Linux

10 Windows features I would like to see in Linux

Jack Wallen recently looked at how certain Linux features might improve Windows. Now he lists some things about Windows that could improve Linux.

Jack Wallen recently looked at how certain Linux features might improve Windows. Now he lists some things about Windows that could improve Linux.


I recently shared my list of 10 Linux features I think should be included in Windows. Today, I'm going to challenge myself by finding 10 features in the Windows operating system that I would like to see make their way to Linux. I am not going to play the typical fanboy and make a joke of this by saying there is nothing in the Windows operating system that would be welcome in the other camp. We all know there are plenty of outstanding features in the Windows operating system. But I might stretch the nature of the word "features" to include a few items that are less inherent in the OS and more about the community or business model.

So with that said, let's dive into this ocean and see what we catch.

Note: This article is also available as a PDF download.

1: Marketing

I have to start with the big guns. There may be only one IT-related business with a better marketing machine than Microsoft -- Apple. But that's a big "may be." And everyone knows how small and inefficient the Linux marketing machine (or lack thereof) is. I feel fairly confident in saying that if the Linux operating system could enjoy the marketing that Windows enjoys, no other operating system would stand a chance.

2: Hardware support

I say this somewhat half-heartedly, because the hardware support Linux enjoys has come such a long, long way. But there are still areas where it could use a huge bump. Specifically, wireless. Most often a lack of a working wireless connection in Linux is due to having an unsupported chipset. And although the list of unsupported chipsets is getting smaller and smaller, it still exists. When potential new users come across an issue like this, they inevitably run back to Windows because they know their hardware will work. They may have to spend an hour (or a day) looking for drivers, but they know they can get it to work.

3: Smart phone syncing

Regardless of the type of smart phone you use, one of the biggest benefits of using it is that you can sync it with your PC. At least you can in Windows. Many smart phones quickly become slightly crippled when plugged into a Linux machine. Even my HTC Hero, which uses the Android operating system, can't sync with Linux. Yes, you can add music to your Android phone. But try to sync contacts, calendars, or email with Evolution or KMail and you're in for a never-ending nightmare. On the Windows operating system, this task is a complete no-brainer.

4: Enterprise presence

On so many levels, Linux is a perfect match for SMB and enterprise usage. Be it the desktop or the server, Linux could help improve the efficiency of workers. But that has not and could not happen without some real change. Exactly what that change is, I am not sure. But I do believe most of the change needed is on the end of the business -- and we all know that is not going to happen. But if Linux could enjoy the presence that Windows has in the enterprise, the whole landscape of IT (from business to home use) would change.

5: Workgroup setup

I can get Samba set up pretty quickly, but that is after years of working with Linux. The average user would seriously be put to task to get this working. Joining a Windows machine to a workgroup is simple. Linux needs to gain this user-friendly ability to see and work with Windows machines with very little setup (and especially no editing of smb.conf).

6: Touchscreen

One of the big to-do's with Windows 7 was the improved touchscreen support. Linux can work with a limited number of touchscreens (see #2), but to do so often requires the user to work with the xorg.conf file. And since X11 is now working with a xorg.conf-less setup, this is even more difficult. Although I'm not a fan, touchscreen could be the future of computing. It has worked majestically for the iPhone, so why not for the desktop PC? If that's the case, Linux better get some Windows-like support worked into the picture.

7: Pre-installs

This could easily dethrone #1 from the top spot. A handful of companies (System76, for instance) offer pre-installed Linux solutions. If anything would give a better boost to Linux acceptance than pre-installs, I'd like to know what it is. Pre-installed operating systems are what gets the OS into the hands of the user. Sure, anyone can install an operating system if they want to (and have the IQ to do so - and we're not looking at Sheldon Cooper levels of IQ), but this doesn't happen on regular basis.

8: Support

This is a tough one. If you have a problem with Windows you can call Microsoft tech support (so long as you have the time). If you have a problem with Linux, who ya gonna call? You can call Canonical for Ubuntu support, if you've purchased a support package. You can call Novell for SuSE support, if you've purchased a support package. You can call Red Hat for Red Hat support, if you've purchased a support package. But what happens when you buy that shiny new computer, wipe off Windows, install Linux, and have a problem? Most likely, you're going to hear that you have invalidated the warranty or support contract by doing so. PC makers need to learn to support the Linux operating system.

9: Software installation

I want to preface this by saying the Ubuntu Software Center will eventually negate this point. But for now, we'll continue on as if USC doesn't exist. To install an application on Windows, you simply download the installer and double-click the file. To install an application on Linux, you have to search for the application in a tool like Synaptic, mark it for installation, and apply the changes. After you click Apply, you have to hope that all dependencies have been met. And if you can't find the software within Synaptic (or whichever tool you use), you have to add the repositories that house the software you need. I am a big fan of how Linux is evolving (thanks to tools like Synaptic and Ubuntu Software Center). New users expect to be able to download a single file and double-click it to install.

10: Direct X

One issue that keeps many features from migrating to Linux is Direct X. What would this do for Linux? In a word, games. Games are the reason so many people will not migrate to Linux. There are a lot of gamers out there, and until Direct X comes to Linux, those games will not find their way outside of any operating system that does not support Direct X.

More features?

That's my list of Windows features that Linux could benefit from. But is it a complete list? If something is missing, tell us what the feature is and how Linux would benefit from having it.


Check out 10 Things... the newsletter

Get the key facts on a wide range of technologies, techniques, strategies, and skills with the help of the concise need-to-know lists featured in TechRepublic's 10 Things newsletter, delivered every Friday. Automatically sign up today.

About

Jack Wallen is an award-winning writer for TechRepublic and Linux.com. He’s an avid promoter of open source and the voice of The Android Expert. For more news about Jack Wallen, visit his website getjackd.net.

138 comments
wgd
wgd

I'm always hearing the mantra that [GNU/]Linux "has to" compete in the desktop space. The value of this is obvious as far as momentum, support, etc. However, similar to that discussed in the various BSD articles, there's a reason why [GNU/]Linux, along with all other Unix-like OSs, is best suited for the server space, "power" users (whatever the heck those are), development environments, and custom installations (e.g. cash registers, library kiosks, etc.). The more fundamental problem here, though, is that [GNU/]Linux and other OSs will *never* lead as long as they are being led by whatever Bill is doing. "Linux needs more marketing because that's how Bill got Windows to be so popular". Pardon me, but this article already mentions how amazingly, ridiculously popular [GNU/]Linux has gotten almost entirely *without* marketing! I've worked in some environments where having a GNU/Linux desktop was not a problem, sometimes even encouraged. The network protocols were open, so what one ran on one's desktop was largely a matter of preference. I've worked in other environments where running other than Windows just wasn't a viable option, though there were usually GNU/Linux and/or BSD servers somewhere in the enterprise. Are "we" trying to force people to use something other than Windows?

RexBallard
RexBallard

The first one - marketing - is probably the biggest. Microsoft and Apple are fanatics about insisting that their trademarks be well displayed. Linus has never been a big one for "marketing Linux". He filed for the trademark to keep some poacher from trying to demand $billions for a fraudulently filed trademark application. Linux doesn't even require that any Linux "logo" or even the Linux trademark be displayed anywhere on the displays or even the packaging of the billions of devices powered by Linux. Linux powers most WiFi hubs, many cable set-top boxes and many DVRs as well as most HDTV systems, many routers, and millions of servers. It's ironic that most enterprises, especially large-cap organizations use Linux servers as well as Unix servers, and often for very high-performance production applications, and Google, which is Linux powered, has become a verb because it's so widely used. The "lack of enterprise presence" is mostly because there is no requirement to put "Tux" (the cute little penguin) somewhere on the screen of the cash register, or a gnome footprint, or the KDE gears - anywhere on the screen. You probably stare an Linux displays when you go to the grocery store, when you order at a drive-through, and when you use a kiosk. Pretty much, if you don't see the Windows logo prominently displayed, the device driving the display is probably Linux. As for nice GUI configuration tools, Linux has them - several of them. Over the last 15 years, Linux has probably had about 20 different GUI interfaces for some of the most common functions such as configuring a network device. Writing a GUI interface for a command line application using python, perl/tk, or one of the other GUI tools, or using a web interface is a relatively easy exercise, something that a freshman in college could do easily - and many have. Hardware support is an interesting situation. Many vendors do offer "binary-only" modules, which can be use with probes that can detect the device and automatically install the drivers. On the other hand, many distributions such as Debian, don't like to support these binary drivers, so the source code available version is used, but this version is often very limited. Most commercial distributions do have support for commercial drivers, and install them automatically, after asking for permission to do so. "Easy Workgroups" is a mixed blessing. Many of the things that make it so easy to share files on Windows workstations is also what makes these systems so easy to infect with viruses, and to infect thousands of other workstations and file servers in a matter of just a few hours. Linux security can be a pain in the neck sometimes. On the other hand, Linux does support much more secure file sharing systems such as NFS-4, AFS, and GFS, which provide features like auditing, security, real-time authentication, and consistent security. Even better, control the access by using your Linux workstation as a web server, which limits the ability of others to put things that shouldn't be there. Linux, Apache, MySQL, and PHP have also made it possible to easily share data, and make sure that it doesn't get corrupted. Coordinating shared spreadsheets can be a real drama, and if you have text field, the spreadsheet paradigm can quickly become quite unruly. With PHP forms and reports, it's trivial to create attractive forms, let user update their own information, and provide consolidated reports in a fraction of the time it would take to try and consolidate and merge spreadsheets. As for pre-installs - there are a number of companies such as Emperor that offer brand-name computers configured with Linux. The top-ten OEMs have tried repeatedly to offer Linux, but Microsoft has it's own ways of dealing with these attempts. Microsoft typically uses steep discounts based on minimum commitments which are far more than the OEM can actually hope to sell. At the same time, if the machines are shipped with Windows the OEMs are not allowed to alter the configuration without prior written permission from Microsoft. Putting in a new device driver, or an anti-virus package usually takes a few days. Getting dual-boot dual-partition Linux/Windows systems approved - well - I don't think it's ever happened. Microsoft did approve Apple's request to support dual-boot, but that was because Microsoft wanted to get into the Mac Market - to keep OS/X Macs from displacing Windows PCs. On the other hand, more and more vendors are making their system "Linux Friendly" or "Linux Ready". Using video, network, and WiFi components that are fully compatible with Linux. Ironically, because Microsoft has tried so hard to resist Linux, Linux distributors have made Linux remarkably easy to install and configure. These days, you can install Linux using a USB flash drive, an external USB drive, or DVDs. You can install virtual Linux to run as a Windows application, and you can install Linux so that it can run Windows as a virtual application. In addition, competition between Linux vendors is very aggressive, with Vendors competing to provide the best blend of performance, power, efficiency, and robust applications for their particular market. The needs of a corporate employee are significantly different from a home user, which are different from a student/gamer. Ironically, Linux has support for USB-3, Windows doesn't. As for touch-screens - the displays used on CNN are UNIX based displays, and X11 based. Multi-touch screens are not that big a challenge. The Windows 7 desktop look remarkably similar to the latest Gnome desktops. Of course, Linux has virtual desktops. Linux can also support multiple X11 screens, a feature that has been around since Motif back in 1992. Software installation is becoming less of an issue. Distributors are pretty good at staying on top of the most popular applications. A commercial "Office" distribution such as SUSE Linux Enterprise Desktop, or Red Hat Enterprise Desktop will typically have the support for commercial and proprietary drivers, commercial software, and will pre-test to make sure that their distributions support the strategic commercial software. On the other hand, a "freebie download" such as the Debian based distributinos are more likely to have a smaller subset, or will download the basics and then let you install the latest versions of lots of free software via the internet. There is generally less concern and less testing of commercial software or drivers, especially on the newest releases. As for Direct-X, Linux supports OpenGL, as do most of the video cards. At one point, Direct-X 10 cards did not support OpenGL and the result was a rather significant drop in the prices of those PCs - because corporate buyers are testing all these new machines with Linux. The video chip makers learned their lesson fast, and now most chips support optimised OpenGL as well as DirectX. Linux has also redefined what is possible in Gaming. SecondLife and other virtual worlds applications have turned into a "killer app" for Linux. Running SecondLife or even the Emerald Viewer on Windows is ugly. Motion is jerky, lag is terrible, and sometimes your screen just hangs there. Running the Linux versions on Linux is a whole different experience, with smooth motion, nice smooth sound, high resolution, and all of it in real-time. There is often little or no lag, and when there is you get a really nice animated display until you "bounce back" to where the server wants you to be. I'm not sure how SecondLife will evolve as a collaboration tool, but I'm almost certain that some form of it will evolve this way, and it will be Linux desktop and laptop users who will benefit.

guygo
guygo

My biggest beef with Linux is still it's inability to handle new hardware easily. Until manufacturers start providing Linux drivers & installers, hardware is at the top of my list.

granholmk
granholmk

Direct X is a Microsoft invention, but Linux does need a similar type product for Games to interface with.

yetisouth
yetisouth

The 11th thing LINUX needs is LINUX users with manners. I tried to make a few humble steps into the LINUX world (I admit, I know nothing about it). I asked some questions on a forum - bad idea! I got answers such as: Use Google you idiot Another Windows Nut (not followed by an answer) I then gave up, like many others before and after me.

FrankH
FrankH

Unsupported wireless is the first, last, and only reason I am not using Ubuntu right now.

colonel.jack.o.neill
colonel.jack.o.neill

I have been "playing" with Linux ,in preparation to eventually switch,but right now I have a major problem............. I can't get my creative labs xfi soundblaster to work.I have downloaded what appears to be a beta driver from creative but I can't install it.........so to listen to anything I have to stick with windows :-(( Hopefully soon someone.........maybe creative ;-) will come up with an easily installible driver for this line of sound cards.

lsorense
lsorense

The software management in ubuntu and debian and such is vastly supperior to what windows has. It is solid and reliable unlike windows where every software company tends to do their own installer design and not all of them know how to do it right. Linux can upgrade every application and library on the whole system from one place. Windows can't. Direct X on linux is not the solution. Using SDL and OpenGL and OpenAL is the solution. It works everywhere (linux, mac, windows, solaris, etc). A few game companies have figured this out and hence have native cross platform games. SDL is also easier to program for than Direct X in most cases. As for hardware support, I find that once a device is 4 or 5 years old, it often stops being supported in windows, while linux will support it just about forever once it has a driver for it. Many devices still don't work with windows 7 (or even vista in some cases) even if they are only a couple of years old. If the manufacturer discontinues selling it, then good luck getting drivers when the next windows version is released.

Neon Samurai
Neon Samurai

Three tries to post and each one: "You must enter text in the field before you can post" What the F does the forum software think the 12 paragraphs in the form field is?

ScarF
ScarF

Well, think like this and Linux will need other 100 years to come close to Windows. Hardware support? Mac does better with less. Where is the software support? There are no killing applications running on Linux. I cannot even think for installing Linux as long as there isn't an AutoCAD version available for this OS. Groupware applications? Forget 'bout't... Adobe whatsoever? What kind of bad jokes are that open applications trying to mime Adobe? Multimedia editing and producing? Ha. Games? One would believe that being a geek passion, Linux should run tones of games. Well,... not quite. There are more games available on an iPhone. Until real applications will be migrated on Linux, this is nothing but a toy - for workstations. What about the user experience? Sad. And, something else: responsibility, liability. Not that MS gives a rat's a** about its customers, but I have someone to throw my anger to. Plus, there isn't always MS to blame. Autodesk has its history, also. When something goes wrong with a Linux workstation, who am I going to blame? The cloud of unknown programmers working from home in their spare time? Linus Thorvald? Jack Wallen? God? Jeez. This is coffee talk. Grow up, Jack.

TheProfessorDan
TheProfessorDan

I think that number eight hits the point. If I decided that tomorrow I wanted to become a Linux Expert, I would have to learn Red Hat and Unbuntu and SuSe. As cumbersome as Windows is, Microsoft has a huge advantage over Linux because essentially each flavor of Linux competes not only with Windows but with each other. I am realistic, I don't forsee each of these entities uniteing against the evil empire. On the other hand, do they need to? One of the trends in business is successful companies going under when they forget what they are. Pepsi will never overtake Coke but that works for them. On the other, the Arena Football League was very successful for years catering to their customers. It wasn't until new investors stepped in and tried to make them bigger that they failed. I see Linux in the same light. Linux has it's niche and as long as they continue to take care of that group, they will do fine.

Neon Samurai
Neon Samurai

Actually, it started with windows. Between the Soundblaster 16 (looong time ago) and the Gravis Ultrasound, the second provided better professional sound quality but the first better supported the software I was running at the time. (Anyone else remember the sb16 simulated compatability in the Ultrasound?) But, the point is buying hardware that supports my needs. One core need is cross platform support. Between the newest graphics board supported by only one platform and the second newest with support across multiple platforms; I'll take the second. The bits of hardware that will only ever be well supported under one platform are few and far between in my purchasing history. Maybe if more consumers with cross platform interests voted with there wallets, hardware vendors would take more notice.

Neon Samurai
Neon Samurai

You'll find equally hostile users in any forum. Had you asked in a different forum or here on TR, you would have gotten a much more civil response. However, if your posting to a forum it may be assumed that you already have a browser available and as such "www.google.com" would be accessible. The real question is; are you still turned off by the experience or would "how to get started" information still be of interest? http://www.linux.com/learn/new-user-guides

Slayer_
Slayer_

Followup: How do use Google without a web browser?

lsorense
lsorense

Given the number of supported wireless chipsets today from intel, ralink and atheros, I think it is simply bad purchasing decisions if you have trouble with wireless under linux. My 3 year old linksys WPA54G only has drivers for XP. Not vista, and certainly not Windows 7. Linksys simply doesn't care. Works perfectly in linux though because ralink supports their chipsets. Now fortunately if you realize ralink cares you can in fact get windows vista and 7 drivers from ralink and manually force them to install for the linksys and get working wireless in the newer windows, but it sure is a lot harder than under linux where it just works.

Neon Samurai
Neon Samurai

1: Marketing Marketing relates to distributions who's goal is retail success. For those, marketing would help. Connonical, Red Hat, Novell, Mandriva.. don't be shy. For the non-retail distributions, I'd suggest marketing in so much as it helps Point #2 2: Hardware support Yes. Ideally vendor support in writing open drivers but even if the vendor wants to make there own secret drivers. Yes. It is getting better but the platform type is still looked at like a second class citizen if not with outright contempt by some of the hardware makers. 3: Smart phone syncing This one is all about the smartphone providers. Thank you for using Linux based firmware. Please provide sync with Linux based OS. This "good enough for the device but not good enough for the user" crap is not acceptable. Seriously.. Android has no Linux sync but it's key selling feature is that it's based on Linux? 4: Enterprise presence Some of this is technical, much of it is political. Of course, part of this would be the ability to connect to MS-LDAP. Right now, Novell has the best chance of gaining near full Active Domain support. Applications are also the big sticking point here. You need an Outlook type app that can make use of the existing Exchange server. You need the office productivity software. 5: Workgroup setup This depends on the distribution. In Mandriva I hit the system settings, open the "windows network shares" utility and it's not really any more complicated than mapping a Windows drive letter to CIFS share. I'd actually rather see this go the other way and be able to replace CIFS with something secure like sshfs. 6: Touchscreen This comes very close to #2 for touchscreens that simply emulate a mouse. My CF27 and CF25 where great to use with keyboard and touch screen. The limitation was hardware support in Windows only. Other than a missing driver, i see no reason why it would behave differently under a Linux boot. 7: Pre-installs One can't underestimate the benefits of preinstalls. The person providing the hardware ensures that the included OS supports the hardware. Maybe this will help #2 along. 8: Support I see support from the parent company about the same on both sides. As you say, call Microsoft, Connonical, Red Hat, Novell for there specific distributions (Windows is a distro after all). More techs willing to support Linux for those who do choose to go it alone would be nice and that is actually on our shoulders as support techs (get out there and learn the platform you win-only techs). The user has to take some responsability here too though; if you choose to go it alone with a non-retail distribution then realize that your first line of support is non-retail sources like forums and reading. 9: Software installation I can't agree here. Mozilla has native packages ready for download. drop the .deb or .rpm on your desktop and double click it. How is this different from downloading a .exe and double clicking it? How is pulling software from a multitude of websites better than bringing up a central list of software and selecting it through the checkboxes? By all means, both platforms should keep the different ways of install available but this is an area that Windows could really learn from the software hippies if MS could fix it's business model enough to allow for software packages that compete against MS own offerings. 10: Direct X It would be nice but it marks this as a whistful wishlist. I don't see MS releasing DX outside of the MS product line any time soon. It's been too valuable to them as a lock-in device for gaming. And if one doesn't think gaming is important, consider how many Linux detractors open with "well, when I can play my Crysis on Linux then I'll consider it". The best we can realistically hope for is enough catch-up development in OpenGL to provide the same sort of complete cradle for games that DX provides.

Slayer_
Slayer_

Eh? Most peoples BIOS's have more native support for hardware than a MAC does. Don't forget, Apple sells computers, Microsoft sells Software. Because Apple sells computers, they only have to support their own hardware. Microsoft and Linux OS's have to support a lot of drivers because they are OS's designed to work on near limitless hardware configurations. I guarantee you if you go plug in a nice shiny new sound card, it likely won't work natively on a Mac. But probably will under Windows (to an extent) or Linux. You also cannot simply install OSX on any system without a large amount of driver difficulties.

j-mart
j-mart

Its not that far away http://www.bricsys.com/en_INTL/bricscad/index.jsp This Autocad clone does the job well. At my place of work our main CAD package is Solid Works but from time to time we require Autocad for dealing with Client's dwg's and for old legacy dwg's in Autocad format. Windows version can be run under wine but Bricscad can see a possible market for a Linux version

Neon Samurai
Neon Samurai

As will professional multimedia production though in this case, it may also dictate the use of Linux based systems in areas. In terms of playing multimedia; it's a non-issue and especially if we're talking retail distributions. Responsability; stick with RHE, Suse or Ubuntu and you can focus your ire on a single entity behind each. Mandriva has a single entity behind it also. Go to town. It'll do as much good as directing ire at MS will do though of course. By using MS software, you agree to absolve MS of any responsibility so what's the difference? Either way your the one left holding the bag. Groupware is not unique to Exchange Server either. There are pleanty of groupware system to look at. eGroupware does great for my needs and won't demand that I use Outlook to gain full functionality. user experience is pretty subjective also. Your discomfort with a linux based system does not make me any less comfortable on a linux or windows based system. I gotta go back to this point to end this out. You recognize that blame can be directed at MS for the OS or Autodesk for AutoCAD issues yet ignore the exact same method of targeting blame with linux distros (distro provider, major application provider, driver provider and so on). That's interesting.

Neon Samurai
Neon Samurai

You'd have to learn each different model of car. Concepts between cars (suck, sqeeze, bang, blow) are not going to change but you'll find that they are all very different product providing compatible functions. With Linux distros, there are differences and common things also. Learn Debian and your covered for Ubuntu and any other .deb derivative. Red Hat covers Mandriva and the rest of the RPM distros based on it. I don't see it as much better on the other side either. Right now you'll need to know XP, Vista and w7 along with each of there personality differences. Add to that the different server features and your talking just as much learning. You may be hitting the mark dead on in your last points. Distributions are destinctly different entities even if they are interoperable and assembled from the same parts bins. It is just as important to recognize the strengths of the distributions your looking at to choose the right one. In that regard, no product is any different and each could fall over by loosing site of it's goals.

CharlieSpencer
CharlieSpencer

Linux is a niche only on the desktop. It's got great installation numbers on servers and embedded devices, where it's competitive with (and arguably more successful than) Microsoft products.

Slayer_
Slayer_

They still have the best backed product. Everything supports them and they support everything. most generic drivers work as well for those DOS games and programs. I notice its often the integrated crap that usually has troubles in Nix, not actual cards from hardware manufactures. The real culprit of Nix hardware problems seems to be caused by cheaply made computers.

ScarF
ScarF

#2, #6, #7: With 1% market share, Linux is a second class citizen for the HW producers. Actually, a third class - following Apple. Linux doesn't even have a declared mission - something like "in 5 years Linux will replace Windows on desktops", and start working around this statement. Regarding your comments about the Windows guys, let's cut the sh*t with this, once and for all: I don't consider myself a Windows guy. I may be called equally a UNIX guy, since I am certified by IBM on AIX. I am just another tech dude earning his money for bills and taxes by helping the business he's working for and its users. Do the Linux adorers have something against it, this is their problem. I don't even care what system I support as long as it satisfies my users and the business I work for. BTW, against the management proposal to migrate to Vista, I opposed vehemently - so, it isn't that I have something against Linux; I have something against the delusion. And, something else. I worked on both sides of the fence. I was employed by IT companies providing services to non-IT businesses. I love working for non-IT, though, and here is where the retirement will find me: giving a hard time to the IT companies always trying to sell green horses.

CharlieSpencer
CharlieSpencer

8. Support. i don't think it's a question of software vendor support as much as hardware vendors. As someone else pointed out, try calling your hardware vendor after you've replaced the pre-installed version of Windows with Linux. Expect to get told you've violated the warranty.

ScarF
ScarF

This is interesting. Thanks for sharing. However, our department using AutoCAD is the company's turntable. Plus, there is another package running over AutoCAD, and the producer doesn't have any plans for Linux. Eh. Thin industrial vertical. Than, puting Windows under Wine doesn't simplify the things. On contrary. Plus it saves no money for licenses. Why should we do this? From love for Linux?

ScarF
ScarF

Since you were so kind to answer to my post, and I have already noticed your interest in Linux, I have a question related to the liability issue: How can one choose a Linux distro? What are the criterias? You understand that there are quite some flavours on the market, and from time to time new distros are launched. Some time one distro may be better than others, other time may be worse. With Linux, this tends to change quite fast - which makes the things messy. For servers, I have a personal preference for Debian. But, what about the workstations? I have always been confused by the too many available distros. One may say "This is the freedom of choice". It may, when using the system for my own activities. But, when I have to decide for the many users I have, I have to choose one only. More distros would become an administrative burden. And, who may say that the selected distro will satisfy all users? Again, Windows is what it is. But, it's one. When I have to upgrade the systems - which is already planned for the next year, I do it for all users at the same time - planning some user training, also. I have to deal with a single OS on the workstations which gives me consistency. Hm. I stick with my opinion. On workstations, Linux is a toy "pour les connoisseurs". And, shortly about the groupware. Exchange is self-proclaimed groupware. We use Domino with Notes clients. Do you have any knowledge about a Notes client on Linux? Again, Notes client is mandatory for leveraging the entire security model provided by Domino, plus all the other goodies. eGroupware doesn't even dare to dream going to that level.

Neon Samurai
Neon Samurai

One unexpected benefit of my 5.1 headphones; I can run different sources to different channels. It would mean giving up the seround sound for your needs but it's a quick easy way to get two sound sources into the same headset. (now if NWN2 would stop resetting my audio to 2.1, I'd be happier. XFI says 5.1, headphones confirm 5.1 with the channel test, run the game, XFI control panel says 2.1 suddenly. Seems I'll need to confirm that the config file setting is still in place.)

Slayer_
Slayer_

I run the integrated to the big screen TV and the SB to my PC speakers. This way I can watch movies and play games at the same time without the audio crossing. Did I mention I multitask the shit out of my computer :p

Neon Samurai
Neon Samurai

Gravis did provide a great card but all the software provided Creative support. I recently stuck with the daughter board shipping with the Asus motherboards. While it's also a good card, the lack of universal games support killed it for me also. It took some slot juggling but the XFI has been rock solid since it went in and there's still lots of air space around the GPU board (it won't fit a second or third without pulling other boards out though). I still remember the winmodem days though so buying hardware selectively is a long ingrained habit now.

Neon Samurai
Neon Samurai

Most people have a preferred brand or model of car. Other's do the research when planning a purchase; read reports, test drive. Some go to a gearhead friend for a recommendation. Choice may also be limited by what dealerships are within walking distance of one's home. You figure out what your going to be using the car for, check what it's maintenance history is like then figure out if you like the feel of it. A new driver is not well served by a high performance car (put your 16 year old two days past the license testing into an F1 and things go badly). If your going to be hauling loads, you won't be looking at small hatchback cars but pickups and other vehicles meant for hauling cargo. With distributions it's the same; figure out what you intend to do with it. This shortens the list. Stick with the top major distributions and the list is reduced again. Test drive the distros (liveCD) and see how they feel and if they come close to filling your needs. The distribution chooser website is a great starting point to figure out what distributions should be on your short list. In this case, the new driver should start with the new user distributions not be expected to install and admin OpenBSD two days after learning to mouse with Solitaire. In both cases, one is not choosing from the entire universe of possible makes and models either. I wouldn't suggest someone's car buy list include everything down to some brazilian or south-asian model any more than I'd suggest someone's distro pick list include everything down to the most obscure specialized distributions out there. Sixty or six thousand doesn't matter because I'm looking at a short list of ten or less with both product purchases. The point you do bring up though is what happens second. If the first talk is about clarifying distributions as the assembled product not the bolts and steal plats it's assembled from. The second talk would be about how to choose a car/distro for one's own needs. It's actually the more interesting talk but like cars, it starts with "what kind of driving will you be doing?"

Slayer_
Slayer_

If we imagined that all OS's functioned perfectly with eachother, its harder to choose. If there is 60 great choices of sports car available, why is that easier than 60 choices of sports OS's available?

Neon Samurai
Neon Samurai

My point was that different distributions including distributions of Windows are different products. This in response to the general demand that "Linux" needs fewer distributions stated by people who ignore or are unaware that separate companies provide distributions which happen to use the Linux kernel. As my Q/A commend a few back starts; we demand consistence across distributions with completely different design goals yet we don't also demand strict consistence among those things which make separate cars different from each other. The distribution is the product, not the platform kernel. The car is the product, not the engine. One does not buy a VW bug for pulling the family camper because the engine is unsuited but because the entire car is unsuited to the task. Of course this comparison falls over at some point but it does provide a more generally understood product category where multiple brands based on the same commodity parts is ok. The comparison with car manufacturers was to highlight the difference in branded distributions equal to the difference in branded vehicles. GM and Ford both use an engine, bolts, headlamps.. commodity parts assembled differently to result in different products. Red Hat and Novell both use a Linux kernel, GNU tool chains and other similar commodity parts assembled differently to result in different products. Both GM and Ford can interact with each other and the same road systems. Both Red Hat and Novell distributions can interact with each other and the same network systems. No one is demanding that GM and Ford merge development efforts yet many want Red Hat and Novell (equally as distinctly separate as GM and Ford) to merge development. The extension of this is the usual generalizations like "Linux sucks" because the speaker looked at one badly done distribution and equated it to anything that happens to use a Linux kernel. I may have been unclear or trumped on other similar discussions. The comparison here has never been about what car represents what OS. That is a far more subjective discussion like two gearheads (car hackers) arguing over GM vs Ford. For one's purposes, Windows could be a Porche 911 provided they realize that Porche has other car models and that cars are also manufactured by companies which compete with Porche and each other. (I'm overly wordy in this post but I think the intentional confusion of different distributions combined with the "too many distros" demand that all production goals be unified benefits only detractors interested in an agenda rather than productive discussion and understanding.)

Slayer_
Slayer_

You lost me... I dont recall criticizing Nix in general in my provious post. I did say that the analogy falls apart for me because I prefer Windows but I hate the Lumina. Windows has better fit and finish but a crappy engine, Lumina has crappy engine and crappy fit and finish. If the car is fun to drive, I could put up with a crappy engine (Isn't that why people buy Arctic Cat snowmobiles, which arguably have the worst engines in snowmobiles today?) And for a bit of bias, I never actually see problems in the Windows kernel that I do not also get in the Linux kernel. They both slow down from uber multi tasking, they both seem to remain slow after extended use, requiring either a logout and log back in, or in the Nix case, often a reboot. (I guess to refresh the daemons *shrugs*) Also, the game engine diversity is dying off. Thankfully, the robust engines are remaining. the Need for Speed series the only series I can think of that still uses a garbage engine. (Not sure what the engine is called, but its performance is terrible)

Neon Samurai
Neon Samurai

It was a while back that you pointed out something I should have thought of on my own. Your lumping all separate products based on the Linux kernel together is like my lumping all games together. By stating that video games are the flakest category of software in general, I ignore the difference between different game development companies. In the same way, by lumping all linux based distros together one ignores the very real differences between Debian, Mandriva and Suse. The experience provided by each of those separate distributions is different. It's no more reasonable to demand that all distro maintainers work together than it is to demand that all video game development companies merge and provide the market with a single title. Why is it more rational to demand two separate platform products based on common commodity parts merge but clearly not rational to demand that Atari, Bungee et all merge and produce one single Need for Speed product?

Slayer_
Slayer_

Better than I could. All cars do not use the same engine. All cars of a brand may use the same engine, but not across brands. Which is once again why my Lumina piston slaps like there is no tomorrow while my Eclipse produces huge power and torque with 2 less cylinders and twice the gas milage while producing a beautiful sound and running perfectly smooth. I changed brands to change engines, no minor tweaking of that POS Lumina's engine will ever fix it, the 3.1's are just utter crap and probably the worst engines they ever produced. If you believe Nix Zealots, the Eclipse should be Linux and the Lumina should be Windows. This is when it really breaks down and the analogy falls apart for me, because I prefer the fit and finish in Windows over the various "Desktop" nix distros. That Lumina is anything but fit and finish, the hood is a full inch higher on the right side than it is on the left.... Just as an example.

darpoke
darpoke

than its engine. Otherwise you'd just buy the engine, right? You seem to be advocating that all cars have their own, unique engine. This makes little sense - are you aware that manufacturers use a common engine to produce a product line of differing price points, vehicle sizes and capacities, each offering further optional extras? The point here is that once an engine has been developed that is deemed satisfactory, it is rolled out for production in all the models where it may be used. Incremental updates may be rolled out over time during production but the engine class remains the same as it does the job. When this ceases to be the case an overhaul is performed and a new model of engine is produced. There are plenty of cases where the same component is used across different vehicles, even of different manufacturers. The Renault Trafic is identical to the Vauxhall Vivaro or Nissan Primastar, as it is the result of a joint venture. The only difference there is the individual maker's badge on the front and rear of the vehicle. Have three massive manufacturers missed something you didn't? The myriad variants of Linux all use a largely common kernel, each extended in different ways to tailor it to a different end user. The strength of the Unix-based core lends stability and extensibility to the entire line of products, while the individual distributions leverage this to provide features for corporate use - such as security and interoperability, or personal use - such as media capability and customisability*. Thus moving from one distro to another with 'the same engine' can yield an extremely novel experience. So much so that users typically have one or more preferred distros - and at least one that they can't stand. Perhaps it's escaped your awareness that various Windows and OS X variants are developed on running generations of kernels? Windows, I believe, is currently on the NT build. OS X on the other hand is based on the mach kernel, and has been a hybrid edging toward full 64-bit capability over the past few iterations. Both these lines, therefore, can be said to run on the same 'engine'. I'm afraid I fail to see where the analogy falls down, at this level (all analogies fail when pushed to their extremity, this is a failure of inductive reasoning, fodder for another discussion altogether). EDIT: typo. *Sorry for all the '-ity' words, it sounds like a badly written product description, but it's hard to describe what I mean otherwise.

Neon Samurai
Neon Samurai

I think your point actually helps the analogy. The car is the product not the engine. The distro is the product not the kernel. Those who would swap engines in there car are as related to the average consumer as those who don't just use the distro provided default kernel. Actually, if we equate your different engines in different cars with different kernels in different distributions it helps it along. The distro is the product and each distribution is a different assembly based on similar commodity parts. Distros can included modified kernels even. This realy highlights the absurdity of equating "Linux" to all distributions along with blaming all distributions for flaws found in specific ones. We don't claim that all cars ar crap because the Ford Pinto doesn't run like a porche.

Slayer_
Slayer_

If my Eclipse used the same POS engine my Lumina used, I wouldn't have bought a new car... Your Car Analogy = Pwned

Neon Samurai
Neon Samurai

Granted, you won't usually see the majority of services running under there own separate user account. But the framework for that is there between current use of "system" account and the winNT kernel's support of true multi-user environments (if userland would only implement it fully without a "value-add" markup). As it is now, I generally have a root terminal open and one or more root GUI apps running regardless of platform. On Windows it's Cain and Netview with an admin cli waiting for whatever else needs privileged. On Debian it's much more so as I'm often using nmap and similar apps that like admin privileged. It's one of the habits from *nix I was able to port back to Windows. True default concurrent users by default in Windows would be nice. I have a few workstations I can't do quick tasks on until late when the user is gone. With true mutli-user, I could simply ssh or remote desktop in as a second active user and do my thing without interrupting the staffer. The alternatives now are cryptic scripting through AD if I wanted it unattended, having the user logoff (no fast switching with AD).

Neon Samurai
Neon Samurai

A: because the car (distro) wrapped around an engine (kernel) is the product rather than the engine (kernel). I can't speak for others but I try to remain consistent in referring to distributions (branding) rather than the commodity part they all happen to be wrapped around. Blindly refering to the entire family when your not discussing a flaw in the kernel which may effect the entire family only leads to confusion. It's better to refer to distributions and provide that clarity about what one is discussing or clarity for people who don't realize that "Linux" based platforms are not all one product from one company. I'm also not going to demand that two separate competing companies suddenly merge product lines; Red Hat and Suse are two separate products regardless of if they both use similar ingredients. It's about as realistic as demanding that Red Hat and Microsoft merge product lines because they both happen to produce an OS.

Slayer_
Slayer_

Truly, the only difference between Home Windows and Home Nix, is that Home Nix CAN login several people at the same time through terminal server or remote desktop, Windows will not allow it unless you have terminal services installed and licensed. Otherwise, the two systems work identical, Windows just requires a Server version and a license. In case you don't believe me, look to virus scanners which use the "System" user so that, even a limited user can still do things like virus scan or update definitions. This can also work via scheduled tasks while there is no one logged into the machine.

darpoke
darpoke

A quibble about your comment regarding the redundancy of a multi-user environment. Even a Unix-based machine with just a single human user has multiple users configured - many services and daemons run as their own user. It's a way of distributing permissions and access, as well as managing the threads that are currently running. Example: we have a new VPS I am currently migrating our mailserver onto here at work (the old VPS is ancient). So far there are only 2 proper user accounts aside from root (the criterion being that they have a directory in /home). One administrator account, set up by default, and my personal account, which I've used to test mail sending and access as a normal, relatively unprivileged user. Care to venture a guess to how many users are listed in /etc/passwd, on this machine which so far has a single "human user" as you would put it? cat /etc/passwd | wc -l returns 29. ftp, sshd, www, mysql... This is what the Unix environment was designed to support. It means you can restrict the access of one process to another processes data, or kill the threads of a stalled process without resorting to halting the system. Kind of like how Google Chrome runs all tabs as their own, sandboxed, process. It's a fantastic way of distributing computation. Far from being a thing of the past, it's one of the core foundations of the stability and scalability of all Unixen.

ScarF
ScarF

Generally, the Linux advocates talk about "Linux". When it comes to the details, Linux becomes - suddenly - what it really is: the kernel. Decide yourselves. When someone say "Linux for desktops", is generic. It actually means "some distro with Linux kernel". Nothing bad with this. Bad is the lack of consistency and common strategy for all these distros. So. All my admiration for the kernel developers. It is superb and I use it on servers whenever possible. But, when I start looking at the distros for the desktop, everything becomes foggy. So, get me right. I am a Linux guy as well as a Windows one. Just, not for desktops. Too much of a headache for me and my users. Too many unknown variables for the near future. And, this multi-user idea. I don't have a single example from everywhere I worked when more than one user needs to work on the same machine, same time. Jeez. The time of Big Blue and damb terminals passed long ago. Here we are talking about Personal Computers.

Neon Samurai
Neon Samurai

If you want a declared mission statement, you have to look at the distro not the kernel that it happens to be wrapped around. If we're talking retail goals only then that reduces the selection down to retail distributions. What is Novell's mission statement or goals for Suse? What is Connonical's mission statement? If we must stick with just the one item. I'd suggest that "Linux" mission statement is to develop a unix like OS kernel supporting a true multi-user posix like envoronment and the widest hardware cross section possible. Kernel devs seem to get doing OK towards that end. Are we talking about the mission statement of the car engine or the entire model of vehicle wrapped around the engine?

CharlieSpencer
CharlieSpencer

Let's try to stick to the 'real world', shall we? :D

Neon Samurai
Neon Samurai

If I replace the preinstalled Linux with a Windows self install then I'm probably chewing my service contract also. This is more about mod'ing the product than any specific OS choice. I don't much expect Cisco to support my linksys router questions after I toss ddWRT onto it but if I buy a ddWRT preinstalled router and replace it's firmware, I'm equally out in the cold. I'm surely not going to suggest that more support and recognition by hardware vendors would be a bad thing though. It could be better regardless of platform. (DLink.. where's the updated drivers for my BT dongle?) Your quibble does put a new perspective on the context of support though.

CharlieSpencer
CharlieSpencer

Remember, these results are just starting points. If you're going to be using these systems in an Active Directory environment, my understanding is that SuSe is the easiest to integrate. Kubuntu vs Ubuntu - my understanding is the only major difference between these two is the desktop. Kubuntu defaults to KDE (hence the 'K'); Ubuntu comes with Gnome. If having a paid external support system is a requirement, Suse or Ubuntu are your best bets on this list. Did that save you some time? :-)

ScarF
ScarF

You know? I worked for IBM a couple of years. Good for the resume. Indedd, they have Lotus Notes for Multiplatform, now. I haven't tried it on Linux, so far. I will. Anyway, IBM have pushed Linux for,... what? Ten years, now? Still couldn't do anything against Windows. Not on desktops. As they couldn't fight back with the superb OS/2. Compared to the OS/2, anyone may see that Linux is the result a part-time job. And, after more than 10 years of struggle, Linux is still at 1% of the market share. It is every day harder to convince that it will succeed. Maybe, Apple with its Mac OS. But, this is like getting rid of the devil and finding his father. Again, I speak from the customer's position. I buy, implement, configure and support IT stuff for the company which pays me to. And I am very devoted to my company and my users. Believe me. If they would need Linux or benefit from it, I wouldn't wait another minute before planing a massive migration. Hm. They don't. In the mean time, I have 4 server out of 7 with Linux. All, application servers.

ScarF
ScarF

For 32 bits: 100% Kubuntu 100% Linux Mint 100% Ubuntu 100% OpenSuSE 100% Mandriva 90% Fedora 90% MEPIS 90% Xandros For 64 bits: 100% OpenSuSe 100% Mandriva 100% Kubuntu 100% Linux Mint 100% Ubuntu 90% Fedora Now. Because we have both processors in the company, after eliminating the 90% results, I intersected the list, and sorted by the score for both categories: 100% Kubuntu 100% OpenSuSe 100% Linux Mint 100% Mandriva 100% Ubuntu So! How to choose? Should I peek OpenSuSe because it's Novell? Should I peek Kubuntu simply because it's in the first position? But Kubuntu is Ubuntu with KDE. Than, shell I peek Ubuntu because it's Debian, and I have Debian on servers? Or Linux Mint which is based on Ubuntu? Or another one not on the list? Eh? Than, comes the GDE. KDE or Gnome? I dislike Gnome, but it is my understanding that KDE hiccups. Same time, there are rumours that KDE will have a come back. Hm. Rumours. Than, the management of installed applications. Here I go in total dark. Virtualizing. Other options to choose from. Need to find more and test all the 5 distros in different configurations, plus the needed applications. This may take one year, also because I have to squeeze it among my other tasks. But, in one year each of the considered distros will suffer at least two new releases. Some will die. Others new will come. Some will split. Others will be rebranded. Start it over? Great. I have something to do until my retirement.

Neon Samurai
Neon Samurai

I've been thinking of something to add but the distro chooser that Palmetto provided the link too asks all the questions I'd ask of you. For what small amount I can add though: If your familiar with Debian already, why not look at that as a base for your standard workstation image? My first task when issued a new notebook was cutting a second partition for Debian. I can't imagine not admining the *nix servers from a *nix workstation; several putty windows on the screen just doesn't compare to a native SSH environment. If your honestly asking for a recomendation for user desktops then that too could be a very interesting conversation. If not starting with the distro chooser, we would have to start with what hardware you have now standard, what user functions are required and what legacy server systems have to be accessed. I also thought Lotus Notes had a *nix client build. Isn't that what IBM likes to recommend instead of Exchange/Outlook? I'll have to read more about Domino to be sure.

Editor's Picks