What Linux developers need to focus on is Windows interoperability. If we can get this OS to work reliably with Microsoft code, Linux will stand a greater chance of survival, and if Linux survives, it means that more and more money will filter down the ranks. That's what it's all about, isn't it? Money. Green. Bread. Pesos. Compensation. We all want to get paid for the blood, sweat, and tears we put into our projects, and once Linux hits the mainstream, the demand will be there to support it financially.
I understand that this is a difficult task to undertake, since Microsoft has a titanium chastity belt around their proprietary code, but I know it can be done. I firmly believe in the Trojan horse theory here; Linux is not the horse itself, but the warriors inside of it. The proverbial horse in this scenario is actually the look, feel, and interoperability of the OS. Windows admins and end-users (is there really any difference?), by nature, are timid creatures that do not like to venture outside the walls of their cozy, fuzzy, cute little operating system. With our Linux warriors waiting inside of the dolled up friendly horse, they stand a better chance of coaxing the Windows folks out to play.
Let's look at a couple of key features and issues that need to be addressed:
As of this moment there is only one client available that even comes close: Ximian Evolution. The developers at Ximian know what makes the corporate world tick: email. Not only is the email functionality there, but the calendar, tasks, contacts, and various other facets of the program function, feel, and look like Outlook. The only thing lacking is the ability to natively import and/or make use of PST and PAB files directly from Outlook. Once this feature is in place, Evolution can also be taken seriously. Don't get me wrong; I use Evolution myself, but I also know what it will take to turn heads and change minds in the corporate world.
That, folks, is everything in a nutshell. There are other concerns that I have, but none as great as the ones I have mentioned in this article. I cannot stress the importance of Windows compatibility enough. It is job Number 1 as far as I am concerned, and you should make it top priority. If we are to conquer the Windows world, there is no other way.
Speaking in terms of money once again, your future depends on it. As time goes by, you will start seeing Windows administrators being replaced by Linux sysadmins around the world, and the only way to survive in this field is to stay on top of the game.
To keep it etched in all of your minds, remember these three simple words:
Note: Comments are owned by the poster. We are not responsible for their content.
Those guys seem to be taking the more effective tactic. . .
Openoffice and Mozilla are meant to be friendly to multi-user installations across platforms. To install openoffice on 50 workstations is no more difficult than doing so for windows apps: hack the scripts burn your own ISO or write an install routine to load OO.org onto the boxes off an NFS or FTP server. Not hard to do and you users will have "preinstalled" OO.org without having to click on som many things. In fact OO.org shows you how to do this<nobr> <wbr></nobr>...
The problem <A HREF="http://www.openoffice.org/">OpenOffice</a openoffice.org> has under Windows is related to the fact that Windows is not a true multiuser operating system.
OpenOffice on a Linux system installs and works very simply for a multi-user configuration (it's even simpler when a distro does the 'hard' part for you, and prepackages it as an RPM).
Just remember to use the <tt>-net</tt> option when installing OpenOffice and everything should work out nicely (why this isn't done through the Setup GUI I have no idea, but there you go).
Finally, the network-based install is only intended to conserve drive space; if you don't care about that, you can cheerfully install the entire app for each individual user. Hopefully, a later version of OOo will become more amenable to the Windows way of handling multi-user configs.
Also, I think that OpenOffice will automatically launch the setup process (for the Workstation install) if it hasn't already been done for a given user, when they attempt to use one of the installed applications (on the other hand, this may just be <A HREF="http://www.mandrakesoft.com/">Mandrake</a mandrakesoft.com> being nice and configuring a bunch of 'preinstall' scripts).
Actually, what's to stop somebody doing the same for a Windows install? Write a bunch of 'front-end' scripts that first check to see if the app has been installed, and then launch Setup automatically if necessary; otherwise, just launch the application as you would normally? This would at least keep the user from having to remember to launch the Setup explicitly. Of course, remember to have the script automatically launch the originally selected app once the Setup is complete<nobr> <wbr></nobr>... something like this could be done using VBScript fairly easily.
The OpenOffice Installation Guide has info on there somewhere describing how to determine whether the software's already installed (under Windows, you just check for a particular INI file; that file also describes which version is currently installed, so your script can determine if the Setup needs to be updated). You can check the Preparing the Installation section on page 11, I think (at least that's in the Win32 version for 1.0.1). You'll probably have to check the INI file itself for specific info, although it looks like it tracks all versions that have been installed<nobr> <wbr></nobr>...
Info on how each application is invoked can be found <A HREF="http://www.openoffice.org/FAQs/faq-other.html#8">here</a openoffice.org>. I'd do this myself, but I use it in a strictly single-user mode, so I've never needed it.
Microsoft only adopts standards until such time as they feel they can usurp and privatize those standards. They "extended" Java with proprietary extensions that would only work on Windows, claiming it was simply more efficient to make use of existing native functionality; therefore, you ended up with Java applications that would only work on Windows (kinda blows away this whole "write one, run anywhere" idea, doesn't it?).
Kerberos was "extended" with a custom capability which, while not actually prohibited according to the spec, meant that non-Windows systems using a fully compliant implementation, couldn't authenticate to a Windows server. On the other hand, a Windows client could freely authenticate against these same non-Windows servers. Granted, they fixed this, but the fact that it was even tried says something, I think.
And there are many more examples, but I don't want to ramble.
Sorry, but Microsoft has repeatedly demonstrated their willingness to follow standards only so long as it is on their terms. If they honestly felt they could preempt a core standard like <tt>TCP/IP</tt> I, for one, do not believe they would hesitate for even a second to do so.
Incidentally, current versions of <A HREF="http://www.linux-mandrake.com/">Mandrake</a linux-mandrake.com> are even easier (if that's possible).
Granted, in a general sense, for the average user (whatever that is), Windows is a more familiar (and therefore "easier") environment to cope with. Linux, however, is rapidly advancing to the point where I wouldn't have any serious qualms introducing a computer newbie to it.
In point of fact, the only devices I've got right now that Mandrake didn't autodetect and configure (without any interaction from me) are the following:
Not bad, considering my hardware list:
And, of course, all the micellaneous stuff on the motherboard and other such items (I've also left out some of the more unusual hardware I've got, which works fine, as far as I can tell).
And no, you don't have to install your own kernels, esp. if you use a distro like Mandrake (which releases new kernel packages almost as fast as you could download and compile 'em anyway). It's nice, though, that we have the option.
BTW, does anyone know if Windows 2000/XP is optimized for a Pentium at least, or is the code still optimized to run on a 486?
6) New motherboard.
Linux: Handled the situation as though there was no upgrade. HD preformance rates went way up.
Windows: Broke the sound driver and had to be reinstalled.
As far as I recall, Photoshop and Illustrator started out on the Mac.
Reading this as a windows user who is interested in linux, i am apalled. The article itself doesn't bother me, it is at least sensible. Make linux like windows, to better 'convert' windows users.
It seems to me that Linux users are 'out to get' Windows users. This bothers me. A community that is this hostile to a pool of people it hopes to assimilate will never come close to touching the kind of numbers Windows draws.
Network: Linux handles a Windows-based network very nicely, through the available GUIs. In <A HREF="http://www.gnome.org/">GNOME</a gnome.org>, try browsing a resource such as <tt>smb:</tt> and see what you get; it's a little slow, but it works. And I know that <A HREF="http://www.kde.org/">KDE</a kde.org> has similar functionality. All somebody has to do is set up appropriate links on the desktop and the average user is set. Of course, this assumes that all dependant packages are installed, but that's another issue (and before somebody suggests "... just install everything<nobr> <wbr></nobr>...", look at how well that really works for Windows. It may be convenient, but it's also very annoying).
Office: Word has become the dominant format through ubiquity and strong-arm tactics. Supporting it only encourages its continued dominance, and puts us in the position of constantly catching up to the latest revision. Sadly, I don't see any real solution to this, except switching to an alternative (of course, this is easier said than done).
E-mail: Linux does, and always has, support e-mail far more effectively than Windows, and probably always will.
Desktop: I don't even want to go there.
Standards: Since when did Microsoft ever care about any standards but their own?
Linux needs to succeed on its own merits. Imitating Windows is a step along the way, not the entire journey. If we follow what you're suggesting here, Linux will never become the dominant force on the desktop simply because people will say "What't the point? I might as well use the original, rather than the imitator."
Education, rather than imitation, is the key. People need to learn that Linux is just as capable, if not more so, than Windows could ever be. Granted, this may involve learning a little, but that's really not that difficult.
And if that's still too much for people to swallow, how difficult would it be to follow the model of another <A HREF="http://www.amiga.com/">operating system</a amiga.com> and simply design acceptable defaults into the environment, while still allowing the flexibility enjoyed by advanced users?
It seems to work for <A HREF="http://www.apple.com/">OS X</a apple.com><nobr> <wbr></nobr>...
bleh
Posted by: Anonymous Coward on September 14, 2002 08:40 PMLinux has its place windows has it's place solaris has it's place OS X has it's place. It's up to the engineers/admins/managers to decide what solution is both COST EFFECTIVE and EASIER to maintain. Cost effective not only includes licenses but training employees to use the software, having available software for the particular os/platform etc. Stop writing articles that either force a choice on someone or whines that linux wasn't the answer. If you like linux then use it<nobr> <wbr></nobr>... stop being like those annoying jeohvas witnesses that run around trying to "save" people.
end rand
#