News | Mail Archive | OS Software Downloads Ad Info ::
Subject: Databases | Java | Linux | Open Source | XML | Data | Tech


Contribute:
· News/Reviews/Release

Misc:
· My Account
· Editorial Feedback
· Logout

Login
 Username
 Password
 Remember me


 Become a Member!
 Login Problems?

Recently Updated Mail Archives
bug-hurd-gnu
debian-announce
hypertable-user
debian-security-announce
fedora-package-announce
solr-user.lucene.apache.org
dev-apr-apache
enterprise-watch-list
gtk-app-devel-list
announce.maven.apache.org
dice-tech-jobs
tracker-list
sqlite-users
user-roller.apache.org
unassigned-bugs-kde
httpd-apache
fedora-docs-list
bug-gsl-gnu
info-gnus-english-gnu
mapserver-development
fedora-infrastructure-list
dev-qpid.apache.org
centos-docs
bug-gnuzilla-gnu
kde-users
savannah-register-public-gnu
centos-es
sage-support
auctex-devel-gnu
user.ode.apache.org
NameLast Updated
Popular Mail Lists: Daily Proxies OpenStack nginx modpagespeed Android windows linux solaris osx ubuntu fedora enterprise crm ruby python java xml perl php cvs subversion version contol db
database mysql postgresql mobile telephony voip apple apache
all
sitemap (mail)




Posted Mar 02, 2005

Part IV: Corporate Desktop Linux - The Hard Truth

by W. McDonald Buck, Retired CTO of World Bank

In previous articles (Part I | Part II | Part III) I've advanced the proposition that the reason Linux has failed to garner any big wins in the corporate desktop market, is that the cost/benefit case is not, after all, very compelling, despite what we would like to believe... Surely a shift to linux will bring advantages that are worth the switch. Unfortunately, the hard truth again is that real dollar benefits are hard to justify. [This series of articles is based on an avg. 250 non-technical seats]

The Hard Truth About Linux on the Desktop – The Benefit Side

In previous articles I've advanced the proposition that the reason Linux has failed to garner any big wins in the corporate desktop market, is that the cost/benefit case is not, after all, very compelling, despite what we would like to believe. The most recent article explored the cost side from the perspective of the corporate CIO, and concluded that the cost picture of switching to Linux on the desktop is not very rosy. While there are savings in the range of $115,000 over three years offsetting costs are present too (we identified at least $50,000 to $60,000) and there are significant risks.

But even if that is true, what about the benefits? Surely a shift to linux will bring advantages that are worth the switch. Unfortunately, the hard truth again is that real dollar benefits are hard to justify. Consider the dimensions of ease of use, security, software availability, stability, support costs, lengthened hardware life and adaptability.

Adaptability: One of the great strength of Linux is its adaptability. There are many ways to do things, and its easy to add additional functionality or tailor things the way you like. One can evolve the software as needed at the point of use. Though you can wait for a new distribution, you need not do so if you need to be on the bleeding edge of something important to you. No waiting for months or years. In a corporate server environment, IT professionals confronted with a new problem can solicit help, download and install a fix in a matter of hours. Even in the home or small office environment, a skilled owner can download and install the latest versions and fixes of individual packages, something completely impossible with Windows.

But here's the rub. These benefits accrue only to those willing and able to change the software environment: to select among packages, switch versions, install upgrades or apply fixes. An owner/user unable to change the environment gets no advantage from software adaptability. Unfortunately these are the conditions in the two places comprising the majority of destktops: the home/small office consumer with no IT staff and the corporate environment. The last thing most unskilled or semi-skilled consumers want is to make changes to their computer, or be confronted with technical choices they don't understand. In the corporporate environment underfunded IT staffs have learned that the key to successful support is lock-down standardization – keep all the desktops configured as much alike as possible; change is disruptive and costly, so it requires careful planning so it can be done quickly in large populations at once. In many corporate environments users are strongly discouraged, and prevented if possible, from making changes to their environments. Of what value is the ability to adapt if adaptation is forbidden?

These are the environments in which Linux is not making headway, because in neither of these environments is Linux' adaptability a benefit.

Ease of use: Nobody could disagree that linux has made enormous strides in ease of use for novices, with better graphical environments. But I think it would be hard to credibly claim that linux has made up Windows' head start of 10 years and hundreds of millions of dollars. The best a linux fan could hope to claim, for the novice user, is a tie. And that isn't even true.

Some commentators have pointed out that the development community just hasn't spent the time and energy on ease of use that Microsoft and Apple have. I think that is true. I think there is another factor too, which has to do with having alternatives. I claim you have to know more to use Linux because the user has to make more choices. I argue that an excess of software choice actually operates to reduce ease of use. The presence of a lot of alternatives means choices must be made, and while it is great to have choice if you know how to choose, novices finds it at best baffling, and at worst mind-numbingly complex.

One reason for so much choice is the absence of any force to drive out the second best or third best idea or implementation. The commercial marketplace sometimes drives out good ideas in favor of others that are not as good, but right or wrong it usually picks winners. Where costs and profits are concerned, a company like Microsoft will settle among alternatives. They may get it wrong, in some sense, but the customer sees only one solution, one way to do things. The open source marketplace is not as good at picking winners and losers, the second or third most popular toolkit can continue to thrive, and be a key component of several packages which may be best of breed in their own space. A rich distribution ends up offering everything. This isn't a bad thing, but it does result in an environment which has lots of choices. You don't have to install it all, but if you don't want it all you have to know how to choose. If you do install it all, then you still have to choose what program or tool to use.

Ease of use is subjective. I could write a whole article on why I think Linux has deficiencies here, and you would have your own view. But when my not so computer literate friends ask for help, I don't try to get them to migrate to Linux. I do download OpenOffice for them, and suggest they give it a try before buying Office. I do suggest they change browsers and mail clients. But I leave them on Windows, at least for now.

The bottom line is, I don't think anybody would claim that increased ease of use can be counted as a benefit of switching to Linux.

Security: Linux has a clear win on security, both practically (less malicious code targets linux, which will be true for the foreseeable future) and theoretically (openness is a superior path to security). Curiously this seems not to be the big selling point that it should be in the smaller and medium sized companies which are hit the hardest.

I think it is important to note that this is not as big a problem for the bigger companies, because the problem doesn't scale linearly. The economics of managing thousands of PCs are different. There is a threshold amount of money that must be spent to achieve an acceptable level of protection in a large population of computers with tools like SMS and Group Policy. As the number of boxes and the size of the risk increases to that threshold the cost of protection becomes cost effective for the population. Sure it would be great to avoid spending that money, but a company fielding 5,000 to 10,000 computers is already spending $10M or $20M on it, and the cost of an effective protection program together with specialists to manage it is not only not the large companies biggest problem, it is a pretty small percentage of the total spend. It is possible that another huge new disruption is right around the corner which will overcome their defenses, but the big company will be able to react pretty fast.

Our hypothetical company with 250 seats arguably has more of a problem with security, since they probably can't justify the cost of such a program. One continuoually reads horror stories of IT staffs spending huge shares of their time fighting recurring malware attacks. This is a big cost. If there is a factor which should drive the switch in small and medium sized companies, I think this would be it. But it isn't happening. Why don't more of the smaller companies switch? I wonder if it could be that smaller companies are watching what the bigger companies are doing? Though the head IT guy may see the fallacy of it, perhaps the CEO or the head of finance is wary of moving to Linux if Gartner isn't reporting that bigger companies are doing so?

In any case, I think a company switching desktops to Linux would get a clear measurable benefit from moving to Linux in terms of time lost to viruses and worms. It is an open question why more small companies don't seem to believe this.

Software availability: This is a weak spot for Linux, since nearly everything is built for the huge Windows market first. Beyond the cases where solutions exist (e.g. OpenOffice), we point at emulation (Wine and such), and at alternative approaches to legacy apps (e.g. Citrix). But even if Linux could claim equality, it could not claim superiority and truthfully equality is not yet attained. A large part of the mainstream software known to non-technical people just isn't available for Linux. Linux has lots of tools, but fewer finished products. I mostly live in Linux, but I still have things in Windows I can't do in Linux.

Stability: On stability, the nod goes to Linux, but the margin is no longer what it once was. Windows 2000 and Windows XP are real operating systems with far better stability than Windows 98 or its crash-dummy ancestors. Instead of saving 3-10 reboots a week, the margin is now probably more on the order of 1-2 a month, maybe less. Some argue there is no significant difference anymore. Without quibbling about the actual number, the dollar savings is no longer a big factor.

Support costs: People who know about Windows are everywhere, as are companies doing support. Linux support people may be a little harder to find. Simple economics suggests Linux support per staff hour won't be a lot cheaper. Will there be fewer hours? Linux advocates are pretty sure administration of Linux servers will be easier once all the scripts are built, and that more servers can be supported per administrator; but desktop support is different. A lot of front line user support for desktop users is really training in disguise, and much of that is on applications. It seems unlikely to me that there will be less support required after changing to a new operating system, particularly one with so many choices. Over a longer time horizon, perhaps this changes, but that is speculative. In the meantime I don't believe Linux will cost less to support (security issues aside).

When support costs are mentioned, an area in which a fair assessment would have to give Windows an advantage in a corporate environment is the large scale management already mentioned, with things like SMS, Active Directory, and Group Policy. It costs a bit of money to set this up, but in a corporate environment the support savings mount up really fast. I don't know about a Linux equivalent (though of course there are tools like LDAP from which you could try to build your own). There could be some solution that I don't know about, and if so that will nullify what otherwise I have to regard as a serious Windows advantage.

Hardware Lifetime: Linux will run on lower end hardware, to be sure. But this argument won't sell many deals. The question whether a new operating system will fit on old hardware doesn't arise for those in the real world who care about applications rather than upgrading operating systems. The computers running a business application will have to be replaced when the computers break or the application no longer performs acceptably. Switching to linux isn't a factor for the first, and only marginal for the second (presumably it is the application, not the operating system which consumes the bulk of the resources, even on Windows). When the upgrade happens, it's true too that linux can be used to rescue the old equipment, and eke out some value, but in a cost benefit analysis this is fully depreciated equipment being used for applications that couldn't otherwise be cost justified. The hardware lifetime argument isn't going to sway many corporate CFOs.

I'm sure there are some benefits I've missed, and I'm sure others will point them out. But my bottom line argument is that the major benefits of Desktop Linux for non-technical people are better protection from viruses and perhaps a little better stability. These are offset by Windows advantages: availability of more applications, including especially legacy apps, familiarity coupled with greater ease of use, and advantages in large scale support costs.

The hard truth is that the benefits that are most important to individual technical people are simply not important to those lacking technical skills. When you couple this with the relatively meager hard dollar cost savings, the prospect of some extra costs of migration, and the large risks of such a move, is it any wonder few corporate customers are making the transition?

We should recognize that this scenario is not going to change very quickly, and stop fighting a defensive unwinnable war for desktop migration. Desktop progress will be slow. The biggest near term wins will come from migrating the applications, not the operating system. You can switch to open applications one at a time, slowly, and save money doing it. And you don't have to migrate everyone. The OpenOffice spreadsheet doesn't have to be as good as Excel (and it isn't), it only has to be good enough for 90% of the users. The wizard with a need for every feature can keep his copy of Excel, or Word or Photoshop, and the bulk of the users migrated. Once that is done switching the OS may become relatively easier.

It is great that Novell and Sun want to invest in the fight, I wish them well and great success, but I'm not expecting them to gain a lot of ground in the near term. In the meantime, the same benefits which are unimportant to unskilled desktop users are powerful arguments for the IT professionals who manage servers and embedded environments. It is here also Microsoft helps by charging truly exorbitant licensing fees for the same Windows. Many observers expect Linux to eat Windows' lunch in the back room, and in the set-top box and other embedded environments. I'm among them.

Corporate Desktop Linux - The Hard Truth is a four part essay: [Part I | Part II | Part III | Part IV]


blog comments powered by Disqus



Re: Part IV: Corporate Desktop Linux - The Hard Truth (Score: 0)
by Anonymous on Mar 02, 2005 - 11:24 AM
How about a discussion of OS X replacing windows desktops? This is more feasible. NOt only do you get a better interface, but it will cost less money (post migration)


Re: Part IV: Corporate Desktop Linux - The Hard Truth (Score: 2, Insightful)
by element on Mar 02, 2005 - 11:47 AM
I can see your point in some instances, like it will be a very slow change and really people should expect results over night. However on the Corporate desktop the only real problem is applications, the others are far more mythical in reality. Application will be the real determining factor in how quickly linux is going to get adopted. Items like easy of use or pretty much a mute point in a Corporate environment where users only use a handful of application and by almost all means should not be playing with configuration and installing software, which is where easy of use is really important, I don't mean to say that the Linux desktop can't be improved, because it can. I'm just saying that the whole training myth is just that a myth, sit down with someone who is non-technical and only knows Windows and show them how the boot up linux and where the menu is, they'll be on their way working in time at all. I've seen this time and time again. To be honest though it kind of depends on the DE though, I have a far more challanging time with KDE and there kitchen sink appoarch to menu, as compared to GNOME. Also as a support person I take far more calls from Windows users than Linux users, this is straight accross the board for technical and non technical types on both OS's. One decent Linux Support person can fill the position of at least Five well trained Windows support people if the environment is switched. Also a help desk receives most of the calls about the network and people lossing passwords, alot of the network problems are traced back to windows in my experience, those that aren't server related and it has nothing to do with the setup, usually someone kick the ethernet cord and then windows doesn't reconnect correctly when they plug it in.

Also the point on too many choice should really never be a problem in a Corporate environment, users should not have a choice, they use what the admin installs and for an admin choices are a good thing.

I can see where you come up with your reasoning and how it better suites a Personal Computer user more than a Corporate one. But by most means the truth is alot of the reason linux won't succeed is because of Myths about support easy of use and FUD, the only really factor keeping linux off Corporate desktops is Applications. Plain and simple Application are lacking, you may have a good office suite, a thousand mp3 players, hundered of desktop options but very few business applications, stuff like gnucash is nice but it's not only showing it's age, i far more complicated than the windows counter parts. More commercial applications would be a huge boost, but therein lies problem, most commercial vendor will not port to Linux as there is not a big enough user base and there is the perception that people who use linux won't pay since they get too much free.


Re: Part IV: Corporate Desktop Linux - The Hard Truth (Score: 1, Insightful)
by Anonymous on Mar 02, 2005 - 12:19 PM
"I think it is important to note that this is not as big a problem for the bigger companies, because the problem doesn't scale linearly. The economics of managing thousands of PCs are different."

There's a lot of hogwash hidden in the details of your security discussion. If a compromised box was only a problem for that user at that station, some of your analysis would be more useful. However the ability of compromised boxes to be staging grounds for compromising other boxes almost reverses the analysis of scale: the larger the number of seats, the more likely that a low probability event happens and the even a single low probability event may lead to further compromises and to catastrophic losses. Sure you can afford more expensive security solutions but they do not eliminate all of the consequences of the rare but costly event.

My quite large regional bank had computer operations totally shut down for customers and nearly totally shut down for branch offices a little over a year ago because of "virus infections." The details were never made public as far as I know, but these compromises occurred at a time of high Windows virus activity and the problem was OBVIOUSLY very expensive to fix. The large scale of desktop use made for a large scale problem that needed lots of time and money to solve.

I think that as large organizations with large numbers of seats get hit and spend huge amounts on recovery the security advantages and administrative advantages of Linux become overwhelming.


Re: Part IV: Corporate Desktop Linux - The Hard Truth (Score: 1, Insightful)
by Anonymous on Mar 02, 2005 - 12:28 PM
I think the "software choice makes the UI harder" is actually not true for the corporate desktop. The fact is that the choices available are governed by the system administrator, which can remove as many choices as necessary for the end-users. That does mean you need a system administration staff skilled in Linux -- but I don't think that people who believe that Linux can be a success meant that system administrators didn't need to already be skilled.

However, your article does hit on the side of one issue that is very present throughout IT -- that many system administrators throughout the industry really have no skills, or very few skills. I have been shocked by what passes as a system administrator or even a developer in corporate environments. It is obvious that the colleges who have become SysAdmin mills have done a great disservice to the industry.

The fact is that the choices in Linux are meant to be made by the system administrator, and that system administrator is expected to be skilled. The end-user would wind up with very few choices -- the choices that they are expected to make at their level, as determined by the system administrators.


Re: Part IV: Corporate Desktop Linux - The Hard Truth (Score: 0)
by Anonymous on Mar 02, 2005 - 12:47 PM
Re: Security "Our hypothetical company with 250 seats"

Why not have a Linux thin client server to run a web browser ( firefox )and a email client ( evolution or thunderbird )applications. And have all output from these applications displayed back on the Windows Desktop PCs ( in a manner like Windows terminal services or Linux thin clients ).

Users who wish to access the Internet "logon" to this server and get a "firefox session" back on their windows machine ( in a manner like Windows terminal services session except CALs are not needed to connect to the server ) .

Files may be transfered or accessed ( quite transparently ) via Samba which will be running on the Linux box. Access to files can be both ways or be restricted .

In this way, you can still have all your Windows applications and familiar Desktop environment, but have a measure of "insulation" against new viruses and Trojans.

Its easily done with Cygwin-X and any Linux distro can be turned into a thin client server with very minor changes.

In this way, Linux performs a both "server and desktop applications" role but in ways which are "least disruptive" to users while helping an organisation address some security issues.



Re: Part IV: Corporate Desktop Linux - The Hard Truth (Score: 1)
by gfkdrobi on Mar 02, 2005 - 02:23 PM
A very thorough, level-headed and practical approach to this subject, which frequently is involved in the prototypical 'religious' discussions that are the bane of our existence it seems. Having served in a variety of technical posts since I first ran Linux on a 286-based desktop as a successful router, to today where I find myself in an advisory role to multiple entities on the subject, I'm happy that the pros/cons of this exciting choice have been so thoroughly examined. Congrats, & I plan to excerpt many of the points presented (with due credit, of course) in some upcoming discussions with financial, technical and end-user communities.

Couldn't have said anything better.......

Cheers,
- gfkdrobi -


Yes and No (Score: 0)
by Anonymous on Mar 02, 2005 - 02:42 PM
I ran a Cybercafe with mainly Linux based systems for one year (15 workstations) and have managed a lab (ten workstations) for nearly three years. While much of what you said seems true to my experience, here is where I can clearly refute your views:

Ease of use: This might sound unbelieveable but no user has EVER had a usability question, pertaining to the operating system itself. My questions pertain to how to (in this order):
1. How to do something on a website
2. How to do something pertaining to OpenOffice

Users in the cyber cafe mostly browsed the web, but they also played games and did homework, while at the lab they do mostly word processing/spreadsheet work, graphics editing, and sometimes web development. Sometimes specialized Windows applications were run at both locations via Windows Terminal Services / rdesktop. Some people would reject the idea of using OpenOffice immediately, but those who were willing to try only needed a little time to adjust. The vast majority of people didn't react at all to the fact that the underlying operating system was Linux. Those who did had "oh really" or "that's nice" like responses.

Our major problems have been mainly printers. Although initially, we also had massive problems keeping mozilla working...when it crashes, it crashes an entire user session. Thunderbird solved the problem sufficiently.

I want to write a lot more, but one other thing important to mention is the assumption that Linux should be deployed in the same way as Windows. Frankly, there is no benefit to not using thin clients with Linux. There are huge savings in dollars and time and capabilities for using thin clients. Both reliability and performance can easily exceed using PCs on every desk. Additional capabilities include (but are not limited to) being able to login to your desktop at anyone's desk, having a lot more computing power at your disposal on demand, far easier management, etc. and a far better ability to provide workstation support from afar.

First, you buy a three or more (at least two for failover and one for testing updates/modifications) redundant application servers on one power SAN; next, buy low-cost (~$200) diskless workstations for each desk. Use KDE and KDE's desktop sharing facility for workstation support. Tech support will never have to personally visit a desk... If hardware goes bad, users can grab another from out of a cabinet, but it isn't likely to go bad being a 100% solid-state device.

Cost Savings:
- Workstation support cost savings should be somewhere around 90%
- Past an initial threshhold, hardware savings kick in at around 50% - 70% for initial purchase price
- Ongoing hardware costs are dramatically reduced, as workstation lifespans increase to 15 to 30 years and servers and storage need only be added, not replaced.

I am wasting precious time I need to spend working right now... But if interesting in how this all works, email me at teddemc@yahoo.com (this email has pretty good spam protection)

Matthew


Re: Part IV: Corporate Desktop Linux - The Hard Truth (Score: 0)
by Anonymous on Mar 02, 2005 - 03:51 PM
ABOUT SECURITY

You don't show any understanding of the MAIN reason why Linux has better security than Windows. It's not the often-claimed (by Steve Ballmer, et al) "less attractive target, 'cause hardly anyone uses it" thing, and it's not the "Open Code has more eyeballs"' although that probably helps.

It's the firm line between user and system context. The difference between Windows "Administrator" and "non-Administrator" context is a fuzzy hairball tied only to the files they get to change, and not the processes they get to run. It's been slapped on to a single-user "system", piece by piece from the days of DOS-- a single user Operating System. In contrast, Linux acquired its basic design roots from time shared multi-user Operating Systems.

Maybe MS-'Longhorn' will be better than Windows-XP in implementing a true User Context. If so, you can probably expect to see significant migration difficulties, along with the inveitable next loop of the never-ending license fee treadmill.



Re: Questioning the assumptions (Score: 0)
by Anonymous on Mar 02, 2005 - 04:09 PM

Mr. Buck obviously knows what he's talking about, and has done his homework. I wouldn't question his analysis.

What I would quibble about is a couple of the assumptions behind this series of articles. First, that buying new hardware is necessarily synchonized to replacing or upgrading software. Second, that the entire organization has to upgrade or migrate at the same time.

One of the reasons software, especially system software, needs to be periodically upgraded is to keep up with security fixes. Unfortunately, with proprietary systems that often means upgrading to much more bloated software, which won't run on the existing hardware, or at least won't run well. This can force a premature hardware replacement. This can happen with open source software too, but it happens less frequently. So a hardware replacement can be postponed for one or two extra major upgrade cycles. For example, what finally forced me to retire my 486 workstation was Mozilla Firefox; getting adequate performance out of it requires more than the maxed-out 32 MB RAM on the old machine.

There have been some accounts by sysadmins who have introduced open source applications and platforms user-by-user, rather than doing a massive cutover. One small government administrator, I think in Florida, started giving new hires OpenOffice and Linux, unless they specially requested MS Office and Windows. Few did. He left existing users alone. Doing it this way obviously takes a long time to gain the full benefits of bringing everybody over, but it greatly mitigates the up-front risks of introducing new software.

A few of the other commentors mentioned web-enabled applications, which make the client side platform-independent. This reduces the need to maintain a legacy desktop platform, but good luck applying the old cattle prod to application vendors, to get it done in any kind of hurry.



Re: Part IV: Corporate Desktop Linux - The Hard Truth (Score: 0)
by Anonymous on Mar 02, 2005 - 04:34 PM
Thanks for an honest, and from my point of view accurate, analysis.

I know something of running small businesses, and the major reason they don't see the benefit of going to Linux is change. A change means money, and I've never seen a small business that wasn't short of money. One reason they would benefit would be because a small business can not afford a 2-3 year equipment lifecycle.

One other comment I'd like to make is on Linux usability from a non-technical user's standpoint. In spite of what Linux fans say, it ain't there yet, folks, and it is still horridly behind for usability. Once it is installed, with all the programs running and the learning curve behind you, everything much closer in quality to Windows and Mac. But getting to that point is much harder than Windows or Mac. Most users don't care what the OS is or the hardware - they want to get the work done first. I'm a Windows and Office expert, and I went thru 5 distros of Linux to get two working installs last year. The average user won't put up with that.

Spokesinger


Re: Part IV: Corporate Desktop Linux - The Hard Truth (Score: 0)
by Anonymous on Mar 02, 2005 - 05:54 PM
I must say I really do enjoy reading your analysis from the corporate side, as usually arguments are fought o­n the technical side between Windows and Linux, which often becomes somewhat irrelevant as to the real challenges an IT department faces.
One issue though that I believe you have missed while talking about security is about the number of security holes being discovered, the time it takes to release a patch and the guarantee that such patches will not disrupt the current workflow. From a Gartner report I saw, reported that in a time frame of couple of months, Windows was publishing much fewer security bulletins than most of the other widely used Linux distributions, and the time to publish a patch was much lower. I am extremely interested in reports that dispute or discuss this, as my experience is low o­n this matter. Another factor, that I only know from the Windows side is that patches go through rigorous testing by Microsoft, as they already have the infrastructure to do this, and they have huge quality expectations from large customers, large number of customers and partners that they must guarantee that workflow won't be broken with the installation of these patches.

Samoil


You call this the "Hard Truth"? (Score: 1)
by zapyon on Mar 03, 2005 - 12:26 AM

I am sorry, Mr. McDonald Buck,
while I do see that some of your points are valid, and I even agree on others, your series of "Hard Truth" articles does leave me with a certain FUDdy feeling.

From part 1 I have been wondering how you can go and -- like those other migration 'studies' -- just set once scenario (that was rightly questioned by comment-posters after the first part) and claim that your personal opinion is "The Hard Truth". Sounds like FUD to me.

Further, you claim that big companies don't migrate. Are you really sure? Have you consideres researching the opinions, experiences -- and successes -- of companies in other countries than the US? Public offices, hospitals, medium and big companies are migrating in Europe, Brazil, France, China and many other countries are going the Open Source way in public. Don't you think there would be many case studies available? I know that quite a few are even publicly available in cases in Europe. And I remember having read such story about an American city as well -- even the police's tablet PCs were migrated successfully.

You advertise your former position as CTO of the World Bank. And you are retired (So I assume you have a lot of time for research). I am sorry, but I had expected a much more thorough approach to the topic when I read this the first time. After reading the fourth part just it just increases the FUDdy taste in my mouth. Please, if you publish an opinion piece, don't call it "The Hard Truth". If you plan to really provide hard facts, do your research properly and quote your sources. THAT would make your article(s) a lot more valuable. A thorough review of existing published data and case studies could have yielded a lot more valuable information. As your articles are right now, I fear they will rather scare off people who are just starting to think about the possibility of migrating their desktops, should they happen to come across your opinion pieces through a Google search.

I disagree with you on the point where you say that too much choice makes Linux somehow difficult to utilise on company desktops. This simply not true: many migrants have proven that the adaptability is a great help, especially when it comes to locking down the machines to a specific setup and appearance. Also, there is the -- admittedly still young -- 2.x branch of GNOME with is making rapid advances towards a really user-friendly desktop. And there already some distributions focussing on this. I am reminded of Ubuntu and the User Linux project.

An example of a big company that managed to migrate even several years ago is DeBeKa, a German health insurance with (IIRC) between 5000 to 10000 PC desktops at the time. They did not only have to migrate standard stuff like browser and office, they had highly customised database stuff to access plus any number of health-insurance-specific stuff. An interesting case study was published in the iX magazine (German language, sorry). But there are definitely many more such stories, and many are available in English. Google is your friend. (And this is excactly one reason why I consider your article series FUD). Just yesterday I saw an article published at ZDNet saying that Firefox sneaks into the enterprise but especially big companies are wary of publicly admitting this because they fear problems with the big monopolist who yet sells them their desktop software. So why shouldn't this be true for companies that 'dare' to go the migration path on a larger scale?

I really hope that someone will make the effort of replying to your opinions with a fact-based article that provides real hard data. What the FLOSS community should hope for and what migration-aspirants need is not opinion, but facts! And what nobody needs is opinions sold as "Hard Truth". It's the facts, stupid!

Kind regards,
Andreas



Re: Part IV: Corporate Desktop Linux - The Hard Truth (Score: 0)
by Anonymous on Mar 03, 2005 - 04:47 AM
I think Mr. Buck's piece is the way opinions on OSes (especially Linux) should be presented. Level-headed and no unecessary flaming or OS bashing. I use both Linux and Windows at home and at work (I prefer SuSE Linux btw), and I agree with many of his points (especially with his opinion on the attitudes of the average computer users). However, I hope that he is aware of the recent large scale migrations to Linux by corporations and governments.

Byron Joseph Hallar
Philippines


Re: Part IV: Corporate Desktop Linux - The Hard Truth (Score: 0)
by Anonymous on Mar 03, 2005 - 06:31 AM
"The OpenOffice spreadsheet doesn't have to be as good as Excel (and it isn't), it only has to be good enough for 90% of the users."

As a matter of fact, studies have shown that most users of MS Office usually use as little as 10% of the product features, so I would say as long as the Linux/open source equivalent aproaches that 10 % of most usable features, a good number of people will have their requirements met. Why are corporations spending $300+ a seat for features that are rarely used? Companies that are more efficient will be more competitive and more solvent.

For those that absolutely can not find an open source equivalent, why not run their windows o/s and apps on top of VMware or Win4Lin; running virtual machines means that they are no longer risks to the rest of the network and Linux provides a stable, secure foundation.



Re: Part IV: Corporate Desktop Linux - The Hard Truth (Score: 0)
by Anonymous on Mar 03, 2005 - 12:18 PM
Zealots not withstanding, the truth can be uncomfortable. And the fact that many don't want to hear it, won't make it go away.
"benefits accrue only to those willing and able to change the software environment"
Check
"An owner/user unable to change the environment gets no advantage from software adaptability. Unfortunately these are the conditions in the two places comprising the majority of desktops:"
Check.
"The last thing most unskilled or semi-skilled consumers want is to make changes to their computer, or be confronted with technical choices they don't understand"
Check.
"Linux will run on lower end hardware, to be sure. But this argument won't sell many deals."
Check.
"the major benefits of Desktop Linux for non-technical people are better protection from viruses and perhaps a little better stability"
Check
"The hard truth is that the benefits that are most important to individual technical people are simply not important to those lacking technical skills"
Check
."stop fighting a defensive unwinnable war"
Check
When I turn the key, will the engine start? Don't tell me about priming the carb, or pumping the pedal, will it start?
If the answer is not an immediate and emphatic "yes", you've lost a sale.
The user is not as concerned as many believe with OS v hardware upgrades. You can have 5+ year old sound and video cards, printers, scanners and they will all be supported by Win95 through WinXP. If not: click, click and you have a new driver installed. Trying to explain to the Windows user how dangerous that is, from a Unix point of view, will at best, fall on deaf ears, at worst, hopeleessly confuse the issue.
Does anyone really think that Steve at CompUSA wants to explain, or will be able to explain, to Mr. and Mrs. customer how to unzip a tarball? Perhaps give them a rundown regarding the pros and cons of rpm. versus apt-get? When the 15 year old's only concern is to be able to play 'Halo' on the internet and the 13 year old just has to be able to 'instant message', moving a copy of SUSE to this family is going to be a real chore.
Replace the hardware, will be the battle cry.
Why should they?. What they have works. To advance the accolades of Linux?
Not likely for the consumer, will never happen for the business.
It should be kept in mind that the majority of consumers do not follow 7 year old battles with the U.S. Dept. of Justice, would not understand them if they did pay attention, don't have a clue as to what 'Sherman Anti Trust' action is and actually don't know, or care, that Bill runs an Evil Empire.
Ease of use isn't as much of a factor when everything is locked down but even the simplest variant can be trouble.
"My floppy doesn't work"
"No problem. Did any of them work"
"The first one did"
"Oh well, you just need to to unmount the drive before you put in another disk"
"Huh"????
It doesn't matter how simple it is, if it is not as simple as it used to be.
Ease of installation should not even be an issue. The overwhelming majority of Windows users didn't install their OS in the first place.
Neither should cost be an issue; one way or another they have already paid for a copy of Windows. Anything else, no matter how low the cost, is an additional expense.
The inevitable questions will be, "easier than what?" " How come I have to install anything." " What do you mean by install anyway?"
If someone with a new Dell, let's say, wants to make the move, how excited are they going to be when they find out, after deleting their copy of XP, that Linux doesn't support their SATA drive? It's been a standard for a couple of years now. Will Linux natively support SATA by the time IDE drives are no longer sold?
If the response to any of this is,
&qu

Read the rest of this comment... Zealots not withstanding, the truth can be uncomfortable. And the fact that many don't want to hear it, won't make it go away.
"benefits accrue only to those willing and able to change the software environment"
Check
"An owner/user unable to change the environment gets no advantage from software adaptability. Unfortunately these are the conditions in the two places comprising the majority of desktops:"
Check.
"The last thing most unskilled or semi-skilled consumers want is to make changes to their computer, or be confronted with technical choices they don't understand"
Check.
"Linux will run on lower end hardware, to be sure. But this argument won't sell many deals."
Check.
"the major benefits of Desktop Linux for non-technical people are better protection from viruses and perhaps a little better stability"
Check
"The hard truth is that the benefits that are most important to individual technical people are simply not important to those lacking technical skills"
Check
."stop fighting a defensive unwinnable war"
Check
When I turn the key, will the engine start? Don't tell me about priming the carb, or pumping the pedal, will it start?
If the answer is not an immediate and emphatic "yes", you've lost a sale.
The user is not as concerned as many believe with OS v hardware upgrades. You can have 5+ year old sound and video cards, printers, scanners and they will all be supported by Win95 through WinXP. If not: click, click and you have a new driver installed. Trying to explain to the Windows user how dangerous that is, from a Unix point of view, will at best, fall on deaf ears, at worst, hopeleessly confuse the issue.
Does anyone really think that Steve at CompUSA wants to explain, or will be able to explain, to Mr. and Mrs. customer how to unzip a tarball? Perhaps give them a rundown regarding the pros and cons of rpm. versus apt-get? When the 15 year old's only concern is to be able to play 'Halo' on the internet and the 13 year old just has to be able to 'instant message', moving a copy of SUSE to this family is going to be a real chore.
Replace the hardware, will be the battle cry.
Why should they?. What they have works. To advance the accolades of Linux?
Not likely for the consumer, will never happen for the business.
It should be kept in mind that the majority of consumers do not follow 7 year old battles with the U.S. Dept. of Justice, would not understand them if they did pay attention, don't have a clue as to what 'Sherman Anti Trust' action is and actually don't know, or care, that Bill runs an Evil Empire.
Ease of use isn't as much of a factor when everything is locked down but even the simplest variant can be trouble.
"My floppy doesn't work"
"No problem. Did any of them work"
"The first one did"
"Oh well, you just need to to unmount the drive before you put in another disk"
"Huh"????
It doesn't matter how simple it is, if it is not as simple as it used to be.
Ease of installation should not even be an issue. The overwhelming majority of Windows users didn't install their OS in the first place.
Neither should cost be an issue; one way or another they have already paid for a copy of Windows. Anything else, no matter how low the cost, is an additional expense.
The inevitable questions will be, "easier than what?" " How come I have to install anything." " What do you mean by install anyway?"
If someone with a new Dell, let's say, wants to make the move, how excited are they going to be when they find out, after deleting their copy of XP, that Linux doesn't support their SATA drive? It's been a standard for a couple of years now. Will Linux natively support SATA by the time IDE drives are no longer sold?
If the response to any of this is,
"they don't understand,"
"they're not smart enough,"
" they're just fools for using MS",
"Linux isn't a gaming platform",
"when they get past the learning curve they will realize",
that's just so much a spinning of wheels. At the end of the day Linux will still be in the same place: behind MS. Now while none of these statements can be totally denied, they have nothing to do with the end game of moving Linux further along. Faulting the user has never been a strong selling point.
Oh, and before it's asked and the flames are lit, KOffice under Debian 3.0 r4.
Tom



Disagree on the Ease of Use Argument (Score: 0)
by Anonymous on Mar 03, 2005 - 04:15 PM
It's not supported by your own comments, in fact.  You speak of Administrators doing a "lockdown" standardization of configuration and apps, and not allowing customization as canceling out Linux's advantage in customizatiopn ability, and then you argue that the customization is what makes it less easy to use!  You can o­nly use o­ne of those arguments at a time, not both!

For novice or nontechnical users, when there is no need to fiddle with setup and configurations, the Linux desktop is EVERY bit as easy to use as is Windows, or even OS X. 

Overall, it was a well-thought out article, and you made some good points.  I think, though, that your pessimism got the better of you o­n this o­ne.


It's not for everyone (Score: 0)
by Anonymous on Mar 04, 2005 - 10:46 AM
The "average" user has enough trouble dealing with Windows (who's UI hasn't changed much in the past ten years). Put them in front a Linux machine and you can expect a number of them to panic only because it's different from Windows.

Most Linux advocates are much more technically savy than the average Windows computer user. What might seem to be simple and straight-forward to the Linux advocate might be very difficult for the average Windows user.

I'd had to support people who've panicked because they couldn't find Word (they deleted the desktop icon and couldn't be bothered to look for the shortcut in the start menu). I've also had users who couldn't log onto a domain because they couldn't figure-out which username/password to use (even though they've logged onto the domain before.... they just forgot). Or people who were completely lost after upgrading from Office 2000 to Office XP (just imagine them with Open Office). I've also had to deal with people (different people at the same company) who were panicking because their "screen was blank"... they forgot to turn it on and assumed that the computer had crashed.

I had setup a few Linux boxes at the company (firewall, Free/SWAN VPN, proxy) and I thought about migrating a number of users (about 150 or so) to Linux because they didn't need any exotic software, only a web browser for Outlook Web Access (this was because the Evolution Exchange connector was free), a terminal client for a legacy application, and Open Office; we also had four Citrix servers that we could use for anything that wouldn't run under Windows. However, it would have been a absolute support nightmare (and I'd have been the one supporting it), so I ultimately said "Ni".

The biggest problem with Linux on the corporate desktop is the potentially large amount of training that the users might require and the associated support costs (because your support people will have to be Linux savy too). If you know that your users have trouble just turning on their computers,

The biggest potentialy problem users are the Big Cheeses. If management tells the users that "this is the way that it's going to be so deal with it" at least you have some backing. On the other hand, if the management is complaining about not being able to do stuff as they used to, you've got a huge problem.

Then you have to consider the issue with the lack of software. Here's an example:

The VP of Sales was crazy about ACT!, but it doesn't run under Linux. And the lastest version of ACT! doesn't run under a Terminal Server at all and being .NET based you can forget about WINE. So for him and his sales team using Linux would be completely impossible. One option would be to use something like SugarCRM instead (which is a heck of a lot better and more functional), but that would again involve a whole lot of training and a number of political battles and a depressed VP of Sales (you'd think that the guy had shares in Best Software).

In business it's all about costs. In some cases Linux (and more specifically OSS) can offer significant ROI and is definately worth the migration (especially for companies that are in the process of migration from Win'9x... because switching to XP ain't cheap). In other cases you might be looking at a whole lot of short-term (or not-so short-term) trouble and you have to ask yourself if you can ride-out the bumps and hope to see some long-term benefits.

On the server-side is very easy to see the ROI and you see it very, very quicky. On the desktop, it depends...





Advertise With Us! | Comments are property of their posters.

Performance Testing by R17 - sister company of OSDir.com

Copyrighted (c) 2000-2014, but we're happy to let you use what you wish with attribution. OSDir.com
All logos and trademarks are the property of their respective owners.
OSDir is an inevitable website. Contact | Privacy Policy | Terms of Service

Page created in 0.249855 seconds.