We should work up some statistics on the flames in these things. The number of people who feel like they have to defend their OS, The number of people who must insult the original poster, the number of strawman arguments… it would be interesting to see if these rise and fall over the years.

IMHO all of the OSs offer something unique and nice. I use Windows because I am most familiar with it and I have to use it at work. I use Linux because its free and I like to learn new things. I like both.

Most of the benefit in any OS is in how it implements something rather than what is implemented. There are work arounds to just about everything and eventually something new is going to be ported in every possible direction.

Overall I’d like Linux to win because it is the underdog, and I always root for the underdog.

You obviously haven’t seen powershell manual. Man, UNIX shells don’t even compare to the new powershell console. They had the lead while there was old NT command prompt, but now with PowerShell, IMO they have definitely lost it. PS has far more advanced architecture. Haven’t you ever felt that passing data in text format is wrong way to do it?? Using regexp and other tools to separate data (from text) is wrong too. Machines aren’t working like that. They work best with binary data, and PS works just like that. It passes data through objects. And because it uses .net behind the scenes it has both the power and speed.

Also, from user’s point of view – MS has cleaned naming convention – unlike how it is in UNIX word where commands are shortened to ambiguous abbreviation with no common rules how to compose them and one has to learn all the commands that he needs, Microsoft has fixed that with “verb-object” naming convention (like ‘get-process’ instead ps, (also abbreviations are possible, so if you like you can use ‘ps’ too)). Also there is also lot less ‘mess’ that is common with UNIX commands.

I work at Microsoft, so I might be a bit biased. :) Of couse I speak only for myself below, not my employer.

1. Putty.
2. Absolutely. I don’t want to be the guy to run such a repository, though.
3. It’s all about muscle memory. Ctrl-arrow keys to move a word, shift as a selection modifier, and ctrl-c to copy and ctrl-v to paste are wired into my hands. I’m much less productive–an order of magititude less productive–in applications that fail to honor these shortcuts. It’s easy to see how the X shortcuts could be as ingrained.
4. http://download.microsoft.com/ does tend to have new stuff all the time, and it’s cool new stuff too! :) The really systemic cool new stuff, like what’s in Vista, has taken too long to come out. Hopefully that’s changing.
5. There are many multi-desktop apps currently out there for Windows and with Vista they will work much better (since it has a desktop compositor.) The root issue, why multiple desktops aren’t part of the Windows environment, is a design, history, and usability decision that I do not know the specifics of–but–like the fancy terminal emulator–it strikes me as a 1% feature.
6. Pretty much every indispensible tool for Unix has a Windows port. Vim is my personal must-have.
7. Yay for Vista, all hail Vista. Far far fewer reboots are now needed.
8. Political issue, rather than technical, isn’t it?
9. Paranoia is always warranted. :) Isn’t that the lesson of OpenBSD?
10. Windows goes farther than giving you choices: it not only gives you choices, it gives you choices about how to want to find those choices. Isn’t that what #2 is complaining about? You’re also going to find, I hope, that the experience of software licensing in Vista is far, far more friendly than WGA was in Windows XP.

I used Linux (Gentoo) and FreeBSD back in college and I still use them occasionally. What I miss from Unix is the ease of the command line. Yes, Windows has cmd.exe and batch scripts and javascript and yes everything can be completed in Windows that can be completed in Unix. BUT–there’s a focus in the Unix world on making command-line tools easy to use that doesnt exist on Windows. The principle of least surprise applies to non-graphical as well as graphical user interfaces. rant off Hopefully PowerShell will do something here.

Thanks for posting this even-handed list.

Yours, Philip

I’ve tried using SuSE and Ubuntu and I’ve run into the same brick wall both times, to whit: if it came with the distro and is pre-listed in the package manager – no problem, but if it is downloaded from the net, it isn’t going to happen. I’ve tried compiling from source, RPM’s, self-extracting binaries and not a single one of the half dozen programs is even close to working. I’ve read man pages, info pages, books by O’Reilly, Sybex and Wiley; I’ve religiously followed recipes from the web and had absolutely no joy of any kind.

The problem seems to me to be that installing unapproved software on Windows is designed to be something any idiot user can do (which has pluses and minuses to it, I know) and Unix is designed so that only the most adept and knowledgeable system administrators can install unapproved software on it.

I’ve found deinstallation of packages from Linux repositories to be cleaner than deinstallation via add/remove programs in Windows. For the person on FreeBSD who has manually installed software, read instructions that came with the program. Oftentimes you can ‘make deinstall’ or similar (or never performa a make install & just put the binaries you’ve made in a subdirectory which you can later rm).

Linux does have many commercial apps. Check out many listed: http://tldp.org/HOWTO/Commercial-HOWTO.html This is from 1999 & the number of apps has only increased. Also, commercial support infrastructures are quite common.

Maki’s anti-shell argument is a riot. The thing with Linux is that it offers you a choice–it does have a very powerful GUI, but also has a very powerful CLI. Windows had a relatively poor CLI by default. Text-based input can still be used to manipulate multimedia & you can listen to them from the CLI or can open them in a graphical viewer after having made or modified them in the CLI.

I’ll gladly take Maki’s challenge of graphical editing in CLI vs GUI. Except, I’ll up the odds: take 100 images, add a line to all of them, add a timestamp under that line, and resize them. Maki can you Paintbrush. I’ll use imagemagick.

This is trite, I know–the point is that Maki is making a false choice–noone HAS to use a CLI in Linux. They have the ability to choose to use a CLI when it would be more efficient and more effective.

1. There are decent term emulators available for Windows, they are just not packaged with the distro from MS. I myself use putty and/or the f-secure client and have no issues getting everything done that I need to. Are all of the various clients free? No. but free isn’t always a requirement either.

2. There are ways to automate the build process for Windoes systems as well as Linux systems. I can have a new Windows based server built in under 3 hours to our corporate standards using automated methods, basically booting to a CD or DVD disk or image, supplying IP and hostname information and walking away. I can do this sitting in front of the machine or from 3000 miles away. Is it perfect? No. Does it work 99.9% of the time? You bet. We have a similar process for out standard Linux build that takes about the same time.

3. As stated before, Power toys from MS can handle all of the gripes here.

4. Frequesnt painless patching? Not on any platform! You try patching a couple hundred Linux boxes and you will see your productivity numbers drop significantly. Do you have to load more patches on Windows? Without a doubt. Can I use a product to automatically update those Windows boxes more easily than my Linux boxes? Also, without a doubt. There are more commercial products available to handle patch management for Windows than for Linux. We also have more issues with patches breaking binary compatibility for our Linux and Unix systems than with the Windows systems.

5. Multiple desktops? Power toys and/or additional monitors. I use both.

6. I use vim on Windows. I also use pspad (www.pspad.com) and pfedit, and of course, notepad in a pinch.

7. No reboots? yeah right. Windows systems do require more reboots for patching or updates, but in a large corporate environment we find that flaky code can exist in either environment. It does more damage to the Windows systems, but we do have to reboot the Linux systems at times too.

8. Open formats are open regardless of the platform. If there isn’t a Windows port, there is generally someone working on one. I haven’t had any issues finding what I need here as well.

9. If you look up the history of viruses (http://www.cknow.com/vtutor/HistoryofViruses.html) you will see that the first virus was written for the Apple II, the second written for Unix. Right now, Windows is a good target for script kiddies and the likes… in the future, who knows?

10. No arguments really, I find that advanced users have to work harder no matter what the focus to get the respect you deserve.

Bare in mind, I am not trying to argue with you on your points, mearly expand on them and offer a different or expanded perspective. I find there are pluses and minuses to all platforms and ultimately there is a reason they all still exist. I have yet to find the perfect platform for everything. I appreciate the hard work you put into your article as well.

http://www.hypersensory.com/linuxvswindows

To those who argued that Shell is almighty and powerful enough for anything – try to do something more fun than looking for a string of code in your files.

THE problem with shell is that the returned information can only be viewed as ASCII. So anything more complex than text of some sort, is impossible. Yes, i know i can make image or even watch a movie in ASCII, but who would do it more than once (to just check it out?).

Someone has mentioned image manipulation – i dare you to edit a picture or even draw one in shell in the same way it can be done in the most primitive editor in the world – MS Paint.

You just can’t do enough useful stuff in Shell that most people do in graphical interface. I’m not saying shell is useless, but it is just not as useful as GUI.

GUI is times and times again more powerful than shell, because people can receive and manage information on the screen easier.

1) Powershell closes the gap easily

I’ve seen this argument a few times now. Powershell still doesn’t compare to Gnome-terminal (or kterm), with the available Gnu (and other unixisms), at least for me. Given time, maybe it will grow on me (it took me years before I appreciated vim).

I actually have found add-ons for most of my nits for XP. The problem is that they’re add-ons, not part of the core distro (hence MS is missing out). For improved terminal goodness, for example, I use poderosa with Cygwin. It’s a vast improvement over the default shell, but I have to install it on every XP machine I work on (takes time).

Even with multi-desktops, there are many solutions for XP. The problem is mapping key bindings and features so that I can switch between them. The default Gnome/Metacity switcher is nice (as is KDE’s). The Window’s powertool is nowhere near as well done, and it breaks many applications (so I end up not using it).

Building this stuff into the base is why I personally find my Ubuntu desktop so much fun. Note that I still use XP regularly: it’s still a very reasonable OS.

And by the way, Maki, you sure have to be kidding about something. I mean you cant really be using both systems and say something like the ‘using the mouse is faster than using the terminal’ or that you have not heard of proprietary formats in windows !

1. The terminal IS faster than using a mouse, if you know how to use it. If you can do something with your mouse, you can do it in the terminal, I guarantee it. Except for 3D games, of course.

2. Finding an app in Linux that does what you want is as easy as $ apt-cache search whateveryouwant That will search the repository of roughly 20,000 apps, for whatever you want.

3. To do an update in Windows, you ave to go to their website, install the update manager stuff, validate your Windows crap with a little .exe, and then you can update (and reboot) In a Linux distro using apt, you get all of that preinstalled, and it just says “# updates Available.” You click install, enter your password, and then they’re up and running. Unless you’re recompiling your kernel, you’ll never have to reboot.

4. I use multiple desktops, and I stay organized. I have 4 desktops. One for Firefox, one for text tools (OpenOffice, vim, etc.,) one for personal stuff, like Gaim and Rhythmbox, and then one for development and such, like Python and the terminal.

5. You just don’t like vim because you don’t know how to use it. If you knew how, I guarantee it would become the most useful tool on your system.

6. The thing with Linux and rebooting, is that when you run an app in Linux, it loads the binaries into the RAM. That way, you can modify anything, you can even delete the program, and it will still run until clear the RAM. This is useful, because you can modify programs (updates, patches) and then just reload the program, and it’s working. The only time you need to reboot in Linux is if you recompile the kernel.

7. Now this is just ridiculous. Linux uses OpenSSH, Windows doesn’t. Linux uses open-source network drivers, etc.

8. What? Are you crazy? OpenSSH and ProFTPd are thousands of times more secure than MS’s RDP.

And most importantly, @ Justin Silverton “Something linux is missing is the ability for developers to write commercial applications (I haven’t seen many successful, commercial, linux apps). This is an important driving force for any operating system.” Oh. Apparently Apache, PHP and MySQL (open-source apps) that serve over 70% of the pages you go to aren’t successful and used commercially. Shall I continue? OpenOffice.org, GNUCash, Firefox, MediaWiki, Blender 3D. What?! More?! SourceForge.net, Samba, PostgreSQL, SQLite, and GNU/Linux.