Unix that he can run on his PC. That's why it was born, and any other uses are purely coincidental. As it happens, Linux is a pretty good server OS. It is also (and here's the controversial bit) the best desktop OS in the world. No, I'm not on drugs -- I seriously believe that it is the best desktop OS in the world for me. It does pretty much everything I want, and it does it better than all the alternatives I've tried. I'm not foolish enough to claim that it's for everyone yet, but given time it will get there. The features that make it a good server OS don't preclude desktop use. Or are you going to try and claim that Win9x is a better desktop OS than NT4 or Win2K?

Re:Building a desktop OS from scratch

[nevets]

First I must say that Unix was not for mainframes (which is a centralized system) but to be more of a network (client/server paradigm).

You do have valid points, but I must say that the problem may not be with the OS but what we are doing with it. We say that Unix/Linux is not for your grandmother, and I would agree. But *nux is very good with networks. Now the problem is, we are trying to get Grandma onto a network. The Internet.

Now this causes all sorts of problems. I saw one poster a few days talking about how their wife complained about having to log in. "I own this machine, and I'm the only one on it". But that my not be the case if you have an Internet access. You see when you take single user methodologies and put them with network ones, you get things like virus and privacy compromises.

So, we need to educate the average user and maybe tweak the OS a little. But I really think that if you are going to have a connection to the internet, you should have basic knowledge about how to use a computer. The normal analogy is to compare computers with cars. You may not know how a car works, but you definitely know what to do and not do with it before you drive. That's why we have licenses, (although I would say there are those that don't know how to drive). A computer is no different (except for being less dangerous). You should have a basic knowledge before surfing the net. And I'm tired of hearing how Joe Schmoe doesn't want to know anything before using a computer. If all you want to do is play games, that's fine, but if you accept email and go out onto the net, then either learn or accept the consequences.

Steven Rostedt

Innovate it however you want

[Greyfox]

You've got the source. Innovate it however you want.

Want to add obscure nifty features like ACLs and Roles? You can do that, kind of like the LIDS guys and the Linux ACL project.

Want to rip underlying kernel out but keep the rest of the look and feel? Kind if like Debian/HURD maybe?

Want to keep the underlying kernel and get rid of the look and feel? Like all the embedded Linux projects like the Tivo do? Go for it!

Want to do something else? You're free to.

This isn't so much a Linux phenominon as an Open Source one. People have said inventions don't happen until we're ready for them. The world wasn't ready for Open Source until everyone got wired via the Internet. Then it just exploded and companies started to realize that with the playing field level everyone can profit.

Something radically better than Linux may come along at some point (Though I rather doubt it'll come out of Redmond) and people will start switching at that point. Innovation will still happen, but now it'll be a lot more people controlling the direction we go.

Heh!

[Greyfox]

I was at a Linux gathering a while back and someone had some words about VIA: He said "They said they want to be the Dell of the Linux world. Well I think only one company will be the Dell of the Linux world, and that will be Dell." Insightful man. I think it was one of the SGI guys.

History of unix repeated?

[w00ly_mammoth]

And then spaketh the lord, "Let there be unix companies galore, each different from the other, so that the standards can multiply and compete amongst each other."

And the companies thus arose, and they fought each other, with various wondrous types of software, none of which worked with the other. And everyone saw that there was competition, and it was good.

Thus began the ten years of drought, when application developers moaned, "Wherefore am I to write my software? None shall buy it, for each is on his own little island, each separate from the rest." And Gates said to them, "Come to me, ye fools! And ye shall be happy, for all my denizens live under one roof. It leaketh sometimes, but ye shall earn gold selling yer stuff." And the developers all flocked to Gates, and the unix companies continued to fling dung at each other, and it was all as it had always been.

w/m

What do you want from a desktop OS?

[Greyfox]

I want an OS that won't self-destruct if you add software packages on a fairly regular basis. It's not a lot to ask (I think.) Right now I'm not entirely positive that Linux provides this (But then, I REALLY abuse my Linux setup) and I'm almost sure Windows doesn't.

My anecdotal evidence is that my room mate's Windows machine (Which I have as little to do with as possible) seems to become increasingly unstable as time progresses. My Linux machines seem much more resistant to this, assuming you can get the software and install it in the first place. If I were to religiously install everything from RPM, I think it would be a lot more stable. The problem with Linux of course is that my room mate can't get all those games she wants and I'd end up having to set up her 3D card and networking and stuff. And not every average person wanting to run Linux can find a guru to room with.

Re:What do you want from a desktop OS?

[Ralph Wiggam]

I work with Windows machines every day and adding many applications over several months with KILL the performance of a Windows box. I've gotten calls from people who think thier computers are broken and all I've needed to do is remove 25 ancient applications they didn't even know they had. I'll bet that 10 million people have MS Money installed on thier computers and don't know that it exists, much less what it does or how to use it. Another thing that will kill your performance is the tendancy of programs like Office and AOL to launch at startup.

-B

Re:Still not easy to find on the web site

[Shotgun]

It takes more time (and hence money) to set up Linux than it does to set up Windows. Flame away, my brethern, if you must, but 't is a fact, that is to say, 't is fact if you want to set up Linux as a desktop OS (which it was not exactly meant to be).

Let's compare:

Windows:
1 man-day to install.
2 man-weeks to actually get everything working
1 man-hour to make a disk image

Linux:
1 man-day to install
2 man-weeks to actually get everything working
1 man-hour to make a disk image

Where is the extra time for Linux?
Note: Dell doesn't run an install from CD on each machine seperately. They create a master image and copy that to thousands of disk which each go into one of thousands of identical machines. If they didn't do it that way, Linux would win hands down, because you would do the install and then run "Dell's Install Script", while the Windows machine require fiddling with the GUI seperately.

Re:Don't dis M$...BUT

[freeBill]

When you put together some of the things Dell has been saying lately, it's pretty clear he has problems with Microsoft, even if he cannot discuss them openly.

Dell (from this interview, with emphasis added): "I think it definitely has the potential for a lot of change -- and disruptive change. Not so much on our business models, but on other business models. The whole open source movement has the potential to really change the way value is created and distributed in the software industry -- the speed at which applications and tools are developed and deployed. And Dell's the perfect hardware platform to do that on."

Dell (from a Charlie Rose interview at a conference in Paris): "If you had a business that was based on tricking your customer -- which, in fact, a lot of businesses were fundamentally, assumed that the customer didn't have much information and the customer was, in fact, uninformed -- well, that's going away."

When you put these together, it seems to me that Dell is inching away from Microsoft because he sees a big fall in their future. And why not? Look at the MS business model from Dell's point of view:

He can remember the days when PCs sold for $3,000 and the OS cost $15. Today he sees computers selling for $500 to $800 with an OS that costs $85 to $250. And the vendor selling the cheap box has to provide support for the OS, even if it turns out to be a piece of crap.

Somebody's getting squeezed here, and Dell knows who it is (even though he probably has a really sweet deal with Microsoft). He knows the main reason he has to keep selling MS OSes is because of the ignorance of its users. But that doesn't mean he has to like it. And he sees the Internet as working against software vendors who rely on the ignorance of their users.

Sure, he dare not speak its name. But Microsoft clearly fits his description of software companies whose business models will not survive in the Internet economy. We should not be too quick to doubt his sincerity since his conversion (though late) seems to be based on solid values and genuine interests (that's interests as in business-type interests, not curiosity-type interests.

Re:Building a desktop OS from scratch

[Black Parrot]

> Rather than that constantly remake Linux in order to compete with Windows, it would make a great deal more sense for the FSF to create a brand new operating system designed from the ground up to be a desktop OS.

Actually, you don't have to remake Linux to be a desktop OS. All you have to do is put a "desktop layer" on top of Linux. That has the following advantages over remaking or starting from scratch -

• Layering is the proper way to design software systems.
• It requires less work.
• People are already doing it.

--

Re:Building a desktop OS from scratch

[dattaway]

I saw his humorous reference to mainframes too. From what I remember of mainframes, their memory capacity at the time had several kilobytes of RAM. Also, I remember Linus hacking out a new kernel for his 386, because DOS just didn't do it for him. When maddog mailed him a very sweet Digital Alpha, cross platform compatibility then became a priority. But I have never heard of Linux running on an intensive I/O machine such as a mainframe, except in emulation.

Re:No M$ Tax ?

[mirko]

This doesn't mean no tax at all.
And what makes Linux exciting to most if us is the DIY stuff involved. If MD really want to listen to the first generation of Linux users he should take this into account and then deliver OS-less laptops (servers, etc.) instead of ones with specific OS installed.
How much money is MD expecting to charge his customers for a Linux enabled laptop ?
Let's evaluate :

1. Internal R&D rentability, including eventual driver development and staff training
2. Technological survey and distribution optimization
3. possible prices raises on the Windows-powered side as Bill might raise his prices because of such a lack of fidelity. As MD didn't especially speak about a divorce, he might not want to lose its existing customers and then amort these raises by increasing the Linux machines prices.
4. Supplemental advertising
5. etc.

So, if you want my 0.01$ guess, I'd say that the Linux laptop might be as expensive as the NT/2000 or at least ME ones.
--

Re:Building a desktop OS from scratch

[orabidoo]

Instead of wasting time trying to turn into Linux into a desktop OS, why isn't someone designing a new one from scratch? Sorry to burst your bubble, folks, but Linux just isn't designed to be a desktop OS.

because OSs are layered things, and the lower levels (the kernel and drivers) are not anymore "designed to be a desktop OS" than designed to be anything else. The Linux kernel can be the foundation for a perfectly good destkop OS, it's just all the Unix command-line and /etc/configuration stuff that's historically tied to it. No-one's preventing YOU from making a desktop OS on top of the Linux kernel, that doesn't even use /bin/sh. For that matter, no one's preventing you from writing your own kernel either, but you'd better have some really good new ideas to put in it, to be worth the effort.

In the meantime, GNOME and KDE are moving forward, and doing the right thing: putting a pretty user interface on the whole system (which means that the average user ultimately doesn't have to touch the command line unless he actually wants to), while stayign reasonably close to the Unix way-of-doing-things, and keeping interoperability with more traditional Unix desktops (X11 + whichever wm).

Re:not much of a true believer

[finkployd]

As a hardware manufacturer he really doesn't have to be. Be is simply using good business sense. What impresses me is that we have progressed to the point where good business sense dictates that Linux is a good risk for a large business like Dell to invest in.

Finkployd

Dell's think-tank

[zzen]

manager: Gee, Michel, IBM is getting a lot of press for pushing this Linux thing.
Dell: Nah, Linux sux
manager: yes, but it gets a lot of coverage lately. We should really come up with something.
Dell: Nah, Linux sux.
manager: Michael, you will slashdoted.
Dell: Darn, you are convicing. Call our PR department...

Building a desktop OS from scratch

[vertical-limit]

Instead of wasting time trying to turn into Linux into a desktop OS, why isn't someone designing a new one from scratch? Sorry to burst your bubble, folks, but Linux just isn't designed to be a desktop OS. Linux is based off of Minix, which was based off of UNIX. And UNIX was designed to run on mainframes, not on my grandmother's PC. Never did any expect that people would be running solitaire games, Outlook, and Instant Messenger on it. It simply wasn't designed for it -- look at all the command-line entry you still have to do, even with using the GNOME or KDE window managers.

This isn't to say that Linux is a bad OS. It's a terrific OS for servers, routers, and other non-end-user computers. But it doesn't make any sense to try to hack shiny, happy desktop features into it. Everything has its purpose right? Linux's purpose is to be the best server operating system available (whether or not it succeeds is your call), not to battle Microsoft.

Rather than that constantly remake Linux in order to compete with Windows, it would make a great deal more sense for the FSF to create a brand new operating system designed from the ground up to be a desktop OS. Not only would this OS include all the necessary components for a desktop OS (GUI support built in from the beginning, no CLI, journaling file system, plug-and-play devices, advanced multimedia support, etc.), it would eliminate all the problems seen in current desktops -- licensing problems with KDE; feature bloat with GNOME. And right now, there's simple no free OS that does this -- sure, there's BeOS, but it's only free as in beer, not as in speech.

Remember P.T. Barnum's famous quote "You can please all of the people some of the time, or some of the people all of the time?" Right now, Linux is trying to please all of the people all of the time, and that just isn't working. It's time to divide and conquer. Leave Linux to the server market and design the efficient, stable, user-friendly, and most importantly, open-source desktop OS that the world has been waiting for.

Re:Building a desktop OS from scratch

[Samrobb]

Throw out X11 - since no desktop user needs its network functionality

I'm a desktop user, and I need its functionality.

There are a couple of different flavors of desktop user...

At one extreme, there's my mother, who expects her computer to boot up and work. She's not going to configure it, at least not intentionally. She wants a truly simple, iOpener-like system that lets her read and send email and see pictures of her granddaughter on the web. She uses a desktop all the time.

At the other are the sysadmins, developers, and IT folks who understand the hardware, the software, and how it's all glued together. They want X, ssh, and a wide selection of tools and capabilities; and if they can't find what they want, they'll build it themselves. They use a desktop when it's convenient.

Halfway in between are the "power users", the folks who know there's more to the machine than the desktop. They might be comfortable learning a bit about the shell, or a scripting language; but that's not their job. They want options and flexibility, but not too much - just enough to keep it all straight in their heads and get their jobs done. They use a desktop unless there's a very big advantage to using a CLI.

To be honest (and I don't mean to flame you), it sounds like you'd be happier using something like BeOS or MacOS. Why go to the trouble of using a complex, powerful Unix-like kernel if you're immediately going to rip out half of its functionality?

For a couple of reasons...

• To use that "powerful kernel" to run powerful software that makes difficult tasks simple.
• To provide a common platform for interoperability.
• To optimize software for certain types of usage patterns.
• To provide a simple "upgrade" path.

That last one is the real killer. I put "upgrade" in quotes because with the same system being used by multiple types of desktop users, it becomes a whole lot simpler for someone to learn a little bit here, a little bit there, and slowly expose the full functionality of a system by adding or exposing new pieces as they feel the need to.

Re:Building a desktop OS from scratch

[finkployd]

And UNIX was designed to run on mainframes

Bullshit. Only within the last few years has unix been ported to mainframes. Unix System Services (a componant of IBM's OS/390) and Linux are the the only two I know of.

MVS and before it MVT were mainframe operating systems.

Finkployd

System administration vs typical usage

[harmonica]

I'm pretty sure you can teach someone without computer knowledge to use Netscape, StarOffice and KMail on KDE (as an example). You show them how to log in, that there are the buttons to start the program, how to open files etc. It's a lot like Windows.

The problem is to set up the system. You must give them a preinstalled Linux and they must have someone who will fix problems for them. If they don't have a tech-savvy friend / neighbor / relative, they're screwed. And frankly, I don't think the people who need that much support (e.g. for every icon that disappeared) are able to solve the same problems under Windows.

Dell's take on the NC...

[SuperKendall]

LM: Do you think that'll come about in a Larry Ellison-type "NC-take-2" Linux-based personal appliance?

Dell: No. As you get more bandwidth the computing power will expand both at the edge of the network and at the center.

Larry's model is that it expands at the center, but the edge gets less and less powerful. I think that actually both will get more and more powerful and as you get a faster connection to the internet, the data streams become more and more complex and there's richer and better data that gets processed and dealt with at the client. You can kind of see this a little bit if you talk to people who have DSL connections or cable modems in the home and you say to them, "Now that you've got this really fast connection, is your computer fast enough? Will you use a slower computer now that you have this really fast connection?"

Nobody wants to go to a slower computer now that they have a fast connection; they want a faster computer because the bottleneck shifts. Right now for a lot of users -- at least in small businesses and at home -- the bottleneck is the line, not the computing power, but when you get a fast connection, the bottleneck shifts back to the computer itself.


I agree with Dell that clients on the edge will grow in power just as clients in the center do. However, I don't think that's at all incompatible with a NC view of computers - NC's will be powerful devices, but more specialized.

For instance, I would love a small computer in the kitchen to look up recipies, and a small computer by the TV to be able to look up related things during TV programs. The kinds of devices I want in each place have a fairly different set of requirements though, which would best be served by an NC type of device. I could see such devices outnumbering normal computers in most households before too long, once easy household connectivity is figured out.

Also, now that I have a DSL line I do not find my computer (only a P450!) to be the bottleneck. I think my bandwith would have to increase by an order of magnitude or more to start thinking that.

What I found more interesting in the interview

[w00ly_mammoth]

Yeah, the chicken+egg problem gets solved when a major vendor gets behind the software, but what I find odd is how it happens. I mean, linux has >20 million users, which is about 10% of the Windows installed base. Now, if you were a CEO, wouldn't you push your company to support it, thereby gaining 10% more customers and getting an edge over your slower competitors?

From the interview:

Then I started going out on the Net and searching for "Dell and Linux," "Optiplex and Linux" "Dimension and Linux," "Inspiron and Linux," "Precision and Linux," and Boom! You'd see hundreds of thousands of user who had figured out how to use our products with Linux, and we hadn't done anything to help those people either.

In reality, though, (and this is something I find very, very surprising) many/most hardware companies are far slower than Michael Dell at catching on. Recently, I bought a logitech cordless mouse, but only after a frustrating search to find out if it works for linux. Now, logitech could hire a couple of web developers to put that info on their site, and maybe get a couple of their programmers to test stuff for linux. Bingo. You're helping out 20+ million customers you weren't before. Why don't CEOs see this?

My guess is that Michael Dell is probably way ahead of the curve, since he actually bothered looking at the search queries, installing linux himself, etc. Maybe most CEOs wait for a Gartner analyst to draw pie charts and wait for their underlings to draft a business plan before they get the point. Any idea how this works?

(As an aside, it's very revealing that Michael Dell actually saw *hundreds of thousands* of customers helping each other out before the company even realised there were all these customers using Dell on linux. Is it any surprise people prefer looking for product info from other users instead of the "official" company site? )

w/m

Re:Still not easy to find on the web site

[uebernewby]

The reason Linux boxes are almost as expensive as Red Hat boxes are:

• OEM licences for Windows are much cheaper than the licences you buy in a store, probably very close to the cost of an OEM licence from Red Hat (which is not exactly the cheapest distro anyways)
• It takes more time (and hence money) to set up Linux than it does to set up Windows. Flame away, my brethern, if you must, but 't is a fact, that is to say, 't is fact if you want to set up Linux as a desktop OS (which it was not exactly meant to be).

Having said that, I'd of course much prefer them not to install anything, be it Windows or Linux, because I'm one of those control freaks who'd rather keep everything in their own hands. I'd much rather put stuff on my harddisk in a way that makes sense to *me* (and probably no one else) than to dig through an entire 13.7 Gig HD trying to find out where Dell has put the sources to the kernel.

Of course they see linux desktops in the future

[Elvii]

Reading the interview, they say they want to take the server market over from sun, imply they want the desktop market from MS, etc. What they really see, imho, is selling computers and making as much as possible, they're a business, after all. So they say linux can do desktops, partner for support or provide their own support, and the chicken and egg problem half solves itself, with linux desktops in shiny TV commericals. Poof, linux has major vendor backing, and their vison works. I'd really not be suprised if it works out this way.

Myself, I'd not mind as long as they charge less for a linux install than for a win98 install, and have options like order 1 year phone support plan or just a cheapbytes cd shipped with the box. Myself, I don't want or need linux support that way, i can get any info I need from the net. So please make it an option, Dell.

bash: ispell: command not found

Games = desktop success

[Dan Hayes]

The real trouble with pushing Linux onto the desktop market as quickly as possible is that the infrastructure required to develop games easily is still in the process of development. Why games? Because games are where mindshare is in the eyes of Joe Sixpack buying a PC for his kids. Sure, he may claim that it's for "educational purposes", but if it doesn't run games as well then he's not going to buy it.

Whilst the latest versions of XFree86 attempt to go further in what hardware is supported and what features can be used, there is still no unified framework for the kind of features required for games and other multimedia applications. Sure, there's OpenGL and now OpenAL, but these are very much a work in progress under Linux, and even then rely on the goodwill of manufacturers in making driver code available for a system where they won't be making any kind of return.

The reason that there are so many games under Windows is that it is, relatively speaking, easy to create them. DirectX provides a unified framework for integrating graphics, sound, music, input devices, network play and more in a single package, allowing designers to concentrate on what matters - the game itself. The difference is very real, and can be seen in the amount of time it takes for Loki to port a game to Linux.

It is safe to say that the importance of games cannot be stressed enough in the public's view of how desirable a desktop system is. And until Linux delivers a unified framework for creating games a la DirectX, it's success on the desktop will always be limited.

Slashdotting as a propaganda tool.

[Ungrounded Lightning]

manager: Gee, Michel, IBM is getting a lot of press for pushing this Linux thing.
Dell: Nah, Linux sux
manager: yes, but it gets a lot of coverage lately. We should really come up with something.
Dell: Nah, Linux sux.
manager: Michael, you will [be?] slashdoted.
Dell: Darn, you are convicing. Call our PR department...

Interesting point...

Are there any Windows-related news media that have such a large following that mere mention of a URL will bring down a server? I don't THINK so... (Otherwise it would be called "WinNewsing" or some such rather than "slashdotting".)

So perhaps this can be turned to Linux's advantage...

"Look, guys. There are so many Linux fans on this one news board that just the mention of your site there will bring so many hits that you'll think you're under a DOS attack. Don't you think there might be a market for a linux port of your ?"

(Yes, I know there's some *BSD, BeOS, and Mac fans here too. But I still think it's a fair characterization.)

Re:Building a desktop OS from scratch [OT?]

[Samrobb]

Because tool is all that computer is. It's not a piece of art, it's not a personal friend -- it is a mere tool.

Leaving aside the fact that traditional tools have been used in art (not just to create art) and that an experienced professional craftsman does come to know "his" tools...

A physical hammer is akin to the original "word processors" (remember those beasts?) or a non-programmable calculator. Each is built for one task; you may be able to use it to do something else, but that's purely an accident of form and design. There's a reason why word processor machines are almost extinct - we (as a culture, possibly as a species) have decided that advantage of increased flexibility is worth the cost of increased complexity.

Computers are the first truly general purpose tool that mankind's invented. Imagine a chunk of matter that responded to your thoughts; if you think clearly enough, and in the right way, you can cause that chunk o' stuff to take the form of any tool you've ever seen, or any tool that you can imagine.

That's what a computer is. Not a hammer - but something that can be a hammer, or a screwdriver, or a socket wrench... as long as you know what you want, and how to turn it into the tool your interested in. The complexity is not in the use or the particular shape that the tool takes, but the fact that it can take many (any!) shape you can imagine, either perfectly or imprefectly suited for the job at hand.

Modern tools already do and will continue to incorporate elements of computers. That's progress for you. But don't confuse the technology that goes into the tool with the tool itself. Just because it's used in a simple tool doesn't mean that it's simple, or that more complex uses of that same technology are wrong.

Re:No M$ Tax ?

[mirko]

In June I bought a laptop. No brand here. Some say it is a Gericom but I didn't buy it for this reason.
I bought it because all of its components were standard (and also because it was twice as less expensive as a Dell laptop) :

1. PIII600
2. ATI Rage LT Pro
3. Maestro 2E
4. (and, BTW, a zone free) Toshiba DVD
5. Tulip compatible Network card

I installed OpenLinux 2.4 on it and I have to say that this works.
So, as I might expect most people here just spend more time choosing a laptop than a car, I am sure that having a dedicated Linux install for their laptop is not *that* crucial. BCNU
--

CNN

[Black Parrot]

CNN's Headline News (the one that recycles every 30 minutes) just did a spot on Linux. They must have spent around two minutes on it. Among other things I noticed,

• the introduced the story with a big Linux logo,
• they pronounced it right^w the way I do,
• they were apparently filming at the show, though they actually focused on content rather than on the event,
• they used the de rigeur "scrappy upstart" description,
• they said it was popular because of its "amazing reliability" (sorry; forgot the exact words),
• they showed barricades labeled "Windows Free Zone",
• they showed systems running GUIs, while voice-over talking about Linux,
• they had a bite from an Eazel guy, mentioned their Appleness, and showed a bit of Nautilus,
• they said Linux was was used on 10% of servers and 4% of desktops,
• they showed an xterm (or the like) with an "uptime" display of 41 days (though I doubt that the uninitiated viewer would have been able to figure that out),
• they showed part of MD's speech (see, I really am on topic!).

CNN repeats some of these spots several times a day, so you may be able to catch it if you're inclined to.

The most interesting thing about this is that it puts Linux in front of a very large audience, and they portrayed it in a very good light while doing so. You may have people asking about that Linux thingy you run, if they saw this.

--