Slashdot Mirror
Gartner Group Squints At Future OS Growth
Icebox writes: "Cnet is offering up this bit from GartnerGroup that includes their predictions for the next few years in the OS market. Their predictions are aimed stricly at the business side of this but it is interesting to see how their ideas stack up against what Slashdot's readership expects. Pay particular attention to Factor #9."
While many users may consider linux more powerful, etc, they often forget to consider the learning curve. Unix systems are harder to use, lets face it. Linux, BSD, etc, will continue to be powerful, and crash less often, but windows being more "assume the user is an idiot when programming" oriented, it will continue to have a good part of the market. IMHO, unix should be used for servers and for the workstations of advanced users, at least until the interface can become more standardized and easy to use.
I am !amused.
From the end of the article:
The total-cost-of-ownership argument on behalf of Linux will disappear (Unix platform vendors like IBM and Sun already offer their Unix OSs at virtually no charge).
*COUGH* Microsoft *COUGH*
This space for rent. All reasonable inquiries will be entertained at proprietors discretion.
They don't forecast the rise of the Amiga? What? Didn't they even attempt any research? :)
10) After the Justice Department demands the breakup of Microsoft, Gates and Ballmer sequester themselves in Redmond with thousands of the company faithful.
Negotiations break down after several days, and a fierce assualt by BATF troops is quickly repulsed by the computer controlled defenses at Microsoft.
A second assuault two days laters succeeds, and Gates and Ballmer are led away in handcuffs. The central defense computer is found with a BSOD.
The big hardware players like IBM, HP and Compaq are likely tired of cow-towing to Bill & Co. every time they sell a machine. Linux means they can free themselves from any outside control in product development. It also levels the playing field for everyones hardware. Lastly, Linux is buying them mindshare from the people they need to support thier products - sysadmins, delvelopers and IT personnel in general - by introducing them to a *NIX variant, they get a toe-hold for thier proprietary versions of *NIX. No wonder they're all solidly behind Linux - it's good business.
"Depression is merely anger without enthusiasm." - Anonymous
Of course, the 2.4 kernel is due soon. But that was true last October, too. ;-)
The author of this report seems to assume that eventually the market will be divided between 2 or 3 dominating OS's, but that doesn't seem plausible. I mean, if all, or even many, users were likely to take the path of least resistance and go with the biggest, most widely supported systems, Linux itself would not have gotten very far.
Much as I like Linux, I don't think it alone will satify those who reject Windows.
--from a not-so-secret fan of BeOS
I think we shouldn't pay attention to the figures released by the Gartner group. Why?
Because you can't aim a prediction strictly at the business side. The business side depends on the customers opinion and the customers(most of them love Linux) opinion is not easy predictable. Maybe Linux is not as userfriendly as it should be, but this is a question of time and another issue.
Predictions of companies like Gartner are also quite dangerous, because Gartner (or other Companies) may support special companies and try (with their predictions) to take influence on the market.
new MS strategy. A way of looking at the
have given up their strategy of trying to convert heterogenous systems
over to MS only systems, and have attampted to go for peaceful
coexistence.
It's a dangerous strategy if OS X takes off, since now there is a
UNIX-like OS with a first rate desktop / GUI development environment.
According to the article BSD will cease to exist. Or at least remain completely unnoticed by journalists.
134340: I am not a number. I am a free planet!
Linux shipped revenue in 2005 will approach 20 percent of the revenue of Unix and 17 percent of that of Windows.
This, of course, says nothing about the number of units shipped or the number of boxes with Linux installed, as the "shipped revenue" of Linux can be as low as the purchaser desires.
well-established OS environments that continue to benefit from the research and development resources of Microsoft, Sun Microsystems, Hewlett-Packard and IBM,
As nearly as I can tell, Linux benefits from their research as well. The lag is a little higher, but the 'geek effect'* means that Linux benefits from any research that benefits any other operating system.
Much of the beneficial backlash Linux has gained at Windows NT's expense will dissipate by 2002, forcing the Linux community to refocus and re-energize its campaign for wide corporate acceptance.
Yup. Market growth of any non-MS OS must eventually slow, as only hardcore MS-ites will continue to buy MS products, and everyone who's convinced that MS is non-optimal has already stopped buying MS products.
Interesting article, but really only worthwhile from the point of view of business people and marketroids. Every business in the world could drop all support of Linux, and guess what?
It would still exist and grow and improve.
*geek effect: Geeks work in technical places. Geeks like Linux. Geeks learn new technical things in technical places. Geeks say "Hey, let's add this to Linux!"
This report says nothing about how Apple products will effect things. Don't forget, way back in 1979 Steve Jobs brought the personal computer to the mass market. Apple is about to release a modern graphical operation system with a true Unix core. My personal prediction is that Apple will steadily gain marketshare in three critical computing markets: small to medium scale data serving (webserving), home clients and business clients. I see Apple's (not Linux's) marketshare equaling Microsoft's by the year 2005, if not sooner.
Gee, machines with a decent OS. What a concept... (Though I saw no mention of OS X (Darwin) in the mix, but that's essentially correct too.)
Look for consumer machines to run Linux (PC style) & OS X (Mac Style,) while Windows gets squeezed further and further (unless they give up and go with Linux [not {expletive deleted} likely? To survive, M$ could mutate & metastacize. They already make more money from their investments than their OS!] like Apple went for BSD/Darwin... )
Now are we surprised?
MSBPodcast.com The opinions expressed here are my own. If you don't like 'em... Think up your own stuff.
Another thing they ignore is that computers are not only used for productive work. This was to be expected from Gartner, them being analysts, but still the fact remains that Windows is a better platform for games and will be, especially with the anticipated rise of technologies like the X-Box.
A final thing they ignore is that OS choice appears to be less deterministic than one thinks it should be - while a number of people know that non-MS platforms are better for some applications, they still use Windows, and Windows does have its advantages, some geeks would rather be hanged than use it and stick to sometimes less comfortable and/or less powerful solutions.
As a state gets corrupt, its laws multiply; the most corrupt states have the most numerous laws. (Tacitus, Annales 3:27)
I just wanted to point something out here. While I agree this does seem to be a good endorsement for Linux's role, I have to wonder about some of these conclusions. One in particular was interesting to me:
7. The performance advantages of RISC over Intel-based servers will decline by about 20 percent to 30 percent each year [..]
They mention nothing to back this up, and it just plain doesn't make sense to me. This kind of conclusion without any presentation of a reason behind if makes me take this whole report with a grain of salt.
Using your sig line to advertise for friends is lame.
1) Microsoft .NET will crash and burn leaving a crater the size of Texas after having millions poured into its development. Nobody wants remote storage of data and software. Look at the Javastation for an example. Microsoft will loose more marketshare as they scramble back towards Windows NT based OSes.
2) Linux will still be around, but the distros will have to rethink their strategies. RedHat shot themselves in the foot with that garbage they released as 7.0. Cutting edge is one thing, stability and compatibility are another. Debian is on the right track, maybe give them a slight boost in the release department. Stability is one thing, staying up to date is another. I predict that IBM and other hardware vendors of the like may release their own versions of linux and give back some to the community.
3) Linux moves more into the desktop. KDE2 is nice and may sway more companies to migrate. Look at non-profits and little-profits that are computer savy to make the migration. Yes, the learning curve is greater, but the cost savings alone allow for hiring Joe Highschool for occasional maintainance and consulting. Yes even the non-profit discount still requires a decent cash outlay for MS, etc.
4) Freebsd and the like make heavy inroads in the server area. I'm still waiting for a distro of linux to install on a server that doesn't require gigs of drive space. Yea, I know you can do min installs and then install what you need, but xxxBSD just seems cleaner to install. And you have to love the ports.
Just remember, Linux and Unix are now at the place now where MS was in the early 90s. There OS was powerful but difficult to learn. Eventually Linux and Unix will get to the point where the normal user never sees the shell prompt.
IOW, Microsoft less, linux/unix more.
What the Gartner Group doesn't understand about TCO (or at least fails to recognize in this report) is that the price of the OS license is only a tiny fraction of TCO. Here's what I think really matters in the calculation of TCO:
Primary documentation. Is the operating system well-documented? Are the manuals (or man pages) accurate, well-written, clear and concise? Linux doesn't score particularly well in the man page department, unfortunately--I think it loses points here. However...
Secondary documentation. How good is the secondary, user-contributed documentation out there relative to other operating systems? Are there well-written, current HOWTOs, user guides, tips and tricks, etc out there? If so, how well are they organized? Linux scores very high here. Other UNIXes are not so lucky. Windows has a lot of users out there--there are a lot of helpful tips on the Net but the documentation is not organized well.
Source code availability. It's an old adage that "the source code is the best documentation," and it's hard to argue otherwise. If I really want to know how a certain function call works, or how the kernel is talking to the hardware, I can dig out the glibc or kernel sources and see for myself, and be 100% certain of the accuracy of my conclusions (well, as long as my understanding of the code is correct). Once again, Linux and other free OSes score high here. With commercial UNIXes and Windows, you have to trust the documentation -- which is often not 100% accurate. In addition, if there's just a bug that needs to be fixed, it's often much easier to fix it yourself if you have access to the source code. Waiting for Microsoft or a commercial UNIX vendor to get a patch out to you can be painfully costly if your product or service's success is dependent on your software vendor's turnaround capability.
In summary, what I think really counts the most towards TCO is the relative understanding of the managers who plan system rollouts and the administrators who maintain them once they're out there. I've seen way too many Windows NT admins out there banging their heads on their desks because something doesn't work the way the documentation says it will, and they have no hope of getting it right because Tech Support doesn't know either. This leads to a lot of costly experimentation (labour is much more expensive than hardware) to make things work.
On the other hand, while Linux documentation isn't always perfect, it's plentiful, it's reasonably organized, and if you just can't find what you need from the HOWTOs or the man pages, you can always fall back on the source code.
Over the next five years, a large number of recent graduates who are in sysadmin positions will start to rise to positions of greater purchasing power in IS departments. Many of these people have grown up with Free operating systems.
Additionally, new recruits into corporate IS departments will also have had significant experience of Free operating systems at University.
Together, this means that a lot of the traditional barriers to Linux/*BSD in the server room will disappear.
Coupled with the increasing quality of desktop tools for X (Gnome, KDE, StarOffice, KOffice etc) this *may* cause a gradual acceptance of Linux etc. on the corporate desktop.
Happy days ahead.... - Mind you, I have been wrong before, and the Gartner Group are not exactly perfect.
This report was aimed at your boss, not at you. Apple is carving out its own niche and it is NOT in the boring 9-to-5 world of the work office.
Apple can rule the living rooms and professional home-offices with style, dash and panache.
If you're enjoying something, do you want to wrestle with failings of the OS, or would you rather just enjoy it.
If you're creating something or advising someone, do you want to wrestle with failings of the OS, or would you rather just get it done.
Linux boxes will rule the garages, "we mean business" home-offices and tinkerer's shops.
If you're tinkering at something you want to do, do you want to wrestle with failings of the OS, or would you get something accomplished.
MSBPodcast.com The opinions expressed here are my own. If you don't like 'em... Think up your own stuff.
This is a possibility. It is also possible that they are trying to displace java and take all of java's promises away from sun. Right now, Sun is poised in a very interesting position, and I am sure that MS strategy revolves around sun to some degree.
It's a dangerous strategy if OS X takes off, since now there is a UNIX-like OS with a first rate desktop / GUI development environment.
OS X is awesome, simply put. It put *BSD right in front of the most GUI driven group yet.. the Mac users. The OS itself is great but the cost of Apple Hardware will keep it from spawning into the next Good Idea(tm). Comparably equipped PCs costs about 1/2 the price of a Mac. I want a G4 cube in all its glory running OS X, but the cost of the thing plus its non-upgradabilityness (word?) pushes me away. I think this is also what will keep OS X from overtaking the desktop as it very well could.
"You'll die up there son, just like I did!" - Abe Simpson
There is no reasonable defense against an idiot with an agenda
:wq
Thanks
Bruce
Bruce Perens.
At one point during that time, they predicted that four Unixes would survive. I believe the winners were Solaris, HPUX, Digital Unix, and SCO. That's right; March 1998 or so and Linux was NOT EVEN MENTIONED AS A PLAYER, much less a survivor.
Now, they sprinkle notes like "0.7 probability" throughout their predictions, so they have an out, but one would rather they show more of their work.
--
So then even if you have the same 50/50 split between *nixes and windows, I think you'll see Linux slow start to gnaw away at the propertiary unixes. Eventually as these OS's lose market share they'll divert their existing IP and research efforts into Linux (see also SGI). With each fallen Unix variant, more energy will be diverted into making Linux better.
So, if my theory holds up, not to far down the road will be down to four major enterprise OS's: Windows 2000, Linux, Monterrey, and Solaris. So then who's next? Monterrey. IBM has invested much in that product, but they've also invested heavily in Linux. Why waste resources on two development efforts when one is clearly growing and beginning to look like a dominant player. So, monterrey fades, and you are down to three.
Sun has been downplaying Linux on big systems, but as this all evolves, if Linux keeps moving full steam ahead, those arguments will cease to be plausible. Sun may resist eventually, but with the majority of Enterprise unix work running on Linux, it will be hard to argue the business case for backing Solaris any further.
So then it will be just Windows vs. Linux. I dunno, come back here in 10 years and see if I'm right
---
This sig has been temporarily disconnected or is no longer in service
Apple's resurgence won't significantly affect the markets that this study predicts.
Apple's only server-based products, OS X Server and AppleShare IP, aren't targetted at the high-end server markets that IBM, Sun, and Compaq are jousting for, and at the low-end, neither product has the installed base of Windows NT or the 'geek chic' of Linux. OS X Server and ASIP probably will have more resilience to migration than Novell Netware, since a company wouldn't invest in either Apple product without a specific reason (say, the ability to NetBoot Mac clients from an OS X server, or ASIP's native support of AppleTalk). Until those needs can be met by a competing product, it's unlikely that current installations of OSXS or ASIP will fall.
On the other hand, Apple can remain a very successful and profitable company without selling a single product in the high-end server market. All those servers have to have somebody to send data to, after all...
--
Another blow struck for decisiveness...or was it clarity?
They seem to have completly forgotten that IBM is on the Linux bandwagon. Supporting it for everything from laptops to S/390 (sorry, z/390) systems. That ALONE is going to give a huge boost to both IBM and Linux if they continue along this policy.
Finkployd
Too bad they think that the Unices can match GNU/Linux in TCO, just because the vendors give away their software for `virtually no cost'.
:) Oh well, maybe this time they are less far off than last. In a few decades they will maybe even learn.
Why is it so difficult to understand for reporters that the cost of acquiring the software for a GNU/Linux system has nothing to do with the total cost of ownership ? That gratis software does not necessarily mean that bugs get fixed within the hour, etc.
No wonder they are having a hard time predicting the Free systems growth
the report doesn't really seem to appreciate the complexities of free software and the unix world - they focused totally on commercial unices and linux, forgetting about freebsd, openbsd, etc... additionally, there is no account for the effect the open source has on buying decisions. sure, AIX, etc. are virtually free with hardware purchase, but that doesn't mean you can re-compile the kernel for smaller memory footprint, fix bugs in-house, or any of the other options open to users of free or open source software...
Much of the beneficial backlash Linux has gained at Windows NT's expense will dissipate by 2002, forcing the Linux community to refocus and re-energize its campaign for wide corporate acceptance.
Gartner factors in the transition to windows 2000, but clearly has failed to factor in the replacement of the Windows NT backlash with newer, improved Windows 2000 backlash. Even more users upset that previous versions!
"Pinky, you've left the lens cap of your mind on again." - P&TB
"I can see my house from here!" - ST:
It's this idea that makes Linux act like an amoeba. It constantly changes shape and slithers around into all kinds of new and unexpected market sectors. I'm constantly surprised and gladdened at the vast array of applications Linux is put to. That's what you get when you have access to all the source code: amoeba-like flexibility. You can't have this in a commercial product. Vendors want to sell you a black box and not try to look inside it. Commercialized software is inflexible.
I expect someday soon Microsoft is going to wake up and realize that Linux psuedopods have occupied all the markets they want to enter. Even worse, big nasty psuedopods are slowly slithering and inching their way into all the markets Microsoft now occupies. They also won't be able to fight it because it's not possible to truly fight such an amorphous enemy. They best they'll be able to do is whine about "that big slimy monster" that has no head to chop off. But we Linuxers will just grin and keep pushing the psuedopods forward. Maybe the Linux mascot shouldn't be a pengion, but an amoeba.
Macophiles rarely seem to recall the real reason why the Apple computers, which were far superior, fell to the Dos/Intel machines: The fact was that the Win/Dos/Intel machines were alot more open, while the Apples were pretty much black-box.
The reason Wintel now feels a threat from Linux is that Linux is to Windows as Windows was to Apple (In terms of openness and accessibility).
Unless apple decides to really open their box and allow anyone to make hardware/software compatible with their's, they have no chance of taking over the market whatsoever.
Toppling the WinTel hegemony is tough- akin to overthrowing the telephone company, or competing with the interstate highway system. The only way another O/S could do that is to really be shockingly better for the tasks a desktop user want to do. Ironic that success in the business arena depends upon how well a platform supports games.
Gartner's clients are CTOs and managers who quote Gartner reports in order to justify pet projects. It works like this:
1. Techies play with cool stuff.
2. The techies start whispering into the ears of their managers about the cool stuff they're playing with. The techies know not to challenge the status-quo too much or they'll be ignored, so the managers only really hear about moderately cool stuff.
3. Managers (many of whom are has-been techies) start to daydream about the cool stuff the techies have mentioned. Some if it sounds like it might be useful, but of course the big boss (CEO, CFO, etc.) will never go for it, oh well.
4. Gartners asks the managers and CTOs what they've been thinking about.
5. Gartner produces a report that reflects what the managers _would_like_to_do, but don't really have the guts for.
6. Managers buy Gartner reports and use them to justify their pet projects.
The conclusion: Gartner is really reporting techie opinions, filtered through a powerful "you can't handle the truth" lens and contaminated with strange manager ideas.
What will happen now is that all the managers who were dreaming of Windows installations will keep doing what they were doing. All the managers who were dreaming of Linux will have some ammo to justify jumping in with both feet.
If the techies in the organization like Linux, the Windows projects will fall strangely behind schedule while the Linux projects will go surprisingly well (it's amazing how happy techies make a project go better). In two years, shortly after the managers have noticed that their Windows projects are going nowhere, Gartner will report that Linux is suddenly the greatest thing since sliced bread.
It is tempting, if the only tool you have is a hammer, to treat everything as if it were a nail. - Abraham Maslow
Unix is much easier to work with and maintain for a server OS, but Windows is better than linux for a desktop. Since the article was about servers, I guess you should read the article next time. :oD
Mas vale cholo, que mal acompañado.
RE: "TCO advantage to Linux will disappear..."
Linux's big strength isn't just FREE AS IN BEER, it's also FREE AS IN SPEECH, and until I can read kernel source for AIX, Solaris ET AL, I think Linux and FreeBSD have defininite advantages not mentioned in the Gartner article...
--- Jump!! Fire!! Bullet time!! - Lego version of the Matrix
I disagree. When I was at University (left 9 years ago) I wasn't aware of *any* free operating systems. Certainly none that were 'industrial strength'. Please let me know which OSs you are thinking about. (I did Maths, so YMMV)
We used SunOS on workstations, BBC Micros (I kid you not) and a chuffing great Control Data monster running something nasty (but it compiled and ran FORTRAN77, which was all we needed).
My point is (and this comes from experience of hiring people) that recent graduates in CS *all* run free operating systems out of choice. These operating systems are now totally capable of earning their food in the datacentre. I am *seeing* this pressure to adopt Linux/*BSD where I work.
Share and Enjoy.
For example, and I wish I saved it, I remember quite well all those nifty graphs around 1989 showing total installations of OS/2 and how they'd overtake MS/DOS, Macs, and anything else you could think of within a few short years.
So, I've committed a /. crime (and tradition) of commenting on the article without reading it.
When I run helix-update, I get all the newest stuff in a neat little window. It gives me warm fuzzies. True, I only get a list of packages that have been blessed by Helix-code, but all I do is double-click on the little soda-can on my desktop. When the smoke clears, all the neat little apps are installed, and I can access them all from my GNOME menu.
Having said that, what's *really* holding Linux back is NOT ease of use. We're there as far as that's concerned. It's strictly a matter of application availability, ie: We either need more support from mainstream application vendors, or OUR applications need to BECOME mainstream. Either would be fine, and either would totally solve our problems. Ease of use (though there is always room for improvement) is yesterday's news.
--- Tao
This report by Gartner Group, in my view, is flaw. By 2005, the majority of users will treat the OS in the same way as they treat virtual memory today -- i.e.: they will ignore it and not known that it exist.
Application and "agents" are all that will matter.
-- George
Karma stuck at 50? Add 2-5 inches.. err.. 2-5x Karmas Count to your pen1es.. err.. Karma all naturally and private
The OS discussion is no different than the game console discussions. It boils down to the software. What "Linux"(Why does everyone here say Linux? My business sells BSD systems *shrug*) is alot of good applications that can be used in business. My company has a few projects, but no investment capital so these are going slow... 1) Bootable CD, it starts off bsd, locates drive, formats, runs BSD off the CD, and data on the drive. It will come with Apache/PHP/PostgreSQL installed. Now on some CDs we will put different versions of our software. Our main one being Glimpse, which is a business management/POS software can be included in the PHP scripts.
So you pop the CD in a computer, boot it up, type in its IP address. And bam! Software installed, go around to all the clients web servers, make the home page be the IP of the server we just installed, and they run all the software from there. There is a profitable world out there, but Linux hits a big hurdle in its license. FreeBSD uses the BSD license, which is a usefull one. Thats why apple is using it in MAC OS. Linux is free to use, but making money off it? I dunno, thats why we use BSD and PostgreSQL, less licensing issues.
Sorry if I offended anyone, but "Linux" as it stands sucks, and everything with it. Everyone is so busy doing nothing, or coding useless features, its disgusting. You have to make devices work, not by recompiling a kernel, but autodetected, and installed on the spot. You have to make standard documentation, (check out the manual on www.freebsd.org as a start, but even that is too technicall at points), you have to make it so the server can be set up without ever seeing a command prompt, steps being explained logically, and lastly developmental tools. Last I checked, I was the only one who even submitted a php editor to www.zend.com, (someone coincidently downloaded that and make a GREAT version of it, when I finish moving and get all my computers out of boxes, I'll post the link to it). PHP could propell the "Linux" world if developers could pop a CD in a machine, grab another computer on the network, open the web browser to the server, and start working immediatly.
TurboRoot GS Data Design bleach@theshop.net is my personal email if you are working on a PHP development library.
I think originally it was designed as (among other things) a way of running multiple OS "kernels" at once while Mach handled all the device driver interfaces and things like that.
Anyone know if this means you could run both OS X and Hurd simultaneously on the same machine? (odd thought but it might be interesting for Hurd development.)
This space for rent. All reasonable inquiries will be entertained at proprietors discretion.
Start with a little ancient history:
Linux will hasten the demise of weaker OS variants, such as OS/2, NetWare, older NT versions, SCO OpenServer and UnixWare, SGI Irix and other nondominant legacy OSs.
Then add some utter stupidity:
Revenue for AS/400 and S/390 systems continuing to decline slowly through 2005 as in previous forecasts - both AS/400 and S/390 will remain viable investments throughout the next five years -- For what, pre-y2k accounting packages that drive dot-matrix printers?!
Plus some beating-around-the-bush:
Solaris momentum "speed bump" in 2000, with increased discount levels until next-generation UltraSPARC III systems appear (especially to replace the UE10000); -- can you say "just don't use the 400 and 450MHz processors with the cache bug?"
Finally, top it off with a buzzword-gasm:
Countering this trend will be accelerating server deployments for e-business and services, Web portals, intranets/extranets, application service providers and Internet service providers. The increased storage needs of business-to-business and business-to-consumer activity will drive torage subsystem costs higher in a ratio to total configuration costs -- blah blah blah
Way to go Gartner Group!
> do you think that the IT people want to fuck around for hours on end trying to get a simple Windowing env. to work?
In my business environment, they fuck around for hours getting Windows (NT) to work. Then fuck around for several more hours trying to lock it down so the users can't do anything to break it. Then they still have to worry that people won't fill their C: drives with MP3s (which can cause SMS updates to break when they run out of space).
Once they've got everything as they like, they just ghost it to all the other machines. It's just as easy to clone/lockdown under Linux.
And, BTW, I installed KDE2 on Tuesday on my Debian machine. apt-get did all the work, including handling dependencies & building menus for all the packages on my system.
-- Don't Tase me, bro!
If you want idiots to be able to run your servers, you will get what you deserve: idiots running your servers. Why managers think this is a good idea, I'll never understand. I guess the point is that idiots are much cheaper to employ. Unfortunately, they're also more expensive to clean up after when they screw up your servers.
Really, I think that most UNIX administrators know more about what they are doing than NT administrators. I consider that a Good Thing.
Software sucks. Open Source sucks less.
I use what gets the job done. I used to have a Win95 486 machine run as a file/print server and internet router. It worked fine, I could set it up with my eyes closed. But the dang thing locked several times a week, or simply stoped dialling out, or the print queues didn't respond anymore, whatever.
/etc and provide me with valid options for settings I couldn't be bothered to remember. That's what Windows does better in my opinion, and that's why I can find my way around quicker there. But once KDE or Gnome get augmented with a DEEP config tool that goes beyond fluff configuration and can work on anything or most things in Linux, that's when I'll use Linux A LOT MORE. Until then it'll just run my print and file server in the closet. Oh, and when Kylix finally ships, I'll switch desktops too.
So I reformatted and installed RedHat. Took me quite a while to wade through HOW-TOs and MINI-HOW-TOs to figure out the IP masq stuff (there seem to be endless conflicting versions floating around). Print serving and Samba was easier, didn't take too long with the right docs. After I finally got it all going, the thing was quite reliable. It worked weeks and months on end without a hitch, even on this paltry platform (486 16MB RAM 250MB HD). However, every once in a while there were glitches with the print queues and dialling out. A quick reboot often worked, but sometimes I had to screw with the config. Fiddling with printcap files only twice a year or so I simply forget all the options and keywords. Same with the diald.conf file. That's when I longed for a GUI. So I got linuxconf which I have mixed feelings about, but it improved matters somewhat.
Basically I want Linux with KDE and a config system that can actually make sense of all the shit in
Paul
Look, you DOWNLOADED a current copy of kde from their website, but didn't DOWNLOAD the latest qt that it uses, instead you took the out-of-date one from the CD in your hand. Of course that doesn't work. That's the same as, say, in Windows trying to use a game that requires DirectX7, and then installing DirectX5 from your old Win install CD. If you had also downloaded the qt that the KDE site *has a URL pointing you to*, then it wouldn't have been as messed up. I'm not denying that it's a lot of steps, and it's not easy, but your example situation seems contrived to be harder than it needs to be on purpose.
Don't label something "offtopic" unless you know the topic well enough to tell what's on topic.
There are a couple of remaining issues, but yes, we're certainly advancing much faster than the competition. Linux is *definitely* easier to install than Windows already, for example [comparing Win2K and WinME with Red Hat 7 - in the WinME install, USB users don't even get mouse support for the first part of the install].
/etc/fstab to mount a disk [/dev/sda] rather than a partiton [/dev/sda4] and look at the error message that pups up during your next reboot. This isn't Linux specific, but it sure does need improvement.
/top confuse things significantly [what exactly is optional anyway]? Is grep optional? Is KDE? Or is it anything that didn't come with your distro - so Acroread for example, should be installed in different places on different distributions. Clueless folk like Adobe and Citrix put entire self contained apps into /usr/lib
But here's what I see as some remaining issues to do with ease of use.
* Many root-requiring GUI apps display error messages rather than asking for authentication details
* Many error messages are too hard to understand, epecially for console apps. Try modifying
* The filesystem is still all over the place. Directories like
* There's no stadard format for configuration files. An XML DTD would help significantly
* Sys Admins have to deal with multiple permission systems because rwx doesn't offer any fine grained control. So filesystem rwx, Squid and Samba ACLs, and probably more all have to be managed using different system. Linux needs POSIX ACLs badly.
* Lack of anti-aliasing or font smoothing causes accessibility issues for those with vision problems
* lack of comprehensive, distributiuon specific documentation. This is improving over time, but certainly the HOWTOs are often too generalized in nature [eg, they often tend to favour kernel recompiles even if they are not necessary for your distribution]. Those that aren't generalized usually are advertising pieces for a certain distro.
* Unfortunately, RPM lacks apt like qualities, and is far more prevalent. Once this is sorted, Linux apps WILL be easier to install than Windows
* Some minor changes need to be made. What the hell is a GNORPM? Why not just call the app `Installer' and put it on people's desktops?
Well, there's my $0.02
-------
Under Debian Linux it's as easy as:
apt-get install task-kde
It figures out all of the dependencies for you and magically installs them on your computer (downloading them from the Internet as necessary). Better yet, when the the KDE team decides to rev KDE, upgrading is as simple as:
apt-get update ; apt-get dist-upgrade
If, for some reason unknown to me, you decided to install KDE on a system that doesn't do this for you, well then, that's your problem I suppose.
Just because you installed KDE the hard way does not necessarily mean that there isn't an easy way.
And Debian's KDE install is probably the hardest one of the lot. After all, you have to install Debian first, and that can be tricky. There are plenty of distributions that install KDE by default (whether you want it or not).
The technically better product does not always win . How many years has Unix existed for X86. Yet there are more copies of Windows sold.
/. flamage...the change isn't readical)
The GPL is one factor that could hamstring Linux. But only if a court case is brought that gets ugly for commercial interests. Look at the VirginConnect Linux box, and the way it is packaged. YET, no lawsuit on this matter exists. (perhaps it is because the VirginConnect members are LEASING the box from Virgin...)
A radical change in the GPL would also cause a problem with Linux. (if the GPL changes, and no effect is noted beyond some
Another thing which can damage Linux is the 180+ versions of Linux all clammoring for marketshare. Smacks of the UNIX versions of the 1980's.
(ok, who has a link to the "linux os versions counter page" I have heard reference to..this way we can all see the 180+ versions)
FreeBSD total users is at about 20% of the TOTAL of Linux. Think of any Linux distro that has 20+%? And BSDi is getting moeny from ISP's to work on and improve FreeBSD. ($5 mil from Yahoo and $5 mil from Livin' on the edge)
Given the final outcome of a Linux ELF binary as the X86 Unix standard Binary, the lasting effect of Linux might just be a standard binary.
If it was said on slashdot, it MUST be true!
There's a lot of good sense in Gartner's projections, which, being their projections, probably err a little on the conservative and hedging side of things.
I think the growing importance of the server appliance market is right on the mark.
Since hardware is getting so cheap, what you'll find is increasing appeal for the low software cost barriers of Linux based solutions. And, since Linux is advancing on the multiple fronts of
- embedded
- real time
- enterprise(big SMP)
these solutions will become more ubiquitous on a faster time scale, IMHO.Indeed, Linux will tunnel thru the mid-range server market to the high end (witness the economical clustering technologies that compensate for Linux current lack of extreme SMP scalability) to eat the lucrative high end business before the conservatives in the mid-range start waking up to the advantages Linux solutions hold for them, too (especially if MS starts charging subscription rates to bring in revenue). The current market divisions of hi/med/lo for Unix/W2K/Linux will cause both the first and the second to disappear, in that order.
"Provided by the management for your protection."
OK, as a new Linux user I have to say something. Linux is not any harder to use for the end user. But, (BUT!) trying to do any sort of changes to the system (something a homeuser would want to do) IS much harder. Even something as simple as installing softare, or even changing a monitor.
I've just recently installed Linux on one of my old systems to run a counter-strike server. I have some limited experience with Solaris from school, so I was not coming into totaly blind.
The initial install was pretty simple. Not something your average user would want to handle (I wouldn't expect my parents to be able to handle it, or even any of my brothers for that matter). The redhat installer detected everything but my monitor just fine. Setting up partitions should be automated a little better. But I've got enough computer experience that I worked through it.
After the install, I added in a network card. Piece of cake. Pluggedd it in, turned it on, and it was detected.
From there things went bad. The redhat recommended partition sizes would not hold all the files I needed to add for the counter-strike server. So, I do a little research on how to add another partion. I play around for a while, figure out how to create a new partition, add a filesystem, and mount it. But then I royally screw up trying to copy the existing usr directory into it.
Time for a reinstall.
So I do a reinstall, setting up the partions better this time. I get the counter-strike server running pretty easily by following a tutorial. great everything is running fine.
A few weeks later my monitor goes out (not linux fault, it was running on my windows box). So I borrow a new from work. Plug it in, turn on my linux machine. No good. As soon as Xwindows starts up the signal gets scrambled (refresh rate too high). But how do I just boot into text mode? (I had set graphical login in the install) No idea.
So I spend all night trying figure out how to NOT boot into the graphical login. By piecing together some info I can find about 'LILO', and xwindows, and 10 other 'tutorials', eventually I figure that I need to:
- hit control-X on the OS loader.
- enter 'linux 3'
- use Xconfigurator (capital X, the rest lower case of course, even though most references to it call it XConfigurator or xconfigurator) to change my monitor settings.
Nice, easy and intuitive eh?
"Freedom in cyberspace'd be fine and dandy if we happened to live there."
OK, I've finally had it with everybody trumpeting the TCO equation. We have a lot of qualified technical people here, and Gartner has a bunch of reasonably savvy business analysts, but does anyone bother to ask the people who REALY KNOW about TCO? Oh, no, we just spout our local dogma as the truth.
Hard facts: the people who know TCO are the System (NOT SOFTWARE) Architects and the MIS/Operations Managers. I've very little faith in Director-level (or above) management knowing about real costs, and Programmers (no matter how smart) don't have a fucking clue about what TCO means.
TCO has four major considerations, which vary by the company involved. There is no such thing as a "global" TCO. There's the TCO for your company, and there's the one for Yahoo, IBM, etc. all of which are different. And each consideration is of a different weight (importance) depending on your situation.
One things I will point out that seems go against a mantra here on /. : Open Source OSes have no real maintenance advantage over closed-UNIXes. Sorry, but that's the truth for 99% of the businesses out there. Nowdays, installed commercial UNIXes use virtually the same userland toolsets as the FreeUNIXes (who cares that SUN doesn't fix bugs in their version of sendmail? Anybody actually run Sun Sendmail? I thought not.) The only real differences are in major applications (which are almost always closed-source, even on FreeUNIXes), and the base kernel itself. Only the largest companies (or those whose business depends on doing kernel-level development e.g. Cobalt) will care. Fixes for important problems in this sort of stuff come out of the commercial vendors as fast as they do from the FreeUNIXes, and honestly, virtually no company has the resources (or even the expertise) in the programming staff to fix subtle bugs in stuff like the Linux kernel (or glibc, or whatever). And if you think the admin staff is going to do this, hey, well, I've got a bridge in Arizona I wanna show you...
-Erik
(yes, I'm a System Architect, and yes, I'm pissed off...)
There are always four sides to every story: your side, their side, the truth, and what really happened.
Definitely. My experience with the analysts, Gartner, Forrester, and the like, is that they take generally available information, make it look pretty and "professional," back it up with some survey data (sometimes), and offer it for very high prices to corporate buyers via a direct sales force that won't stop calling you once they get your name. Fortunately, the news sites are more useful every day, so you really don't need to subscribe, unless you really don't get technology and would rather pay someone else to get it for you.
Remember, the analysts all said "B2B" would be the next big thing...
sulli
RTFJ.
To reach today's 20,000,000 users Linux had to grow at an anual rate of 537%. (5.37^10 ~ 20M) Gartner is forcasting that it will slow down by 511%. I wonder how much they were paid to think that?
--
Life's a bitch but somebody's gotta do it.
XFCE is ok
that's my _opinion_