I don't know about DVDs but old CD technology only lasts a couple of years without data degradation.
So if you want to be sure that your backups of your applications and data on CD really works you have to make new copies now and then.
I would guess that the same goes for DVDs. So unless you can't get a replacement media of your movies, at media price, there really is a need for DVD copying programs.
Actually even if this was possible, there is still a need for copying. What if the copyright holder of your movie goes out of business. Who is then to supply you with replacement media?
But what is not so great is that there is no simple way of verifying that the patched code available for download in fact is the patched code.
We really need some kind of digital signature system so that we can ensure that we are not downloading fixes from a hijacked ftp server.
The best thing to do would of course to validate the code yourself. But many people doesn't have that kind of skills, so they would probably go for just trusting the auther, but without a digital signature how dow we know that the code we download is the real thing.
Your point would have bin valid 10 years ago, but now people have got used to what a GUI should look like and there is not much we can do to change it. If new elements are introduced to the computing environment (e.g. use of database instead of files ) then we can make brave new designs for those parts, but other than that, to much novelty will only alienate the users.
You should also know that Xerox Parc, Apple and finally MS have done a lot or usability research in this field, why throw all that out the window.
I certainly wouldn't mind a simpler standardized way to install apps. But in most distros it is just a double click to do it. I really can't imagine how we can simplify it any more. Unlike windows you don't even have to answer questions like where you want to install it. It just adds the functioality clean and simple. There is a problem with library dependences, but you have that problem in windows too, and people do manage to install programs in that environment. In fact some Linux install systems even address this by automagiclly downloading the missing libs.
I think that the install problem has more to do with all those compressed tar files floating around on the net. Users simply don't realize that they are supposed to be used by developers and not end users. And after trying some of those they tell all their friends how difficult it was.
So I would say that most of the problem is in user perception, even though a standardized GUI to pop up when a user doubleclicks on that install package would not hurt. But the main problem is education.
Or perhaps the code just isn't useable. E.g. It may lack documentation, or it may need build tools or frameworks that is not publically available.
Just look how long it took before something useful came out of the Mozilla project. And I even think they started more or less from scratch after giving up on the original code.
Perhaps there are better sources elsewhere. E.g. if you were to develop a TCP/IP stack would you look at the code from MS or the original BSD code.
And last how do we no that it hasn't leaked. As you really can't do anything meaningful with the leaked code, because MS would sue you if they became aware of it. And that only takes one unsatisfied employee.
The only thing it could be used for is breaking the system. And if you are into that business, you lose your technological advantage if you spread the code further once you have got it. And emploees in the information collection business usually have very strong non disclosure agrements with the governments they work for.
One thing is for sure, there is a lot of windows compatible viruses out there.
To the user, the uptime of the OS is rather irrellevant. The user want to access the serveice and it is the uptime of the service that counts.
Assume that you have a linux box that provide the service with a probability of 0.99 at any given time you could replace that with two Windows boxes that each independently provide the service at a probability of only 0.9 and you would still have 0.99 probability of successful access to your service.
Of course you could do this trick with Linux too, and get 0.9999 probability to acces your servic And you wouldn't have had to pay for two windows licenses to do it.
What's the most cost effective depends on your costs for downtime and your sysadmin. So if you run windows, perhaps you are your salery is to low.
Actually this time they said it was more expensive on the server side. This could be true or not, it all depends. If you are a small company and still have some windows boxes and just use Linux as a complement for some specific service it might be true.
Linux is typically good at handle many users, and a Linux admin can handle thousands of them. But what if you only have 100 people in your company, and still have to use windows for something on the server side. This could result in that you had to hire one extra sysadmin for Linux that was idle most of the time while you still had to have a couple of full time windows admins.
And to be fair modern windows version can keep alive for several months provided you only run one service/box. And if you have tvow boxes for each service you provide you can do schedueled reboots regularly like MS recommends and get very high uptime figures
This gives you slightly higher hardware costs and licensing costs but compared to hiring one more person it is still cheaper.
It is another matter if you apply Linux thin client systems for your desktops then you could expect large savings with Linux. This is waht the city of Largo did and cut their costs in half.
So what system that gives you the lowest TCO is something only you can calculate,it depends on your specific needs.
No, I would not agree that windows always is the obvious chois for the desktop.
In many companies the employees just run a few applications. If you set up thin clients with clean desktop with just one menu contining the approved applications, and all the apps the user needs to do his work starts up at different virtual desktops then the training costs will be very low, as the OS training is reduced to learning to handle one normally very short menu. And application training shold differ very little between the two systems. E.g. a secretary would probaly not need more training when switching from MS-Office to OpenOffice than she would if she upgraded to the next MS-Office version.
You also have the benefit that your emploees will be working instead of swapping screensavers adding pornographic background screens that embarras visiting customers, etc.
The probem is that there so far are less apps available for the Linux desktop. Some of that could be solved by using X11 based windows terminal server clients on the Linux workstations bu not all windows programs works well on terminal server.
But as long as the users only need word processing , spreadsheets, presentation programs, email, webbrowsing, database applications, sofware development tools or accounting programs Linux would be fine.
No it doesn't take many litterate propeler heads, but it will take one or two, to replace four illiterat newbie admins. And the good thing is they only cost you only slightly more/employee.
Your problem is that windows just as hard as Linux to admin. Your newbie admins just don't know it yet.
In fact Windows might be even harder as you must apply schedueled reboot shcemes and failover systems to get the same availability as you get on one Linux box.
So switch to Linux. Fire your newbie windows admins and let them to do windows somewhere else. You will never regret it.
Even if you have to pay slightly more for a Unix expert.
But in many cases this is because he may have more experience, After all each windows version is significantly different from the previous and it only lasts a few years, So you just can't get 20 years of valid Windows experience
At most I guess that the difference could be 10 to 20% more for the Unix guy. But then you don't need four of them to manage your system.
My favourite is net booted department servers that way you could have less skilled personel to change or replace a failing server.
Assuming that the disks are of the hotswapping kind that comes with most IBM xSeries models, you could even have local disks containing data just needed locally, and still use unskilled labor to change a failing server and be up and running in minutes.
Net booting also makes it simpler to do centralized upgrades to software.
OpenOffice do support versioning. It has done that since at least version 1.0. It's one of the more advanced systems I have seen, I haven't used MS-Office for a while so MS may have added functionality that I am not aware of. But I doubt that MS-Office rivals the functionality of Oo in this area.
From the Oo manual: "When several authors are working on the same text or spreadsheet, the review function (marking changes, redlining) makes it possible to document who made which changes, so that this is clear to everyone. On the final edit of the document, it is then possible to look at each individual change and decide whether it should be accepted or rejected."
Hovewer I don't know if versioned documents can be exported/imported to/from MS-Office with versioning intac
It does work, I have tried it. But it is better to refer to security reasons, rather than say that you can't read it.
If they ask what's so insecure, tell them about macro viruses, and that deleted parts of a office document may turn up if you import the document into non MS software.
You say that Linux is less userfriendly than windows. That may be true in some kind of general purpose application, or for the home user.
But this is a bank, they have trained sysadmins that takes care of the things that the users finds difficult. And normally you need less sysadmins to handle a certain number of users on Unix than you do for windows.
The sysadmins can configure user interfiaces that are extremely easy to use. E.g. if the users use six different applications, the sysadmin could have them start on different virtual desktops as the user logs in. And in the rare case that some of those apps would die he could configure a desktop menu with one entry for each program that the user could use to restart the app withour logging out.
Having a simpler work environment like the one I described also reduces the number of support calls.
A environment also makes your emploees focused on their tasks, instead of changeing screen savers, background images, listening to CDs, playing games or other none work related things.
When they say java is faster on Linux they are probably talking about server side applications. And java on an IBM mainfraim is probably a lot faster than on x86 regardless of OS.
As for client side java I'm not sure. The performance have increased enormously on Linux lately, but I'm not sure what have happened on the windows side.
If Microsoft tries to limit the use of windows in non MS software name then you migh here the following dialog in your local software store:
Customer: - What platform does program XYZ run on? Salesperson: -It runs on MacOS-X, Linux and other commnyly know systems Customer: - Does it run on windows? Salesperson: - I really can't say
I agree with your "Don't buy it" part. but you shouldn't download or gain access to it in some other illegal way. If you do you steal the music. By doing so you prove to Sony that they need copy protection.
Instead go down to your local CD store. Find out if they have any copy proteced disks. If they do talk to the customers in the store and people trying to enter the store. Convince them to buy elsewhere. This will mean that the dealer will lose sales and will hopefully avoid ordering copy protected disks.
Talk to the owner and warn him that he could be sued if he doesn't put out warning sings that informs his customer that the productt he is selling as CDs doesn't conform to the CD standard. And sombody might complain.
But if you do a things like this remember to be polite and informative or you risk to come out the bad guy.
Re:It's not just laziness...
on
Due Diligence?
·
· Score: 2
It's not a private matter, just as airplane hijackings are not a private matter.
By gaining access to many hosts on the net an attacker may launch distributed attacks against specific targets.
Imagine if sombody took down all top level DNS servers in such an attack. Today internet is part of the information infrastructure government and press use to communicate with the citizens. And if that information channel is at risk, governments will act.
Re:It's not just laziness...
on
Due Diligence?
·
· Score: 1
Most distros have systems like apt-get. The problem is that when a package for those systems are available, the bug has already bin around for a some time. While the situation is much better than in windows. Microsoft usually spend the same time denying a problem the same time Linux distributers do trying to package a fix. Still, it might be too slow.
You really have to go for the tar archives or even cvs changes that the developers of the offending package usually release very quickly, often within hours when a vulnerabily becomes known.
The problem is how can we ensure that the fix really is a fix and not a trojan put out by sombody else. The MD5 check sums that often acompanies the source code for download only verifies that the file you get over the internet actually is the one you intended to download. This isn't enough. We need to make sure that the code really comes from the developer himself. (I assume that we trust him, or we wouldn't have installed the program in the first place)
So to be sure developers need to digitally sign their code. But the signature is of little value if you can't associate it with a specific person or organization with any certainty.
My idea is that RSA certificates for e-mail could be used. Such certificates can be aquired from e.g. Thawte inc free of charge. In the Thawte Freemail system a subject to be certified is required to in person show a photo ID to at least two notaries that verify his identity.
In this system we would have a way to make sure that we get the real thing when we download something. Provided developers keep the passwords to their private keys to themselves.
You don't give us much facts to go on. But if you don't already use some kind of thin client system for your desktops, this is an area where Linux can help you save money. Look at the Linux Terminal Server Project (LTSP) to get a quick start.
Use native Linux apps where you find good ones. E.g. let Open/StarOffice replace MS-Office, let Evolution replace Outlook. Run the rest of your windows apps from a standard windows terminal server. Yes there are clients for Linux that makes this possible. Then phase out those last windows apps as good Linux alternatives becomes available.
Lock down your thin net booted Linux clients tighly, and your support cost will go down dramatically.
Make sure that you set up a very simple user interface on the clients, don't get tempted to make a windows look alike. Users usually don't need to set background images or new screen savers, play audio CDs etc on their office hours. But streamline the desktop to do business tasks and nothing else. E.g if your users need 6 apps for their daily work, make sure that all of them start automagically at login, perhaps on different virtual desktops. Add some menu choises to restart them if they should fail to your desktop menu. This way you get a much more effective work environment than in windows. And as the uses have less things to play around with they can screw up less things. This is why your support costs will be lower.
On the server side there are good alternatives for filesharing, and print service (use CUPS).
The big question is what you do with exchange if you currently use it. There are good replacement for SMTP and IMAP mail services. But when it comes to calendering and address books the choises are somewhat limited at least if you want to use free software. If you are prepared to have a propriatory solution you could use bynari that is a more or less drop in solution for exchange, available on anything from x86 to z390 platform. You could also go for some web based solution, and there is lots of them.
For authentication and user management use LDAP. E.g. OpenLdap. But if yuo still have some Windows competense left in your company e.g. to handle your terminal servers you could integrate your Linux clients into Microsoft Active Directory. There is an AD extension called AD4Unix that makes this possible.
Things like database system could also be moved but what systems you should use very much depend on what you already use. E.g. if you use SQL Server you should probably look at Sybase in the Linux world. Oracle and DB2 runs well on Linux so they should be no problem to move.
To do all this you will have to spend a lot on education of your sysadmins if they currently have no unix experience. But as you probably will be able to fire three out of five sysadmins when you switched it might be a good idea anyway. Correctly set up you could expect each sysadmin to handle thousands of users.
If you replace the application for your desktop users they will need education on those apps too. But fortunately many Linux applications try to mimic their windows equivalents (do OSS developers lack fantasy?) so the costs may actually be quite low. E.g. I would expect that a switch from MS-Office to Star/OpenOffice would be about the same as when upgrading from one version of MS-Office to the other.
You must remember that so far Linux is mostly used in server rooms where stability and maturity is highly valued. I would expect that many people will upgrade from the 2.2 kernel they currnently use, to 2.4 once 2.6/3.0 is released.
That way they don't need to worry that new features that they probably doesn't need anyway pops up and break things. The only minor number upgrades will be security fixes.
From the Linux point of view it would be better to create good games for the standard Linux platform. That way both the Xboxes and the signed MS software can remain on the shelf.
Slashdot Mirror
I don't know about DVDs but old CD technology only lasts a couple of years without data degradation.
So if you want to be sure that your backups of your applications and data on CD really works you have to make new copies now and then.
I would guess that the same goes for DVDs. So unless you can't get a replacement media of your movies, at media price, there really is a need for DVD copying programs.
Actually even if this was possible, there is still a need for copying. What if the copyright holder of your movie goes out of business. Who is then to supply you with replacement media?
Yes, the quick reaction to this was great.
But what is not so great is that there is no simple way of verifying that the patched code available for download in fact is the patched code.
We really need some kind of digital signature system so that we can ensure that we are not downloading fixes from a hijacked ftp server.
The best thing to do would of course to validate the code yourself. But many people doesn't have that kind of skills, so they would probably go for just trusting the auther, but without a digital signature how dow we know that the code we download is the real thing.
Your point would have bin valid 10 years ago, but now people have got used to what a GUI should look like and there is not much we can do to change it.
If new elements are introduced to the computing environment (e.g. use of database instead of files ) then we can make brave new designs for those parts, but other than that, to much novelty will only alienate the users.
You should also know that Xerox Parc, Apple and finally MS have done a lot or usability research in this field, why throw all that out the window.
I certainly wouldn't mind a simpler standardized way to install apps. But in most distros it is just a double click to do it. I really can't imagine how we can simplify it any more. Unlike windows you don't even have to answer questions like where you want to install it. It just adds the functioality clean and simple. There is a problem with library dependences, but you have that problem in windows too, and people do manage to install programs in that environment. In fact some Linux install systems even address this by automagiclly downloading the missing libs.
I think that the install problem has more to do with all those compressed tar files floating around on the net. Users simply don't realize that they are supposed to be used by developers and not end users. And after trying some of those they tell all their friends how difficult it was.
So I would say that most of the problem is in user perception, even though a standardized GUI to pop up when a user doubleclicks on that install package would not hurt. But the main problem is education.
Standardization doesn't need to mean taking away configurability. It just means that we have a common starting point to begin with.
Or perhaps the code just isn't useable. E.g. It may lack documentation, or it may need build tools or frameworks that is not publically available.
Just look how long it took before something useful came out of the Mozilla project. And I even think they started more or less from scratch after giving up on the original code.
Perhaps there are better sources elsewhere. E.g. if you were to develop a TCP/IP stack would you look at the code from MS or the original BSD code.
And last how do we no that it hasn't leaked. As you really can't do anything meaningful with the leaked code, because MS would sue you if they became aware of it. And that only takes one unsatisfied employee.
The only thing it could be used for is breaking the system. And if you are into that business, you lose your technological advantage if you spread the code further once you have got it. And emploees in the information collection business usually have very strong non disclosure agrements with the governments they work for.
One thing is for sure, there is a lot of windows compatible viruses out there.
To the user, the uptime of the OS is rather irrellevant. The user want to access the serveice and it is the uptime of the service that counts.
Assume that you have a linux box that provide the service with a probability of 0.99 at any given time you could replace that with two Windows boxes that each independently provide the service at a probability of only 0.9 and you would still have 0.99 probability of successful access to your service.
Of course you could do this trick with Linux too, and get 0.9999 probability to acces your servic And you wouldn't have had to pay for two windows licenses to do it.
What's the most cost effective depends on your costs for downtime and your sysadmin. So if you run windows, perhaps you are your salery is to low.
Actually this time they said it was more expensive on the server side. This could be true or not, it all depends. If you are a small company and still have some windows boxes and just use Linux as a complement for some specific service it might be true.
Linux is typically good at handle many users, and a Linux admin can handle thousands of them. But what if you only have 100 people in your company, and still have to use windows for something on the server side. This could result in that you had to hire one extra sysadmin for Linux that was idle most of the time while you still had to have a couple of full time windows admins.
And to be fair modern windows version can keep alive for several months provided you only run one service/box. And if you have tvow boxes for each service you provide you can do schedueled reboots regularly like MS recommends and get very high uptime figures
This gives you slightly higher hardware costs and licensing costs but compared to hiring one more person it is still cheaper.
It is another matter if you apply Linux thin client systems for your desktops then you could expect large savings with Linux. This is waht the city of Largo did and cut their costs in half.
So what system that gives you the lowest TCO is something only you can calculate,it depends on your specific needs.
No, I would not agree that windows always is the obvious chois for the desktop.
In many companies the employees just run a few applications. If you set up thin clients with clean desktop with just one menu contining the approved applications, and all the apps the user needs to do his work starts up at different virtual desktops then the training costs will be very low, as the OS training is reduced to learning to handle one normally very short menu.
And application training shold differ very little between the two systems. E.g. a secretary would probaly not need more training when switching from MS-Office to OpenOffice than she would if she upgraded to the next MS-Office version.
You also have the benefit that your emploees will be working instead of swapping screensavers adding pornographic background screens that embarras visiting customers, etc.
The probem is that there so far are less apps available for the Linux desktop. Some of that could be solved by using X11 based windows terminal server clients on the Linux workstations bu not all windows programs works well on terminal server.
But as long as the users only need word processing , spreadsheets, presentation programs, email, webbrowsing, database applications, sofware development tools or accounting programs Linux would be fine.
No it doesn't take many litterate propeler heads,
but it will take one or two, to replace four illiterat newbie admins. And the good thing is they only cost you only slightly more/employee.
Your problem is that windows just as hard as Linux to admin. Your newbie admins just don't know it yet.
In fact Windows might be even harder as you must apply schedueled reboot shcemes and failover systems to get the same availability as you get on one Linux box.
So switch to Linux. Fire your newbie windows admins and let them to do windows somewhere else. You will never regret it.
Even if you have to pay slightly more for a Unix
expert.
But in many cases this is because he may have more experience, After all each windows version is significantly different from the previous and it only lasts a few years, So you just can't get 20 years of valid Windows experience
At most I guess that the difference could be 10 to 20% more for the Unix guy. But then you don't need four of them to manage your system.
My favourite is net booted department servers that way you could have less skilled personel to change or replace a failing server.
Assuming that the disks are of the hotswapping kind that comes with most IBM xSeries models, you could even have local disks containing data just needed locally, and still use unskilled labor to change a failing server and be up and running in minutes.
Net booting also makes it simpler to do centralized upgrades to software.
OpenOffice do support versioning. It has done that since at least version 1.0. It's one of the more advanced systems I have seen, I haven't used MS-Office for a while so MS may have added functionality that I am not aware of. But I doubt that MS-Office rivals the functionality of Oo in this area.
From the Oo manual:
"When several authors are working on the same text or spreadsheet, the review function (marking changes, redlining) makes it possible to document who made which changes, so that this is clear to everyone. On the final edit of the document, it is then possible to look at each individual change and decide whether it should be accepted or rejected."
Hovewer I don't know if versioned documents can be exported/imported to/from MS-Office with versioning intac
It does work, I have tried it. But it is better to refer to security reasons, rather than say that you can't read it.
If they ask what's so insecure, tell them about macro viruses, and that deleted parts of a office document may turn up if you import the document into non MS software.
You say that Linux is less userfriendly than windows. That may be true in some kind of general purpose application, or for the home user.
But this is a bank, they have trained sysadmins that takes care of the things that the users finds difficult. And normally you need less sysadmins to handle a certain number of users on Unix than you do for windows.
The sysadmins can configure user interfiaces that are extremely easy to use. E.g. if the users use six different applications, the sysadmin could have them start on different virtual desktops as the user logs in. And in the rare case that some of those apps would die he could configure a desktop menu with one entry for each program that the user could use to restart the app withour logging out.
Having a simpler work environment like the one I described also reduces the number of support calls.
A environment also makes your emploees focused on their tasks, instead of changeing screen savers, background images, listening to CDs, playing games or other none work related things.
When they say java is faster on Linux they are probably talking about server side applications. And java on an IBM mainfraim is probably a lot faster than on x86 regardless of OS.
As for client side java I'm not sure. The performance have increased enormously on Linux lately, but I'm not sure what have happened on the windows side.
Perhaps Linux can be used for cleaning windows
If Microsoft tries to limit the use of windows in non MS software name then you migh here the following dialog in your local software store:
Customer: - What platform does program XYZ run on?
Salesperson: -It runs on MacOS-X, Linux and other commnyly know systems
Customer: - Does it run on windows?
Salesperson: - I really can't say
I agree with your "Don't buy it" part. but you shouldn't download or gain access to it in some other illegal way. If you do you steal the music. By doing so you prove to Sony that they need copy protection.
Instead go down to your local CD store. Find out if they have any copy proteced disks. If they do talk to the customers in the store and people trying to enter the store. Convince them to buy elsewhere. This will mean that the dealer will lose sales and will hopefully avoid ordering copy protected disks.
Talk to the owner and warn him that he could be sued if he doesn't put out warning sings that informs his customer that the productt he is selling as CDs doesn't conform to the CD standard. And sombody might complain.
But if you do a things like this remember to be polite and informative or you risk to come out the bad guy.
It's not a private matter, just as airplane hijackings are not a private matter.
By gaining access to many hosts on the net an attacker may launch distributed attacks against specific targets.
Imagine if sombody took down all top level DNS servers in such an attack. Today internet is part of the information infrastructure government and press use to communicate with the citizens. And if that information channel is at risk, governments will act.
Most distros have systems like apt-get.
The problem is that when a package for those systems are available, the bug has already bin around for a some time. While the situation is much better than in windows. Microsoft usually spend the same time denying a problem the same time Linux distributers do trying to package a fix. Still, it might be too slow.
You really have to go for the tar archives or even cvs changes that the developers of the offending package usually release very quickly, often within hours when a vulnerabily becomes known.
The problem is how can we ensure that the fix really is a fix and not a trojan put out by sombody else. The MD5 check sums that often acompanies the source code for download only verifies that the file you get over the internet actually is the one you intended to download.
This isn't enough. We need to make sure that the code really comes from the developer himself.
(I assume that we trust him, or we wouldn't have installed the program in the first place)
So to be sure developers need to digitally sign their code. But the signature is of little value if you can't associate it with a specific person or organization with any certainty.
My idea is that RSA certificates for e-mail could be used. Such certificates can be aquired from e.g. Thawte inc free of charge. In the Thawte Freemail system a subject to be certified is required to in person show a photo ID to at least two notaries that verify his identity.
In this system we would have a way to make sure that we get the real thing when we download something. Provided developers keep the passwords to their private keys to themselves.
You don't give us much facts to go on. But if you don't already use some kind of thin client system for your desktops, this is an area where Linux can help you save money. Look at the Linux Terminal Server Project (LTSP) to get a quick start.
Use native Linux apps where you find good ones. E.g. let Open/StarOffice replace MS-Office, let Evolution replace Outlook. Run the rest of your windows apps from a standard windows terminal server. Yes there are clients for Linux that makes this possible. Then phase out those last windows apps as good Linux alternatives becomes available.
Lock down your thin net booted Linux clients tighly, and your support cost will go down dramatically.
Make sure that you set up a very simple user interface on the clients, don't get tempted to make a windows look alike. Users usually don't need to set background images or new screen savers, play audio CDs etc on their office hours. But streamline the desktop to do business tasks and nothing else. E.g if your users need 6 apps for their daily work, make sure that all of them start automagically at login, perhaps on different virtual desktops. Add some menu choises to restart them if they should fail to your desktop menu. This way you get a much more effective work environment than in windows.
And as the uses have less things to play around with they can screw up less things. This is why your support costs will be lower.
On the server side there are good alternatives for filesharing, and print service (use CUPS).
The big question is what you do with exchange if you currently use it. There are good replacement for SMTP and IMAP mail services. But when it comes to calendering and address books the choises are somewhat limited at least if you want to use free software. If you are prepared to have a propriatory solution you could use bynari that is a more or less drop in solution for exchange, available on anything from x86 to z390 platform.
You could also go for some web based solution, and there is lots of them.
For authentication and user management use LDAP.
E.g. OpenLdap. But if yuo still have some Windows competense left in your company e.g. to handle your terminal servers you could integrate your Linux clients into Microsoft Active Directory.
There is an AD extension called AD4Unix that makes this possible.
Things like database system could also be moved
but what systems you should use very much depend on what you already use. E.g. if you use SQL Server you should probably look at Sybase in the Linux world. Oracle and DB2 runs well on Linux so they should be no problem to move.
To do all this you will have to spend a lot on education of your sysadmins if they currently have no unix experience. But as you probably will be able to fire three out of five sysadmins when you switched it might be a good idea anyway.
Correctly set up you could expect each sysadmin to handle thousands of users.
If you replace the application for your desktop users they will need education on those apps too.
But fortunately many Linux applications try to mimic their windows equivalents (do OSS developers lack fantasy?) so the costs may actually be quite low. E.g. I would expect that a switch from MS-Office to Star/OpenOffice would be about the same as when upgrading from one version of MS-Office to the other.
The most important thing is that you are not
over 25 and that you have at least 15 years
of experience in the field of the job you
are applying for.
You must remember that so far Linux is mostly used in server rooms where stability and maturity is highly valued. I would expect that many people will upgrade from the 2.2 kernel they currnently use, to 2.4 once 2.6/3.0 is released.
That way they don't need to worry that new features that they probably doesn't need anyway
pops up and break things. The only minor number upgrades will be security fixes.
From the Linux point of view it would be better to create good games for the standard Linux platform. That way both the Xboxes and the signed MS software can remain on the shelf.