Follow Slashdot blog updates by subscribing to our blog RSS feed

 


Forgot your password?
Close
typodupeerror
×
Linux Business

Linux on the Desktop 495

Posted by michael from the penguin-in-every-pot dept.
Dhar writes: "Rob Valliere has posted a Windows vs. Linux review: "This review focused on Linux Red Hat 7.1 from a business user's view and attempted to answer my client's question "Can Linux be used as a replacement for Windows 2000". After an intensive hands-on Linux project lasting several months, I was able to provide my client with a pertinent answer to this question." I like the answer. ;-)" It's good that he covers the pitfalls he encountered; opportunities for improvement.
This discussion has been archived. No new comments can be posted.

Linux on the Desktop More

Linux on the Desktop

Comments Filter:
  • You will be stunned by the bang for the buck that Linux bundled free "open source" software offers.

    Since Linux is free, wouldn't any amount of "bang" mean that the ratio of "bang" to "buck" is infinite? Q.E.D. Better than Windows!!! (not trolling.... well.... maybe a little bit... )
    :)
    • It costs to deploy Linux. Deployment costs include licensing (the only free thing in Linux), hardware, and support (anybody else want to add to this list?).


      Linux is free. Filling an office iwth Linux boxen is not.


      • It costs to deploy Linux. Deployment costs include licensing (the only free thing in Linux), hardware, and support (anybody else want to add to this list?).

        Yep. User training on new tools is definitely another cost and not a trivial one. Practically, I think Linux would be a no-brainer choice except for the investment of user training in existing MS products.

        Some of these cost comparisons are interesting not only from the standpoint of the specific environments Win2K vs RH 6.2, but also from the general perspective.

        That is, few people know what it really costs to change OS environment because of all the far-reaching implications in terms of training and support. And, because the OS changes most environments have experience with are less radical: NT4 to 2K, for example.

        Along the same lines, very few IT organizations really understand all of their current costs. The costs of keeping with the existing environment.

        These costs includes the usual well-known costs of licensing, hardware, IT support staff time, costs of training classes for users, but also includes less tangible items like how much time do the secretaries spend changing fonts in Outlook or fidgeting with Word documents to get them to look right. They get benefits from these applications, too, that should be considered in the overall cost assessment, but those are usually pretty well explained by marketing brochures and advertisements.

        Fortunately for the sales of MS products, the training costs are frequently absorbed by someone besides the corporation buying the licenses. Many secretaries take classes in Word, Excel, etc at their own expense to make themselves more marketable to employers. If all of these costs were made visible, then I think a better basis for comparison could be made.

        • less tangible items like how much time do the secretaries spend changing fonts in Outlook or fidgeting with Word documents to get them to look right.

          Exactly. The design of software -- its "user-friendliness" to employ an overused term -- is a major factor in total cost of ownership. That can't be answered by a once-over-lightly review by a system administrator mostly concerned with ease of installation, which is what the article amounts to. Evaluation of costs created by usability factors requires extensive user testing, both in the field through ethnographic methods and in controlled situations in usability labs.

          Instead, what we got in the review was a lot of unsupported superlatives like "excellent," "great," etc. The personal opinion of a system administrator does not have a great deal of relevance to the costs that the software will create in an actual user environment.

          The recently discussed GNOME Usability Study [slashdot.org] showed that even at the current fairly advanced level of development, GNOME is significantly harder to use for most people than Windows or the Mac. That's a cost that impacts overall productivity of the users.

          And GNOME is considerably more usable than most of the applications available for Linux. Granted this is not an opinion I am able to support empirically, since the studies have not yet been done, but it seems obvious that, for instance, Microsoft Office is more usable than StarOffice at their current stages of development, and that the continued reliance on the command line for routine tasks presents a very steep obstacle for the average user.

          So it seems to me quite inaccurate to describe this software as free. Low up-front costs are being traded for persistent usability costs. This may be an equation that makes sense in some offices but as a general rule, it would be a money-losing proposition.

          Tim
        • Continuing with these thoughts...

          The IT guy was set to save $10,000 on 25 machines by moving to Linux. But there was no obvious consideration of the labor cost in that move.
          - A new DB has to be created to replace the current MS one.
          - 25 machines must be completely reconfigured.
          - 25 users must be re-trained for *all* their software usage.
          - Some tasks will be slower, always, even after re-training (e.g. making PDFs)

          What's the cost of that? Just looking at 25 users x $10/hr x 10 hours effective training = $2500. A quarter of the savings is eaten just by basic staff transition costs.

          Without further information about the real costs involved, this move may be penny-wise, pound-foolish.

      • Yes, but as the folks in Key Largo, Florida have shown us it is not only possible to put Linux on the desktops. It is possible to put inexpensive thin clients on the desktop connecting to a Linux server via X. That means that not only is the licensing free, but for $200 bucks a desk (sans monitor, but you can reuse those) and the price of a server you can get rid of your desktop troubles all together. So not only would your Linux deployment cost you less than the next upgrade to MS Office, but your long term support costs would almost certainly decrease.

        In Key Largo they currently have 300 users hung off one commodity Intel server, but even at half that many users imagine the savings in support and maintenance. There would be precisely one box to administer per hundreds of users. Even small offices can benefit from this sort of arrangement.

        The fact of the matter is that if you were really cheap it is possible to turn your existing Windows based computers into Linux based "X terminals." If you already have standardized hardware this sort of thing wouldn't even be difficult to do. When a machine fails replace it with a ThinkNic or an X terminal and throw the old clunker away.

        Thin clients have never taken off because there was never really any useable software to run on them. The combination of Linux, StarOffice, and Mozilla, while it certainly has some problems, is far and away the most compelling set of thin client applications that have ever been available. You can bet that people are going to start actually implementing this stuff.

        • The Key Largo installation is using commercial software on their servers, including WordPerfect, MS Office and Citrix. That's not factored into your cost estimates.

          Tim

          • Some commercial software is going to be necessary in nearly any large installation. Removing Windows, and the Windows desktops that are required, is a huge benefit cost wise whether Linux or Free software is involved at all.

            For example, in this particular case, the admins at Key Largo have two boxes that they need to administer. A Windows Citrix server (for legacy applications), and a Linux server for the standard desktop. That is still a huge win for thin clients, even if all of the software is proprietary.

            Companies are used to paying for software, and I have no problem with that. I was merely pointing out the costs savings that are derived from removing Windows from the desktops and replacing them with thin clients.

            Windows could do the same thing with Citrix, or with Terminal Server. Unfortunately instead of being cheaper to deploy this way, it is more expensive. Not to mention the fact that Windows terminal server will only support a fraction of the users that Linux + X will on the same hardware, and Windows applications generally don't expect to be run from a terminal server.

            In other words, my point had more to do with the beauty of X and thin clients than with Linux. Linux just happens to be the first OS suitable for thin client deployment that actually had some competitive applications. The fact that it has several competitive office suites (all much less expensive than MS Office), and the fact that Citrix would allow you to mix in some Windows legacy applications only sweetens the deal.

        • The idea of windows based X Clients is a great idea. It lets you slowly migrate applications to the Linux server, while still keeping windows to run other software. Once the migration is complete, you can start scrapping your windows boxes and replace them with X terminals.
    • Since Linux is free, wouldn't any amount of "bang" mean that the ratio of "bang" to "buck" is infinite?

      I doubt he worked for free, so his salary was part of the cost for the Linux solution. As are a handful of other costs. Standard accounting practices in the professional world.

      --
      Evan

    • ...there are always costs. Licensing fees are only one cost of using a particular piece of software, and usually not the biggest. But you knew that.

      I've grown fond of Linux over the years, but for a long time I had a hard time taking it seriously. The was mainly due to my first experience installing it, many years ago. The basic OS installation wasn't too hard (though not for the computer newbie), but I couldn't believe how much time I spent on silly little configuration issues. Bad design, inconsistent design, undocumented design. It was actually worse than Windows 3.1! It took me an entire day to figure out why Netscape always mapped Backspace/Delete backwards from other applications. That's the sort of thing that drives up cost of ownership. My first response to Linux was, "OK, it's free, but can anybody afford it?"

      Well, Linux has gotten easier to administer (thought it's still too complicated, and there's too much undocumented stuff) and Windows has gotten much much harder. But the total cost of ownership issue is still hard to answer. One big item is retraining everybody to use the Linux equivalents of MS Office. That's assuming you can persuade people to make such a basic change!

      • The basic OS installation wasn't too hard (though not for the computer newbie), but I couldn't believe how much time I spent on silly little configuration issues.

        Yup. My biggest problem with Linux is this sort of thing. The HOWTOs are very well done but they expect that everything goes smoothly. Most of the time, it doesn't. I'll get to step 4, for instance, in a specific procedure and it will say "type this, and this will happen." I'll do it, but get an error. Then what? The HOWTOs usually do not take this into consideration. It simply says "x will happen", but there needs to be some extra help there: "If you get this error, you need to do this."

        So I end up on the web, searching for that particular error. Usually I have good luck finding it and things have always worked out in the end, but not without spending a considerable amount of time. It took me most of a day to upgrade sendmail over one stupid little issue; I had to download and install a couple of other things before the newer version of sendmail would work correctly. Not a problem, but the howto didn't tell me what to do in case of error x and I had to figure it out on my own. 99% of computer users can't even follow simple directions. How are they ever going to run Linux, where you actually have to use some common sense?

        Most reasonably intelligent people would have no problem deploying a Windows 2000 domain with a few dozen clients on their own. It's all very basic, and everything pretty much works as expected. Linux, on the other hand... I couldn't imagine most people trying to setup even something as basic as a Samba share. Sad, but true.

        The fact is that most users can barely handle Windows. We set my mother in law up with Internet access last week. She walked away and left the thing connected, so her ISP dropped her after some amount of time. She called me up and asked why she couldn't get to any web pages, and wanted to know why there was a button on her screen asking her to reconnect. She was too afraid to just hit the damn button to see what would happen. This is the kind of bullshit that most people pull on a daily basis. If they can't even use Windows, they'll never figure out Linux until it's dumbed way down.
  • The fact is that the guy was very in-depth about it. I'd have liked to see his MSAccess replacement linked to, but oh well.

    The problem with desktop applications these days is that they're gigantic for no obvious reason. I still want to see someone write a full-featured office suite that takes up a grand total of no more than a megabyte for the source tarball -- I would be inclined to think that a fairly nice word processor could be put together with nothing more than Perl and Tk, using standard command line tools like ispell for the more specialized services and (I've heard suggested) HTML4+CSS as the file format.

    Nobody seems to have tried, though...

    /Brian

  • *Cough* This is a great example of someone writing an article to match what they "already know".

    Did he actually do an analysis of what his office needs for word processing? I don't see a list of required features. The alternatives are feature-poor, so we simply don't know.

    Although he says that StarOffice is "fully compatible with Word/2000", other experience has shown otherwise. Did he test with some complex documents?

    What about printing? Did he test with all the printer types in his office? If he is 100% Postscript that he has some chance, but if there are any low-end Epson color printers, his users could be in for a big surprise.

    And on and on...

    It's pretty obvious this guy has never done a feasibilty study in his life. I give it a D-.

    • What about printing? Did he test with all the printer types in his office? If he is 100% Postscript that he has some chance, but if there are any low-end Epson color printers, his users could be in for a big surprise.

      You chose utterly the wrong argument here. I have an Epson Stylus Photo 1290 and previously had an Epson Stylus Color 850, and the GIMP-Print drivers for these have totally blown me away - the output I get from them is simply stunning, and considerably better than the official Windows drivers. They also support every feature and resolution of my Stylus Photo, even doing colour matching using Postscript.

      Also, they don't crash, unlike the Win2k drivers...

      Better yet, I'm using these drivers with CUPS as the print spooler and the KDE2.2 print framework. Using this combination, it is just as easy to add, manage and remove printers as it is under Windows. In fact, for networked printers, it is even easier, as I can also configure CUPS through a web browser from anywhere. The print dialog in KDE apps is fully comprehensive, easily customizable by each app and supports things that Windows doesn't - for instance, post-processing of print data through arbitrary commands, which means every print driver has the capability to print multiple pages per sheet, and every app can print straight to a PDF file. Truly, it is a joy to use. I haven't seen a comparable print framework anywhere else.

      For more information, check out the GIMP-Print [sourceforge.net], CUPS [cups.org] and KDE Print framework [swing.be] websites.

      Printing under Linux has finally come of age - and it is better than Windows!

      • In fact, broken printing is what really ended my reliance on MS. Win98 let me down several times in school and Network printing here at work is just a freaking mess. I quit using MS alltogether at home about a year ago when printing became unreliable. Machines that die stay that way. Those that refuse to boot, get Debinated. Things have been much easier that way.

        Red Hat 7.1's good printing was a pleasant supprise. I gave it a try after the MS printing died on my wife's K6-2. Configuration was easy, and the output was just as good as MS ever was. Images from GIMP, documents from KDE are outstanding. Considering that the legacy alternative does not work at all, the ouput is infinitely better.

    • If there are any low-end Epson color printers, his users could be in for a big surprise.


      Mine works fine.


      Although he says that StarOffice is "fully compatible with Word/2000", other experience has shown otherwise.


      Can you provide examples? I admit, I haven't been sent very complex documents, but it has handled anything anyone ever sent me.


      It's pretty obvious this guy has never done a feasibilty study in his life. I give it a D-.


      I agree, but I think he was mostly correct anyway. Not knowing the exact needs is a problem, but the given solution seems to provide for most basic needs.

  • Nice cost savings, but it only works when you don't factor in lost productivity for retraining people to use the software.

    The main problem is a lot of implementations try to look like windows, and that really only aggravates the issue.

    So, if they saved 10k on licensing and hardware, how much did they lose in time devoted to people learning the new system? What costs are involved when hiring people? Using temp agencies?

    The reason Linux is having a hard time moving into the mainstream office is not because of price, but simply the fact people would have to retrained to use it. Sometimes time is more than money
    • Recently my girlfriend moved in with me. When she needed to use a computer, I set up an account on my Linux box, showed her how to log in, showed her where Abiword where in the menus, and didn't worry any more.

      A couple of days later she was about as familiar with Linux as she is with Windows, had found all my games, had found gphoto and downloaded some pictures from my digital camera and set one of them as her new background.

      So far I've hardly done anything to help her, except giving her names of some of the apps.

      And she don't even like computers...

      Yes, there are things on Linux that have a steep learning curve. But once a reasonably recent Linux distribution is set up on your machine, most beginners don't run into many more difficulties than they do under Windows - in fact I'd dare say that many beginners won't really realize that there are things that are more difficult under Linux, since most such things are sufficiently difficult under Windows too that they'd ask someone for help in either case.

      For intermediate users, though, the switch might be a little more painful.

    • You know, the author did consider that. win3.1 is not win3.11 is not win 95 is not win 98 is not win ME is not win 2k is not win XP. The interfaces to the programs that run within these vastly different OSs are just as poorly documented as the OS, but even less consistent. Productivity is lost with every "upgrade". Even more is lost to broken Word macros and mindless formating issues where I work. The author implicityly considers Linux as an escape from all that.

      He's right. That GNOME or KDE are any more difficult for the average user to work has got to be the #1 troll of the year. If anything, the interfaces to GNOME and KDE or any other window manager are easier because they do not suffer needless market droid type changes. Also, user customizations are much easier to save out and move from machine to machine. MS will never catch up.

  • Hmmm... not sure how to take the article (Score:3, Insightful)

    by weslocke ( 240386 ) on Thursday September 20, 2001 @11:14AM (#2325162)
    Afterall, the entire article shows where and how Linux and Linux alternatives can be used to replace Windows 2k... as well as it should since that's the primary thrust of the article.

    I would think that everyone here would agree that Linux can more than admirably replace a Win2k desktop.

    But my question would be, should it?

    The wrtier uses Star Office for his example. (While there are better out there, I'll use it since he did) He even points out that the suite is missing quite a few of the Office features, almost all of which you and I will probably never use. But what about the secretary for your group? Or that person that uses Word to create forms? You might be able to do a lot of the things that these people rely on in Star, but what sort of hoops to do it? And could a computer 'illiterate' really catch on to those hoops?

    That brings me to his PDF creation solution. Print to PostScript, then use a 3rd party utility to convert to PDF. It's not very user friendly, "but it works." If you look through the article, you'll find that phrase quite a bit.

    My whole point is that Linux can replace/supplant a desktop for those of us on Slashdot, and typically do a far better job than the desktop it replaced. However 99% of the office workers out there are not the typical bored geek that hangs out on here.

    So yeah you can replace Win2k with Linux for a business user workstation, but I personally don't think that you should. At least not yet.
    • That brings me to his PDF creation solution. Print to PostScript, then use a 3rd party utility to convert to PDF. It's not very user friendly, "but it works." If you look through the article, you'll find that phrase quite a bit.

      I don't know what desktop this guy was using, but KDE 2.2 has a system wide "Print to PDF" option integrated into all KDE apps, including KOffice. That seems pretty user friendly to me.


    • "but it works." If you look through the article, you'll find that phrase quite a bit.

      If only NT would work at all! "But it works" is the sloppy catch phrase that MS folks used to throw out as they smashed down their crappy software on cluefull users. I work for a large company, with many databases THAT WILL NEVER TALK TO EACH OTHER, mail that gets stored in a propriatory format, and documents that never print the same twice. All of this is because of propriatory "standards" that never stand still. I'd love something that acutally worked around here freaking ever.

      But my question would be, should it?

      Of course companies should put Linux on the desktop and soon. Just reverse the question and see how obvious it is. Imagine your company was using FREE software and data formats. Try asking your boss, "Should we move to propriatory software that we have less control of, costs us more, is less secure, and does not work as well?"


    • So yeah you can replace Win2k with Linux for a business user workstation, but I personally don't think that you should. At least not yet.

      I agree completely.

      I think KDE and Gnome are excellent desktops for even the typical business users, so that's one box checked. But there are a couple more critical items IMHO.

      I'm not ready to stake my reputation on a transition to Linux until several more key items are a little further along than now:

      1. Evolution (or equivalent email/PIM organizer)
      2. Mozilla 0.9.9 (something fast, standard, that doesn't crash much (or Galeon, Konqueror, Kameleon)
      3. StarOffice 6.0 (something with less integration that 5.2, with standard XML file formats and better compatibility with MS Office formats; maybe KOffice) Something with better Excel compatibility.
      For now, a Linux desktop is a great way for a SOHO to save some bucks if they have someone with some technical expertise and some time and not a lot of cash. Maybe that's why I got the impression this review was for a client in Thailand where you can get more people per dollar than in the industrialized world.

      For the general corporate desktop in the U.S., I think about 6-8 months from now should provide the point at which Linux is really a great option for the majority that don't want to be too close to the bleeding edge. Until then, for many users the transition from Windows is more uncomfortable than the dollar savings alone can justify.

  • The author is right to pick on XConfig - X setup is still abyssmal, and simply needs to be replaced.

    I also agree with upgrading sets of associated packages - like KDE - this is still not anywhere near the level of ease of use that could be automated fairly easily.

    I guess the real issue is - who is going to tackle this work? These are big problems that require the attention of a group with enough clout and authority to push their solution into the linux mainstream. Red Hat, IBM, Ximian and the FSF are the only groups I can think of.

    • If you look at Madrake 8.1 (I have RC1 running on my test box right now from the default install), the setup/install is much improved - no intervention required for X setup... all it asks you is the res and color depth you want to run at (they provide decent defaults). Once running, the controls provided allow res/color switches as easily as windows (though there is a disconcerting pause while switching)... X could use a lot of things (from setup to complete re-architecting), but if the distro installers get the installs done properly and provide good config utils, it really helps things along nicely).

      I'll agree that upgrading sets of packages isn't always as easy as we'd like... the Mandrake Update tools are nice (I'm sure Ximian has something, but I haven't looked...), and are quite easy to use... auto select update server (granted not always the freshest mirror), lists the packages - user selects, dependencies are (usually) caught, and a while later things are humming. Heck, I updated the kernel that way (I didn't have a lot of faith that would work) and it actually came back up fine. There are distros that focus towards the experienced, power users, and there are those that really do make things amazingly easy for newbies (without sacrificing too much).

  • Try be inovative instead of just replicate ? (Score:5, Insightful)

    by boaworm ( 180781 ) <boaworm@gmail.com> on Thursday September 20, 2001 @11:16AM (#2325176) Homepage Journal
    This topic seems to come back over and over on slashdot. Continously, people try to say that "Linux is desktop ready". (Or in some cases they say its not ready, and thus treated like a flaimbait :-)


    I dont really think this is the interresting issue to discuss. Whether Linux will be a good desktop environment/replacement for Windows2000 wont really rely on "Linux", but on all the apps that will run on linux.
    What i'm saying is simply that the Linux kernel has been "desktop ready" for a long time, its just the easy task of networking, supporting some common hardware, and not crash.
    What IS the issue is what programs are available. Why do people choose to use Windows 2000 on the desktop ? Because they want to run the Windows OS ?.. Dont think so. People are looking for the programs that runs on Windows, like MS Office, Internet Explorer, Visual Studio, Borland tools [add a long list here]. They simply look for a good platform to run their favorite software on.

    The Open Source community has a little dilemma here, first they try to say "We can replace windows 2000" by providing products that can communicate with [major software company] products. Then they try to offer software that works just like those products they are trying to replace.. Why should someone want to use the copy when there's the original ? Of course you cannot provide an Office Suit for Linux that will always be 100 % compatible with MS Office, simply because even Microsoft cant :-).


    The core issue is, dont try to walk your way to the desktop by making [almost as good] replicas of existing desktop software. Instead, offer something better! Something like a uniform word processor that uses the XML standard. Maybe use the same XML for spreadsheats, email programs, etc etc etc.


    As long as we try to copy/replicate leading software we will always be one step behind. Why dont we take the lead and provide new things instead ? :-)

    • printing on a non-postscript printer still sucks in linux. It's more trouble than needed, and the results arent as nice as with NT.

      I know this is a problem with drivers and applications and not the kernel, and I know it seems like a minor issue (not!), but besides games (unsolvable) printing is still an issue that makes Linux less fit for the desktop.
      • printing on a non-postscript printer still sucks in linux. It's more trouble than needed, and the results arent as nice as with NT.

        Try using cups [cups.org]. It does a *very* good job of bringing printing in *nix to the same level of nice features that you see in winders and macos. I don't think it's all the way there, yet, but it's a dramatic improvement. Couple it with a large array of drivers from gimp-print [sourceforge.net] and you've got a much more sane printing environment under *nix.

    • The core issue is, dont try to walk your way to the desktop by making [almost as good] replicas of existing desktop software. Instead, offer something better! Something like a uniform word processor that uses the XML standard. Maybe use the same XML for spreadsheats, email programs, etc etc etc.
      Been done.

      escaflowne:/home/taliesin(0)> file judmon.abw
      judmon.abw: XML document text

      AbiWord already saves in XML format. It's been around for a while. Kword saves in compressed XML format. I believe some of the spreadsheets do likewise.

      As far as going beyond the functionality of You Know Who... that remains to be seen.

      • OK, XML is arguably better than other formats because it's somewhat self-documenting and there's commonly available tools.

        But, if these formats aren't documented, and if OSS App #1 doesn't have the code to open OSS App #2's formats (and visa-versa), the XML-ness of the format means nothing in the real world.

        I had to say it, but businesses care more about defacto standards than open specs. They want to hear that "If you switch to a Linux desktop, you can use Sun StarOffice (or whatever) as MS Office replacement and send documents to any other Linux user." They don't want to hear "We've got 10 incompatible and half-finished office suites! Pick any one you like and start hacking the code. But emacs and TeX rulez!".

        Problem is tho in a 'free software' environment, nobody's going to pick a winner and centralize development efforts on that product. This point is where the culture of Linux runs counter to the marketing efforts of it's advocates that want to see it as a Microsoft replacement (when it's not intended to be such).
        • Problem is tho in a 'free software' environment, nobody's going to pick a winner and centralize development efforts on that product. This point is where the culture of Linux runs counter to the marketing efforts of it's advocates that want to see it as a Microsoft replacement (when it's not intended to be such).
          No, that is the point. There are multiple ways to do it. We don't like it when Microsoft tries to be The Only Way, why should we do as they do? Competition is good.

          It really cheeses me off when Stallman et al try to push people into working on The One True Gnu instead of continuing work on things they're already familiar with.... Listar [listar.org], for example... while the web interface is actually a little clunky, the underpinnings are actually better, but RMS wanted the author to drop it and work on MailMan, nevermind whose code was better.

          The fact that we have both Gnome and KDE, Mozilla and Galeon, Perl and Python, several viable Java Virtual Machines, and gods knows how many Linuxes and BSDs to run them on, is one of the Best Things about Open Source.

          Interestingly enough, the only "standards" I can think of that Linux apps conform to are those foisted on us by software monopolies.... Microsoft and Adobe. Perhaps we do need to come up with a more robust XML document standard... but I don't think any one outfit should do it. That's asking for trouble.

    • The core issue is, dont try to walk your way to the desktop by making [almost as good] replicas of existing desktop software. Instead, offer something better!

      Unfortunately, the free software community at large is just too damn helpful for this to work. Every time a better product is produced for *nix systems, some jerk invariably ports it to Win32 ;-). And then, there goes another reason to switch to Linux. The examples are countless: Apache, Perl, GIMP, etc.

      I don't really know if this impacts anyone's decision to switch to *nix, but it is true.

    • Why dont we take the lead and provide new things instead ? :-)

      Like the standard *nix tools?

      IMHO it's not a question of desperately trying to invent something new. It's more about forgetting Windows as the desktop paradigm. In fact many of us know that an efficient workstation environment is not necessarily called 'desktop'.

      \begin{rant}
      [thinking of the new planned version of Enlightenment] I don't want a desktop environment, I don't want a fancy panel, I just want a window manager, dammit!
      \end{rant}

    • 3 year-old Gateway Solo 9100 notebook (Pentium II 300MHz with 96 MB). The RAM on this notebook was insufficient to run Windows 2000

    Hmm... I had Win2K running just fine on a 96Mb Acer Travelmate. Sure, it paged like a bastard on startup, but soon settled down (and I did go back to SuSE/KDE). Minor niggle though, otherwise a great article, well presented, and commendably objective. Cheers!

    • Heh.

      My old room mate ran win2k on a P200 with 32 megs. It took forever to start and paged constantly when loading an application, but it ran ok once a program was started, so long as you didn't try to switch tasks.

      We did try linux on the machine (fvwm + mozilla), but it was unusably slow with the ten thousand ton monster running. xterms were fine, but really, a terminal program on windows provided all the needed functionality (as in it was faster to run programs remotely over a cable modem).

      As of three months ago, the unpageable kernel of windows 2000 was smaller than the unpageable kernel of linux, and Internet explorer had a smaller memory footprint than mozilla. Using an embedded configuration may change the first, and the mozilla team is constantly working on the second, but my guess is that both of those will continue to hold in the forseeable future for a desktop configuration.

      I'd really have to recommend windows 2000 as the environment for memory constrained computing if a graphical environment is necessary. Linux is really an option in this situation only if you can ditch the web browser and office programs.
      • > the unpageable kernel of windows 2000 was smaller than the unpageable kernel of linux,

        Using what configuration options of Linux? How big are we talking about here?
        • Linux's memory footprint is a lot larger because of several things. First, Linux can't page out kernel memory. On my Win2K machine, of the 36MB of kernel memory, only 8MB is unpagable. Since most of the OS's features aren't being used all the time, this can lower the memory footprint of a running machine. Also, Linux can't page out the page tables, while Windows 2000 can. This becomes a larger problem in something like XCOM (used in Mozilla) where lots of sizable files are mapped in at the same time, but not necessarily used all the time.

  • He used StarOffice 5.2. The old StarOffice was incredibly annoying. Recent OpenOffice builds are much better. I know they're not officially released yet, but I think the development builds would have been a better choice.

    Also, Evolution, Balsa, or Mozilla would have been better for e-mail. KDE probably has a native app that's better as well (KMail?).

    He said he couldn't get the drives to automount on RedHat. Odd, that usually drives me mad until I remember to turn it off. Not sure what his difficulty was.

    Trouble with dependancies for RPMs: use Debian or Ximian's Red Carpet if that matters. I think somebody told me Mandrake had apt-get for RPMs - that sounds interesting too. Ditto for the upgrade problem. RedCarpet and apt-get are miles ahead of Windows in this respect. On the plus side, once you get everything the way you like it it's really easy to set up painless network installs using RedHat.

    On the whole a pretty positive piece, even if I don't agree with everything he did. You can probably chalk this up to his being a Windows administrator and not really knowing his way around the Linux world (I wouldn't do much better trying to fine-tune a Windows installation). It would be interesting to have one of these comparisons every six months or so just to see the progress. I bet it would be impressive. As far as I know most of the annoyances he mentioned in his article have already been addressed.

    Funniest Quote: My customized KDE desktop is better than Windows 2000! He seems so surprized :-).

  • It's not so much how difficult that Linux is to install that makes it, to me, difficult to use. It's the community that feeds the perception that Linux is a geek's OS.

    As long as the "more able", on-the-fence Windows users continue to get responses like, "Read the fookin esotericHOWTO, crypticHOWTO, and horriblyexcitingtoreadHOWTO, you idiot!" a la Jimmy Fallon on SNL as they take their first look at Linux and post newbie questions on usenet, the wall will never crack. Heck, when I first installed Linux I didn't even know where the HOWTOs were!

    If I'm Joe Computer User and my "expert friend" says Linux is too complicated in no small part due to the "newbie flame" s/he got when trying to become part of the community, I'm not about to try out this new OS.

    Ruffin Bailey

  • Anti-Virus Norton Anti-virus Not required with Linux, according to the experts.

    Enuff said.
    • Why is that 'enough said'? Can you name a *single* Linux-based virus?

      It's been mentioned above, but I'll say it again - root exploits and worms are NOT viruses, and anti-virus software won't protect against them on any platform.
  • As much as I'd love to ditch MS, the truth is that for the last X number of years, most all 3rd party apps.. specifically specialty apps writtten for specific industries, not just "Word", have been written for Windows.
    I would love to get rid of MS in my work environment, but the apps I need to use simply don't exist under any operating system other than Windows, and it's extremely unlikely that the 3rd party vendors will blow millions of dollars to make a Linux version to please the small (albeit growing) percentage of computer users that despise MS.
    You could make a wonderful office suite, a billion times better than MS Office, but it doesn't matter, because the office suite software is a minor set of utilities in many, many workplaces.
  • Instead of KDE, a less complicated and easier to use and understand environment is Windowmaker + ROX-Filer.

    Instead of the apps he chose, here are my personal favorites:

    Database: MySQL and Ksql for end-users, embedded perl for web-based company-wide databases.

    PIM with Email: a web-based app tailored to the company's business processes for scheduling, and Pronto for mail.

    Browser: Mozilla

    Image Viewer: GQView

    PDF Viewer: ghostview (gv) & Acrobat for stuff gv can't handle

    Umm...what does a mp3 player have to do with business? :)

  • Redhat/Linux confusion again (Score:3, Insightful)

    by RPoet ( 20693 ) on Thursday September 20, 2001 @11:29AM (#2325259) Journal
    From the article:

    And here lies one of the biggest challenges in the Linux world. The Red Hat installation is outstanding - you basically push a button and 500 or so rpms are installed and configured correctly. But upgrading individual applications, especially for a large package like KDE, is far from pushing a button. Upgrades in Linux have a long way before they will be as easy as upgrading Windows applications.

    This guy re-installed Redhat because he couldn't manage to install new KDE 2.2 packages. And he takes it out on "Linux," when he should of course take it out on Redhat. I know I updated KDE with a few commands involving urpmi on my Mandrake install, and it should be even simpler to do on Debian -- certainly much simpler than upgrading something equivalent on MS Windows.
  • So it can be done but why should I do it? I'm not for Win specifically, but I see no reason to change. What are the advantages to changing to Linux?

  • Windows is getting a lot of user level security tools now--easily configurable via a gui, not hard to use.

    Linux Security is powerful, but is a morass of application and configuration issues.

    Someone oughta make a GUI like gtp with tabs to logically split this stuff out and configure it, maybe even a tab for doing saint probes, nmaps, etc.

  • CVS ? (Score:3, Insightful)

    by Etyenne ( 4915 ) on Thursday September 20, 2001 @11:45AM (#2325347)

    When copying files under Linux, original timestamps are replaced with the current date. So the "date last modified" file attribute becomes "date last copied". This becomes a nightmare for anyone dealing with many files - how can you keep track of when a file was last modified. You can force the original timestamps using cp -p., but this means not using the GUI file manager. Very poor Linux design feature!

    Please somebody buy this guy "CVS Pocket Reference" !!! :)

  • Good ideas. (Score:3, Insightful)

    by supabeast! ( 84658 ) on Thursday September 20, 2001 @11:55AM (#2325397)
    This article does a good job of pointing out many of the flaws in using Linux software on the desktop. Linux developers would be well advised to read it and take the author's complaints in mind.

    One thing the author pointed out repeatedly was the problems involved in installing and configuring XFree86. People have been saying this for a long time. I know that just about every time I have installed Linux, one of the first things I have to do is rerun xf86config and then manually edit the files to get things working right. If the Linux companies out there really want Linux to take over, the most important thing they could possibly do would be creating an entirely new configuration tool for X that is easy to use, configures scroll wheels, and has a better interface for less technical people.

    Another sore point was StarOffice all being integrated together. I know that big changes are planned for OpenOffice 6, but we really need the Linux vendors to rally behind Koffice and Star/Openoffice for speed, ease of use, and file portability, as well as better Microsoft Office compatibility.

    Of course, what I saw above all that really stood out was the fact that Linux is being compared to Windows, as if Windows needs to be as good or better than Windows at what Windows does. Linux will never be better at being Windows than Windows. Linux desktop developers need to stop cramming every little tool that might be able to fill a Windows-like function onto Linux desktops and start doing something special and innovative. Microsoft has spent years ripping off Apple's ideas, and all we get are good knockoffs of a knockoff - which is never going to put Linux up front where it needs to be.

  • LINUX VIRUS DETECTOR SAMPLE CODE (Score:3, Funny)

    by mcrbids ( 148650 ) on Thursday September 20, 2001 @11:58AM (#2325414) Journal
    #! /bin/bash

    # Let user know we've started
    echo "Scanning system for viruses"

    # Make him think we're actually doing something!
    find / -type f >> /dev/null 2> /dev/null

    # Report the results
    echo "No viruses found"
  • He mentions Outlook as an e-mail client for windows and that there is Linux alternatives. Sure, mail and pine are great alternatives for MAIL. What about the OTHER functions of Outlook/Exchange? Tasks, Calender, Folder sharing, Centralized Contact database, Journal, notes.... Okay *some* of these things can be fixed by implementing an Imap server and ldap (folder sharing and contact database respectively), but what about everything else? Granted, this can all be done in OTHER programs but the fact that outlook has all of these features in 1 program makes it superior to anything currently on Linux!

    Not to mention, how easy is it to setup sendmail, an imap daeemon, an Ldap server? You need REAL TECHNICAL ABILITY. For an IT firm this is fine, but for the general business use it will not work. Most corporations will not be willing to spend hundreds of man hours setting something up and then now have a support contract to fall back on.

    Now... what about domain logins? Sure, this can be done under Linux... but it's not a turnkey solution like under Windows 2000. In Windows 2000, start the configuration wizard, make the machine a domain controller and log the other machines onto it. Under linux... uhm kerberos? Sure, now figure out how to use kerberos... modify those config files! Or NIS? NIS+? Even worse... sure, possible, but not a turn key solution!

    Now we go to Samba... why use samba? Because it's GREAT! But the original intent of Samba was to allow filesharing between Windows and UNIX (not just Linux). It's great for that and has expanded, but what about access controls? You can control access via samba using /etc/passwd on the server... or by using the db file you can specify in smb.conf... or maybe, if you're a real hacker and AGAIN want to spend hundreds of man hours, you can implement an ACL system based on the same LDAP implementation you used for contact information....

    Windows 2000 + Exchange = 8 hours w/ all M$ patches they have released WITH implementation!

    Now we go to the firewall on the desktop. First, why would one want such a thing in a corporate environment. And even if we had one, anything under Windows is better than ipchains or iptables. Maybe ipchains and iptables is BETTER but it's also HARDER and more COMPLEX... this is a big difference. It requires a higher level of skill from the end user to configure their firewall for everyday use. Now, I do work in an IT environment and even though I work with some smart people, they still fudge up their ipchains rules at home and ask for help! What about the corporate end user? Yeah, I'd be getting calls all day AND night with a nice queue that grows exponentially.

    What about version management? Sure, redhat has it.. but does it come w/ something like SMS where you can rollout patches and programs to every RedHat box on the network with a click of the mouse?

    One thing Mr. Valliere didn't mention in all his pricing was the amount of time required to set all these systems up. The number of man hours MUST be included in the cost of setting this up. Even the most skilled IT person with years of Linux experience in all the products I mentioned would take a good 2 to 3 weeks to implement w/ all the hacking and testing. At the same time, admittingly having NEVER really worked with Exchange or Windows 2000 server, I could guarentee a Windows 2000 network with Exchange and Office on the same number of desktops in less than 1 week!

    My point being, although I will be mod'd down for trolling, is that although M$ isn't the BEST solution for everyone, it's the BEST solution for most corporate environments because it's 1) easy to implement 2) is designed for the corporation )file sharing, task sharing, calender sharing, etc) and 3) they can get support contracts in case ANYTHING goes wrong!

  • It's all about the money... (Score:3, Insightful)

    by Medievalist ( 16032 ) on Thursday September 20, 2001 @12:37PM (#2325623)
    /.
    Somebody was asking about the cost of training. In most businesses today there is no end-user training for office apps. So that issue is usually a washout regardless of other factors; the few places that do train will do so on whatever software is in use.

    Somebody else said "Why use an open-source clone of a M$ application? Why not use the real thing?"

    Here's why: Four years ago my employer was paying over $800,000 US per year for software licensing costs. Today, that figure is less than $50,000 yearly. Linux, samba, rsync, and OpenSSH are the reasons why.

    What's being done with all the money available from the avoided costs? Well, some of it is in my wallet right now. The M$-addicted IT directors out there don't control your paycheck nearly as much as the penny-pinching Smiling Men. And the accountants like to see those recurring costs dropping.....

    Another thing I'm doing with that money is killing off Solaris, NT, and SCO-unix (and of course that dreaded train-wreck of a unix, HP-UX). This results in easier maintenance, and thus more time to work on the holy grail of a totally free desktop. I honestly don't care if it's linux as long as it's open source and doesn't require constant re-purchase.

    --Charlie
  • Nobody uses a computer in order to use an operating system. An operating system enables people to write other software, and that's the software that is important. Focusing on the OS above all else is a distraction.

    The trouble with attempting to clone the Windows environment is that we're getting 80% of the way there in almost all cases. Star Office is 80% of Microsoft Office. KDE is 80% of the Windows desktop. Mozilla is 80% of Internet Explorer. The total result is that Linux environments feel unfinished and shabby compared to Windows.

    Cloning, by definition, is doomed to fail because it is a game of catch-up. A better approach is to think "What exactly do people _need_ to do?" as opposed to providing big and bloated toolsets which do all sorts of irrelevant things (for example, people don't want or need to be able to design a custom GUI for each application). As the risk of being considered a troll (and I guess light criticism is always considered trolling at Slashdot), I think that many people developing for Linux are not looking to "scratch as itch," as ESR likes to say. Rather they're gung-ho about putting Microsoft out of business by attempting to reproduce a popular Windows application.
  • you all are hitting that poor windows/asp server and it is giving "Server to busy"... gee if people only knew how to scale ...
  • The real pitfall of Linux is it isn't very user friendly. In this I don't mean it's complex or clunky that is more or less incorrect.
    Linux needs a considerable amount of computer skills to install configure and use. That is what makes it not friendly. But once you have those skills Linux runs as smooth as silk.

    That is what makes it not suted for the desktop. A desktop should run smooth as silk with limited office skills. Not computer trainned computer experts but the high school or collage interns. The avrage low end office population.

    So the best solution is Windows?
    I think not..
    As of late Microsoft has favored 'skilled end user' solutions to problems in Windows over real solutions. In part becouse some of those problems can not be fixed with out removing a significant feature.
    The results are an office worker shutting things down for a few days.

    The reality is you train your office workers. You have no choice. With this in mind the whole 'user friendly' advantage is a fantacy.

    The defects in Windows really eats into productivity.

    Finnally when you update the hardware get Mac hardware not more PCs. You don't have to train your users and they work hand in hand with Linux systems. Both are Unix like so you can run most of the same code on both platforms. Throw in BSDs and Solarus boxes and everything works perfictly. Just drop the right box on the appropreate desk. Not everyone need run the same system when they all work together.
  • I know this horse has been beat down to the ground, but it still bugs me. How can Linux get anywhere on the desktop without good performance? Win2K runs great on my relatively old 300MHz machine, but Linux/KDE seems much less responsive. It just seems to me that the Linux desktop is too much of a compromise. You can have cool features with KDE2, but you abandon performance. You can have a super-fast desktop with E17, but you have to put up with a tiny software base that actually uses E17's features natively. Win2K manages to get lots of features with good performance, so there is no reason to have to choose the lesser of several evils on the Linux desktop!
  • From the web page:


    "Booting with Lilo


    Linux provides a very good and easy to use booting facility using Lilo. On the dual OS desktop unit that had an existing Windows 2000 installation, the desktop boots to Linux and a menu appears: I can select Linux (now my default) or Windows."

    From personal experience getting NT/2000 and Linux to dualboot properly for many years, I have to say this statement is a bit of a misnomer. RedHat 7.1 clearly doesn't create a dualboot properly with NT (Disk Druid barely recognizes that the drive is NTFS!) The author comes off as making it sound easy to dual boot between the two completely different OS's, whereas doing anything close with 2000/XP today is a bitch and a half.

    My best experience is to simply make a boot floppy for Linux and load as necessary. No muss, no fuss. Until one of these distros properly configures lilo, anyhow.

Slashdot Top Deals

"Here's something to think about: How come you never see a headline like `Psychic Wins Lottery.'" -- Comedian Jay Leno

Close