How about when you install package B, and package A, C, and D, ceases to work because package B over-wrote a DLL file with an older version, in the SYSTEM directory.
I have not found any 3rd party application that (by default) installed in/usr/bin,/usr/lib, and in no-way conflict with or writes to the same directory that system files are in.
Another story...
If I had to install a package on 50 windows machines (not even 1000) I would scream. With any *nix flavour, I can install this package from over the network, and not have any fear that the installation will screw up my remote connection to that machine.
Then I can (in the space of an hour or two) write a script that will update each machine without my even having to babysit.
When jabber or any clone client receives the request for the md5sum for bytes A thru B of the aim.exe, we simply proxy this request to another server... which could be the local machine.
Therefore, we simply setup a simple daemon someplace (aimsumd) which given a version number, and two byte locations returns the md5sum.
client A, upon receiving this request from OSCAR, makes a similar request to an aimsumd server, and sends the result back to OSCAR.
Like I said, simple in principal, pain in the a$$ in practice.
If you can show me an independant, complete, fair and honest evaluation of TCO that shows that NT is less expensive in TCO than UNIX I would be happy to read it.
However, A document written by M$ comparing it's product to any other product can only be rated as maketing propoganda. If for example, you can find something by an oil company, or by foreign military, or something that independantly evalutates it, I would llike to see it.
I have found reports where Win2K wins over unix because they decided (correctly) that one machine on everybody's desk would lead to lower TCO over having a dedicated unix machine on everybody's desk in addition to a windows based PC. This same report did not evaluate PC Unix variants, and did not mention why.
I of course saw MS$ report, which is only marketing propoganda.
I saw some replies to this report, which can only be described as a rave in response to a rant.
A complete TCO comparison would include:
Hardware Costs
Actual Software Licensing Costs
Actual Training Costs
Actual Support Costs
Actual Downtime Costs (Ok I know this is hard to measure)
And not just theories or assumptions. Ideally, the comparison would be done in an engineering type shop, where CAD/CAM/CAE is done instead of just paper pushing. The people who use these engineering type products tend to push the resources hareder, and also tend to be more evenly trained in Unix or Windows. If you go to a paper pushing shop (insurance offices for example). Everybody has been pre-trained in the Microsoft world, therefore training costs would be significantly lower for Windows.
Java is positioned for the wide release market... Not the inhouse (corporate) environment. Your thinking visual (ugh) BASIC (Beginners All-Purpose Symbolic Instruction Code) (it says it all).
Though I am not a java developer myself (yet). I've spent 10 years doing C++ code in X/Motif and for MS Windoze using a variety of different cross platform methodologies. I can already see Java maturing as a language and moving into wide use as a general software development tool for all kinds (especially wide-release) of applications.
The difference between Java and C/C++ comes in age. C++ has been around for 10 years or so now, and is well established in the market. Java is just now coming into maturity. Java is at the place where C++ was 10 years ago.
Simple, putting the standard Linux vs Windoze arguments behind.
For embedded systems, ie: Something controlling valves, switches, and reading meters. Typically sitting in a shed somewhere in the middle of nowhere, I don't need a GUI... And I definately shouldn't need a 100Mhz processor just because the GUI OS requires it.
Embedded Linux (Free), QNX (tried and true), or some other RealTimeOS is significantly more reliable, cheaper to maintain and cheaper to install.
I can't tell you how many times my buddies Dreamcast crashed (stopped couting at two dozen one day).
And I definately don't need a GUI to operate my 64MB MP3 Player.
The only problem with pranks like this is that every once in a while, they do go wrong... and they do hurt somebody.
IMHO It comes down to un-necessary risk.
If these pranksters had hurt anybody, then (if caught) the would and should prosecuted to the full extent of the law...
However, since they didn't hurt anybody, and as usually most of the inconvenience was in the fact that the authorities blocked traffic while looking at the thing, the penalty if these people are caught should be a simple slap on the wrist.
The place (IMHO) where you will get the most mileage is in scheduling appointments.
It would be nice If I could schedule an appointment (and cancel it) online, and if the receptionist could (on an opt in basis) get ahold of me via an instant messenger, or SMS page to reschedule or confirm my appointment.
Records online would be efficient for in the office, but I don't want even the chance that somebody can hack in to your network and get my records, so therefore keep your internal network separate from your internet stuff.
The next step though (and this can be fairly secure) is to be able to communicate with my doctor (not some substitute) via email, so that I can ask non life-threatening questions without having to phone and wait on hold. The doctor can then tell me if I should book an appointment or not.
Even email communications should be secure so that they aren't kept in plain text anywhere for some hacker to find.
R&D costs are probably one reason why NOKIA decided on linux... Most of the software is already there (except maybe for a few firewire drivers, etc) or almost there... There R&D costs would be (IMHO) much lower than trying to get the same thing Stable in a windoze box.
Also, because of GPL, there are no SW royalties (they don't have to pay Micro$oft for each box).
Sounds like a smart decision (business wise) to me.
Here is further proof that X sucks. It seems to me that it is frequently the impossible obstacle hindering the capabilities of almost every UN*X OS. I'd (conservatively) wager that its back asswards client/server model is actually useful 5% of the time, if that much.
<i><p>For those that need remote displays, LET THEM USE VNC. Let the rest of us move forward. I despise configuring X to do anything reasonable, and if it doesn't work at all, it's saner to simply break out the installation CD and try the install again than it is to try putting in monitor/video configuration by hand.
</p></i>
<p>I am currently in France sitting at an SGI, and remotely displaying Netscape that is running on a SUN (I eliminate colour flash problem that way, because the SGI has much better colour capabilities than older suns). I am displaying two xterms on remote sgi machines sitting somewhere in this office. I am displaying rxvt from my Mandrake7.1 machine in Calgary Canada that I am logged into via ssh. I can run licq from there, and can administer my mail/web/samba server, and my firewall. That's the power of X</p>
<p>
For games like Quake, Descent and the like, I agree X probably isn't the solution. But the power of Unix/Linux is that you can use the right tool for the job. You aren't constrained to one window manager that has to try to fill all roles, and ends up doing all of them half-assed.</p>
<p>And, just to get nit picky... At least if I shut down my machine and slap in a different video card, I can re-configure my Linux machine without hours of pain because of incompatible drivers. I simply reboot to a non X mode, configure and test X (I <b>might</b> have to install more modules off of my distribution CD). And go back into the default runlevel(5)</p>
<p>If you simply shut down your win9x box, swap a different video card, and reboot... You could be in for a world of hurt</p>
Slashdot Mirror
mmmmmmmmmmmm, bacon
How about when you install package B, and package A, C, and D, ceases to work because package B over-wrote a DLL file with an older version, in the SYSTEM directory.
/usr/bin, /usr/lib, and in no-way conflict with or writes to the same directory that system files are in.
I have not found any 3rd party application that (by default) installed in
Another story...
If I had to install a package on 50 windows machines (not even 1000) I would scream. With any *nix flavour, I can install this package from over the network, and not have any fear that the installation will screw up my remote connection to that machine.
Then I can (in the space of an hour or two) write a script that will update each machine without my even having to babysit.
When jabber or any clone client receives the request for the md5sum for bytes A thru B of the aim.exe, we simply proxy this request to another server... which could be the local machine.
Therefore, we simply setup a simple daemon someplace (aimsumd) which given a version number, and two byte locations returns the md5sum.
client A, upon receiving this request from OSCAR, makes a similar request to an aimsumd server, and sends the result back to OSCAR.
Like I said, simple in principal, pain in the a$$ in practice.
If you can show me an independant, complete, fair and honest evaluation of TCO that shows that NT is less expensive in TCO than UNIX I would be happy to read it.
However, A document written by M$ comparing it's product to any other product can only be rated as maketing propoganda. If for example, you can find something by an oil company, or by foreign military, or something that independantly evalutates it, I would llike to see it.
I have found reports where Win2K wins over unix because they decided (correctly) that one machine on everybody's desk would lead to lower TCO over having a dedicated unix machine on everybody's desk in addition to a windows based PC. This same report did not evaluate PC Unix variants, and did not mention why.
I of course saw MS$ report, which is only marketing propoganda.
I saw some replies to this report, which can only be described as a rave in response to a rant.
A complete TCO comparison would include:
Hardware Costs
Actual Software Licensing Costs
Actual Training Costs
Actual Support Costs
Actual Downtime Costs (Ok I know this is hard to measure)
And not just theories or assumptions. Ideally, the comparison would be done in an engineering type shop, where CAD/CAM/CAE is done instead of just paper pushing. The people who use these engineering type products tend to push the resources hareder, and also tend to be more evenly trained in Unix or Windows. If you go to a paper pushing shop (insurance offices for example). Everybody has been pre-trained in the Microsoft world, therefore training costs would be significantly lower for Windows.
Just my $0.02.
Java is positioned for the wide release market... Not the inhouse (corporate) environment. Your thinking visual (ugh) BASIC (Beginners All-Purpose Symbolic Instruction Code) (it says it all).
Though I am not a java developer myself (yet). I've spent 10 years doing C++ code in X/Motif and for MS Windoze using a variety of different cross platform methodologies. I can already see Java maturing as a language and moving into wide use as a general software development tool for all kinds (especially wide-release) of applications.
The difference between Java and C/C++ comes in age. C++ has been around for 10 years or so now, and is well established in the market. Java is just now coming into maturity. Java is at the place where C++ was 10 years ago.
Simple, putting the standard Linux vs Windoze arguments behind. For embedded systems, ie: Something controlling valves, switches, and reading meters. Typically sitting in a shed somewhere in the middle of nowhere, I don't need a GUI... And I definately shouldn't need a 100Mhz processor just because the GUI OS requires it. Embedded Linux (Free), QNX (tried and true), or some other RealTimeOS is significantly more reliable, cheaper to maintain and cheaper to install. I can't tell you how many times my buddies Dreamcast crashed (stopped couting at two dozen one day). And I definately don't need a GUI to operate my 64MB MP3 Player.
The only problem with pranks like this is that every once in a while, they do go wrong... and they do hurt somebody. IMHO It comes down to un-necessary risk. If these pranksters had hurt anybody, then (if caught) the would and should prosecuted to the full extent of the law... However, since they didn't hurt anybody, and as usually most of the inconvenience was in the fact that the authorities blocked traffic while looking at the thing, the penalty if these people are caught should be a simple slap on the wrist.
Keep it Simple & Secure...
The place (IMHO) where you will get the most mileage is in scheduling appointments.
It would be nice If I could schedule an appointment (and cancel it) online, and if the receptionist could (on an opt in basis) get ahold of me via an instant messenger, or SMS page to reschedule or confirm my appointment.
Records online would be efficient for in the office, but I don't want even the chance that somebody can hack in to your network and get my records, so therefore keep your internal network separate from your internet stuff.
The next step though (and this can be fairly secure) is to be able to communicate with my doctor (not some substitute) via email, so that I can ask non life-threatening questions without having to phone and wait on hold. The doctor can then tell me if I should book an appointment or not.
Even email communications should be secure so that they aren't kept in plain text anywhere for some hacker to find.
my $0.02 worth
Definately the last one ;-)
Another comment...
R&D costs are probably one reason why NOKIA decided on linux... Most of the software is already there (except maybe for a few firewire drivers, etc) or almost there... There R&D costs would be (IMHO) much lower than trying to get the same thing Stable in a windoze box.
Also, because of GPL, there are no SW royalties (they don't have to pay Micro$oft for each box).
Sounds like a smart decision (business wise) to me.
I don't know.... I'm a hacker (at least I like to think I am), and I would buy a few for:
-X terminals/web browsing in any room of my house (if this thing doesn't come with ethernet built in, then a USBethernet device would do).
-Cheap clustering
-Web servers (I could serve alot of stuff off of 20GB)
-Video On Demand (I could hook this up to my CD Tower to serve VCD's)
Here is further proof that X sucks. It seems to me that it is frequently the impossible obstacle hindering the capabilities of almost every UN*X OS. I'd (conservatively) wager that its back asswards client/server model is actually useful 5% of the time, if that much.
<i><p>For those that need remote displays, LET THEM USE VNC. Let the rest of us move forward. I despise configuring X to do anything reasonable, and if it doesn't work at all, it's saner to simply break out the installation CD and try the install again than it is to try putting in monitor/video configuration by hand.
</p></i>
<p>I am currently in France sitting at an SGI, and remotely displaying Netscape that is running on a SUN (I eliminate colour flash problem that way, because the SGI has much better colour capabilities than older suns). I am displaying two xterms on remote sgi machines sitting somewhere in this office. I am displaying rxvt from my Mandrake7.1 machine in Calgary Canada that I am logged into via ssh. I can run licq from there, and can administer my mail/web/samba server, and my firewall. That's the power of X</p>
<p>
For games like Quake, Descent and the like, I agree X probably isn't the solution. But the power of Unix/Linux is that you can use the right tool for the job. You aren't constrained to one window manager that has to try to fill all roles, and ends up doing all of them half-assed.</p>
<p>And, just to get nit picky... At least if I shut down my machine and slap in a different video card, I can re-configure my Linux machine without hours of pain because of incompatible drivers. I simply reboot to a non X mode, configure and test X (I <b>might</b> have to install more modules off of my distribution CD). And go back into the default runlevel(5)</p>
<p>If you simply shut down your win9x box, swap a different video card, and reboot... You could be in for a world of hurt</p>