Slashdot Mirror
Linux Desktop Myths Examined
Call Me Black Cloud writes "NewsFactor Network has an overview of the $95.00 Gartner report titled, "Myths of Linux on the Desktop". It's a good look at several points from the perspective of a corporate user, not a home user."
The Gartner link is registration required, but not the overview. There are TWO links ....
Infuriate left and right
Or VBScript, JavaScript, even PerlScript if you have the ActiveState Perl distribution installed. Windows Scripting is capable of some neat things.
"People that quote themselves in their signatures bother me" - athakur999
I think that any enterprise rolling out linux on a large scale would be smart enough to go to a thin client, heavy server model. That's where the TCO argument starts to really support linux on the desktop. LTSP and such. As for your myths:
Linux will be less expensive:
Office is the lockin tool, much more than Windows itself. Running OpenOffice on Windows sounds like a great way to wade into a transition, taking the most bitter part of the medicine up front. If that's over with, the OS switch doesn't look nearly so daunting. Licence per license, linux is cheaper up front. Feel free to disprove that.
Linux is free:
Paying for an ERP package isn't much of a show-stopper. You're talking about buying expensive OS licenses + expensive ERP licenses versus buying inexpensive OS licenses + expensive ERP licenses. A shortage of available ERP programs for linux is a better argument, although there are several ways to access ERP systems running on windows or commercial Unix server from linux clients - thin or otherwise (so you're buying licenses for one server, and its client access rights to access that program). And with popularity in the enterprise will come native ERP programs.
Linux means no forced upgrades:
Of course linux shops will upgrade their systems to get newer, nicer software. An enterprise won't be running Redhat 9.0 in 2045 any more than they'll be running Windows XP or Mac OSX. The difference is whether you're paying out windows prices or linux distro prices every 3 years, and whether your company upgrades for business reasons or contractual Licensing 6.0 obligations. And whether you have a choice of vendors.
Linux Management is Easier:
This is where the thin client setup really pays off. Dumb graphics terminals with 5-years-ago pc hardware minus a hard drive (to fail) connected to top-notch, dependable server hardware, centrally managed. An extra 5,000 spent on a server for 300 less spent per client (x40). And good performance for the majority of 'enterprise' tasks.
Linux Has a Lower TCO:
Again LTSP. Simple, very-few-moving-parts, interchangeable-in-5-minutes clients and real server hardware with trinity dies RAID and multi-processors, and hot-swap power supplies.
Linux Means Longer Hardware Life:
Again, of course enterprises will do OS upgrades during a 6 to 8 year lifespan. They would with Windows too. Anyone know how many scheduled upgrades you'd have to go through with MS Licensing in that period of time? Again the licensing price difference. And variance in hardware makes life harder (and more expensive) for IT. For Windows, Linux, or any other OS. How is this a linux-myth-debunker?
Skills are transferrable:
This is a real hurdle for linux. But for how long? It seems like a matter of momentum. The more enterprises switch (in whole or in part) to linux, the more IT people will build their careers around it. The bigger hurdle is nick is back end-user skills and perceptions. Linux desktop environments have come a long way in the last few years, though.
Bottom Line:
Linux isn't going to dominate the desktop anytime soon, enterprise, personal, or other. And it won't be the end-all be-all bliss of computing nirvana where enterprises never upgrade software, and linux solves "cultural and political issues" (ha! that was my favorite part of your article) for companies. But I think it looks like a feasible way to reduce headaches and lower costs, and your article did nothing to change my mind.
The original article is on zdnet here
The AC writes:
"Your statements are nothing but platitiudes with no basis in fact. In short, thanks, but no thanks for the FUD. You are not helping the Linux movement by offering such ignorant statements. It makes us all look stupid."
So the web server management, user account managment service startup, firewall managment, hardware configuration and the like can all be configured in Win2K using PERL and other commandline utilities?
You make yourself look pretty silly when you champion the utility of an Operating System whose designers themselves admit the inflexibility of. As we argue, Windows engineers are trying to figure out a way to add a usefull file based configuration and command line shell to the next release of windows.
.bat isn't a script, it's a batch file.
And an MCSD wouldn't write one to do anything in windows - it's a DOS construct, he'd write a VBScript, JavaScript, or - if he wanted to be just like you - write it in perl or bash or tcl.
Every userland language/app that exists for linux exists for NT.
I don't need no instructions to know how to rock!!!!
Although this has nothing to do with Linux TCO vs. Windows TCO, here goes:
list of open file handles - how about Handle
list of bound ports - how about TCPView
robust scripting language - how about ActivePerl
All of the above tools are free, high quality, and easy to install/use on WIndows 2K/XP. I automatically install them along with many other tools whenever I prep a new Windows 2K/XP machine. And to think I'm not even an MCSE...
That's only true of the Win95/98/ME series. WinNT/2K/XP has the capability to set permissions so that not every program has access to every piece of the system. In fact, Windows ACLs are much finer grained than traditional Unix rwx type permissions; it's easy for any user to set access to his files on a person-by-person basis. I don't think that they're usually used very well, but it's certainly possible for a competent admin to lock down a Windows system pretty tightly against abuse by ordinary users.
There's no point in questioning authority if you aren't going to listen to the answers.
You don't have to buy Office for Windows, you can download Open Office for Windows just like you can for Linux.
You think all those apps on Linux are from a single vendor? If MS tried to bundle all that with the OS they'd be hauled into court again.
The whole issue with support is infected with the proprietary thinking model, which is based on the idea that you are stupid, the vendor is smart, and even if you aren't stupid the vendor has access to the real information, which they won't share because it would harm their support revenue.
Support for open source IS FREE (dammit!). The support process is as follows.
1) Use google to search the web for keywords on your question. This give you access to info in howtos, FAQ's, and basic documentation. IF your question is basic it will be solved here.
2) If this doesn't work, use google groups tp search for keywords on your question. This gives you access to USENET posting. If your question isn't answered in the basic docs, but is common enough that somebody else has had the issue, then it will be solved here. This is about the level of question that standard 1st level for-pay support will be able to solve.
3) If the above two steps fail, it's probably a hard problem or an obscure problem or a very new problem. Join the mailing list for the project in question. Search and read the archives for related issues. Then POST to the list and state your question and what you were able to find. Often you will be able to solve your issue. This is about the level of 2nd level for-pay support.
4) If contacting the mailing list reveals that the problem is a known issue or can't diagnose it, then you should at least be able to find out what part of the code causes the problem. Go get the code and figure out who the maintainer is. Write to them personally and explain what the problem is and what you have done and what steps can demonstrate the problem. If the project has a bug tracking database, enter your bug in it and refer the maintainer to it. Some times you will get a patch emailed to you. This is about the level of very expensive 3rd-level for-pay support.
5) If all of the above fails, download the source and start hacking. You have everything you need with open source. While you are doing this, continue to use methods 1 to 4, and for god's sake, post the patch back to the mailing list when you are done. This level of support is not available at any price from proprietary software vendors.
Your script is, enterprise speaking, useless. It doesn't handle computers that are disconnected or offline. It also doesn't show any error output. If an install fails, you won't know. It doesn't check for nor deal with dependencies. It doesn't allow for static nor dynamic groups. It assumes the rpm is on the end nodes, and even if you pull from a central site it can't load balance or throttle. There is no metering or monitoring capabilities. That is not a mass deployment script. I'm talking about 5000, 10000 or 50000 nodes, not 50.
Ok, you need to readup on the DOSkey command
/MACROFILE='. You can create a custom command prompt icon which will set up your environment with that, etc.
doskey c2 = cd "c:/winnt/system32/documents and settings/administrator/my documents/code/C++/ch02"
doskey p4 = cd "c:\winnt\documents and settings\administrator\my documents\code\perl\ch04"
etc... You can save these associations to a file and recall them using 'DOSKEY
"Still Unix rocks in this regard."
Ok next complaint.
If you going to consider OpenOffice and other OSS then why not consider one of the Linux distros....If your going to use OpenOffice then you are half way there .... Take some time learn one of the Distros (RedHat, SuSe, Mandrake) or customize one yourself. It is truly your option.
.... Hell, I was/am a Microsoft Sys Admin for 8years. My point is that busnesses do not fork out thousands upon thousands of dollars to Microsoft and turn around and use Open Office. They follow the "support/ compatibility" trail and buy Office, SQL, Exchange, and the development tools recommened by those same vendors.
....But there again I have to rely on third party vendors to solve this problem.... Not a simple little button click like mozilla.
.... was it easy the first time ? No ....Is it easy now? Yes. All it took was time and effort (and no money).
I have come from many Microsoft Shops
An OS is used to deliver applications that is what Linux does. It is not a vehicle for big companies to try to sell me something everytime if I open a browser. yeah yeah I know I can get pop-up blockers and such
Is Linux ready for the Desktop?
It Is if you want it to be.... I am completly MS free on my desktop
My tools are free..... My Skills are not. If my customer has to by my tools before I can go to work (XP, Office, Ghost, WIN2K server, etc) that means less money for me and makes me noting more than service tech. (Like the copier guy)
If I bring my tools with me ie..Linux, Apache, MySQL, etc, etc. Then that money comes to me in the form of my skills. Why should the client care what it runs as long as it runs.
THAT IS WHERE TRUE TCO STARTS.... It takes the same skills to manage MS and Linux..... But the middle man (Microsoft) gets a third of you potential income/revenue....
There once was a cluster of 450 Win2k servers on Dell Servers running a coputational application.
They could run a 100 nodes with a 75% reliability and anything more it crashed and required a FULL DAY to reboot the servers.
I was tasked to convert those server to RH7.3.
Once I had Kickstart figured out I redeployed those same servers with Linux, ready to run production jobs in A DAY.... And I did at no cost to the company.... It was so successful the company is now trying to get there money back on the MS licenses. (FAT CHANCE) but they felt so betryed by MS that they felt it was at least worth the effort.
Choose your tools well my fellow craftsman... Choose the ones that will benefit you not relagate you just a handyman for Microsoft.
You can argue perfomance and ease of use all you want but at the end of the day it comes down to money and how much you get. And if the company you work for uses a Microsoft for thier tools, that makes you Microsoft's bitch. Microsoft gets paid first.
Large businesses who use windows don't have this problem. All data is stored on the network. A ghost install is created and a machine can be re-ghosted in a matter of minutes. Problems are rarley fixed. The machine is just re-ghosted. If I spend more than 15 minutes trying to fix a problem it's time for a re-ghost. I know that sounds like I just can't fix anything but it really turns out to be ths most cost-effective way to do things for us. We have 1100 machines in my building and we support them with 4 techs who are really not very busy.
Windows defragmentation toll was written by a good christian company and banned by Nazi germany.
DisKeeper was created by Executive Software, who are a Scientologist company. The Nazi Germany predates Scientology. It was the clearly non-Nazi German government who made doing business with the Church of Scientology illegal.
Ahh. I found the KnowledgeBase article I was looking for:
You can see this is a pretty lousy way of doing things, and I think it's a pretty typical example of how things are in Microsoftland.
Assuming that this is an accurate synopsis, I have some things to say about the report.
Let's move away from the "Staroffice" thing and look at the fact that you don't have to pay for the OS, and that linux server support costs less than windows desktop support.
"Supported versions" of linux aren't free? Uh, you can't make a version of linux that ain't free, because it depends on GPL'd code. You do have to pay for autoupdate services... But at the worst, you might have to pay for the "supported" version once. Paying for service contracts is a separate issue, touched upon above. It's not part of the free/not free issue since you have to pay for it above and beyond purchasing the software in the cases of both windows and linux.
That really is a myth. If nothing else you will have to upgrade glibc and the toolchain; upgrade glibc enough and you have problems.
Depends. If you're the kind of person who is at the level of scripting but not programming then this is probably not a myth. Linux is easier to script together because it's Unix. On the other hand, Windows has pretty good automation ability, you just have to use COM (or .NET now, I guess, in 2k3?) to use some of it. Then again, most of it can be vbscripted (shudder) so this is probably somewhere in the middle. Certainly it is easier to write tools to automate Unix than Windows NT.
I just wanted to display the whole thing to try to point out how this is a meaningless argument. TCO doesn't just cover permissions and procurement issues, it also involves initial cost and support cost which are demonstrably lower with Linux in simple dollar amounts based on support costs.
On the other hand, Windows' permissions system would seem to be better suited to management of organizations than Linux's, since Windows uses ACLs, and while Linux is capable of it, no one does it and the user space tools just aren't there.
Okay, this is so ignorant I don't even know where to start, so I guess I'll begin at the top. The very first paragraph starts out confirming the "Myth". It says right at the beginning that it's true that you can use older PCs with a current version of Linux to run software which windows won't run on the same hardware. Thanks, Silver! Then it is ruined by saying
"You're right," Fisheye says. "I should have set it on 'whip' or 'chop.'"
If all you have are OEM licenses, you cannot use GHOST:
g ust_IIA_2002.ppt
http://www.microsoft.com/usa/presentations/SAM_Au
Engineering and the Ultimate
By basis for speculating that they only contacted RedHat out of the Linux vendors is quote from the overview that you seem to have dropped from my statement, namely that the Linux vendors only support their releases for two years. This is RedHat's new policy, but not the policy of most major Linux vendors; in fact it wasn't even RedHat's policy a year ago.
I don't know they didn't do a TCO study, but it's strongly implied because rather than saying that their TCO study supports or disagrees with the "myth", they merely say they don't "beleive" that the Linux world is enough different from the Linux world to make a difference here. Since it's something that they can check, and it's something that they ought to have checked, why didn't they check? Of course I don't expect them to give their TCO study for free, but I also think it's pointless to pay for a study that says that the price savings of Linux are mythological unless they give some evidence that they've actually looked at the real world examples in front of them.
I say this is poorly researched because the overview contains some glaring untruths like that one, and they strike me as the untruths that result from lack of research. I say this is poorly researched because the overview talks about belief when they could have talked real world figures (or at least advertised that real world figures were inside).
The fact that Gartner Group is more listned to than Bryan Consulting or Paul Murphy is a sad truth of modern times, but that doesn't change the fact that I wasn't linking to them because they are better, I was linking to the information they were discussing. Specifically, Bryan Consulting was discussing the Burlington Coat Factory enterprise migration to Linux and the cost savings associated with it, and Paul Murphy has done a comprehensive and detailed TCO study that comes up with a vastly different result than Gartner Group's "belief".
I don't care how many clients or employees or links Gartner Group has compared to Bryan Consulting or Paul Murphy (at least call the man by his name, not by his website). I was not recommending one over the other in terms of consulting services. I was just pointing out that there is real information out there while the Gartner Group sounds like they're guessing.
"The $95 is for the real report NOT this overview that you have read!"
Wow! What a revelation! I never would have realized that, given the fact that I didn't spend $95 to read the overview.
"Are you trolling or are you serious? I really don't know?"
I am quite serious, the Garter Group sounds like they are charging $95 for a report, and the overview of that report gives every indication that they didn't do their homework.
"but when I learnt about IT I was told it was Gartner who invented the Term TCO."
Whether they coined the term it or not, of course Gartner knows what a TCO study is, it's one of the things that they sell as a business. The fact that they know what one is makes no difference if they haven't done one.
----
Open mind, insert foot.