I used to skateboard when I was a teenager. This was during the times when Tony Hawk was in his prime…Powell Peralta was the number one skateboard company on the planet, and Thrasher magazine was the number one choice of reading material.
Most of my friends at that time all rode Powell Peralta boards. The thing is…I was always looking for an advantage…something that could give me a competitive street skating advantage or something that just plainly worked better.
I found that advantage in H-Street equipment. I began riding a naked H-Street board with H-Street Arrow wheels. Switched from tracker trucks to independent and changed my bearings from German to Swiss. I watched Hokus Pokus and idolized Danny Way. I was ridiculed. I was told that I didn’t know what I was doing. I was told that H-Street was no Powell Peralta. A year later, everyone had a Hokus Pokus poster on their wall and were trying to get the gear and equipment I had already purchased.
I’m not saying I’m a trend setter. I’m saying I recognized quality and functionality before most did. Many businesses today are exactly like my friends. They don’t want to change. They don’t recognize quality or something that can give them a competitve advantage (at least not until its too late in most cases).
Why is this? Why is it that many corporations and small to medium businesses cannot or will not take a step back and look at the competitive advantage and cost savings Linux and Open Source software will give their business?
Bureaucracy
Many large corporations are loaded with managers. Every section has a manager. A group of 2-3 sections has a manager. That manager has a manager. That manager also has a Deputy Director. That Deputy Director has an Executive Director, who in turn has a CIO/CTO who in turn has a CEO above them. Like an onion, so many layers it makes you want to cry.
I worked in a Fortune 500 company as a project manager and saw bureaucracy at work. In order to get anything accomplished, you have to portray a clear and present danger to how they operate with downtime or a large monetary savings/loss. Then and only then would your manager listen to you. After your manager listened, you had to wait some time before that information trickled up to the level of manager it needed to in order to have action take place.
There were times when half a million dollars wasted away while I waited on a manager to make a decision. Inefficiency eat your heart out.
As for Linux and a business like this…getting them to see the savings isn’t hard. They know if they didn’t have to buy antivirus for over 6 thousand desktops they’d save tons of money. They know that if they didn’t have to license the same number of office installs that they’d save an enormous amount of cash. This isn’t the problem. The problem is they don’t believe in the product. They don’t believe in the product because they haven’t used it before. It’s new. Anything new is a risk.
While you may get your manager to see these savings, getting his manager and the manager above them to see similar savings is like pulling teeth from a crocodile…dangerous and difficult. Change in organizations like this start from the top down. It’s hard for open source developers to swallow this because change for us, starts from the bottom up.
Equating Value with Cost, Price with Quality
Most businesses equate value with cost. Something is valuable to them if it costs them money. Linux of course, is free. So how can it be valuable to them? If it doesn’t cost them money…then it must be lower quality right? The open source model of software does not fit turnkey into the standard business model and defies the way analysts measure things.
How to solve this dilemma? Your guess is as good as mine. No one can force businesses to change their mind when it comes to value…it comes with time. There are some heads being turned by various open source programs such as OpenOffice and Firefox…but the tide is slow to roll.
Is it faulty to equate value with cost? I believe so. Value should be a quantifiable measurement and not a monetary one.
How can you help? Make open source valuable to your computing. Make open source valuable to the corporation where you work. Get the open source/linux foot in your door and let your managers and their managers see the value.
Change is Bad
Changing is always equated as costing money. This is due to the age old premise in project management that change midway through a project will cause scope creep and add cost/resource expenditure. For the most part, with projects, this is true. For many open source projects and programs however, it isn’t.
Take Firefox for example…it is a browser that has more control and configurability than Internet Explorer and is more secure than Internet Explorer. It has captured over a third of the browser market…nearing 2/5ths of the market. So why aren’t businesses embracing it?
They fear change.
This despite the fact that a browser is a browser. This despite the fact that it can be made to look exactly like IE. This despite the fact that my fifty something mother-in-law can use it with such ease that when she switched from IE it took her 30 minutes to adapt.
Embracing Change
Linux on the desktop is one of the areas where business could harness such power. Saving anti-virus cost each year. Using the power of user/group management. Locking users down with ONE setting instead of having to lock users at multiple virtual layers. The list of benefits far outweighs (for many businesses…there are always exceptions to the rule) the leading edge time spent ramping up.
The time will come soon where managers who are in charge at all levels of big business are people who have never experienced life without a computer. They’ll be more savvy with technology. They’ll listen to ipods, use itunes, scrobble music, use bittorrent, plugin to facebook/plaxo/myspace. These managers are in the wings now and slowly making their way up corporate ladders. The older managers who can remember life without computers or with early computers are still holding on.
The change and fear is still there…the tide will slowly turn.
These new managers won’t fear change such as the one we mentioned above with Firefox. These new managers understand the technology and how long it takes people to adapt to it. To them, change isn’t a bad thing…it’s a good thing. The minimalistic time spent ramping up is counter to the amount of benefits you see returned by the switch.
These new managers recognize value in new things…they’re looking for an edge…just like I was with skateboarding. They’re not afraid to try something new. While it won’t happen over night, the ball is rolling and in the next 2-4 years the face of computing will change to embrace this.
I think one of the other factors in the lack of adoption of alternative desktops in the business world is what boils down to laziness. Many people in IT or Tech Support or whatever it’s called in any given company learned how to do things with one system and do not want to take the time to learn another way to do things.
Seriously.
Actual conversation (and I realize it’s Mac, not Linux being discussed, but you could just as easily swap “Mac” for any other non-Windows system):
“There will never be any Macs in this building.”
“Why not?”
“Because I don’t know how to use them, and I don’t feel like taking the time to learn how.”
My first reaction at seeing older managers blamed for this was minor outrage. I’m in my mid-40s and I felt unfairly tarred with your brush of age.
But that passed, and I have to admit you are on to something. I code. I run Linux and OpenOffice and Firefox and nothing from Microsoft. I use MySpace and torrents and pretty much live my life online. But I know that most people my age don’t.
I hear friends say, Oh, I don’t know anything about that computer stuff, how do you know all this? And in my mind I’m thinking, how come you don’t? But I run into far too many people my age who can’t even type with more than two fingers. So unfortunately you are right, these are not the people who will make bold informed decisions about open source adoption.
I just hope the transition can be accomplished without throwing away people like me just because we’re old. Not everyone over 30 is a technophobe. Let’s try to remember that, shall we? 🙂
When I say “older” I also mean younger managers with the “keep things as they are” mindset. So you can be young and still be a part of this older way of managing.
Thanks for reading 😛
Unfortunately, there are more barriers to enterprise Open Source adoption than this article suggests.
For example, many firm drank the MS Kool-Aid and developed internal ActiveX applications. It was a bad idea when they did it, it’s a worse idea now, but there it is.
Entrenched mindset is only one of the barriers Open Source faces in conquering the desktop, (or should I say Konqueroring the desktop?) That mindset is rapidly changing even among the most senior of management. Entrenched systems are a stickier obstacle.
This reluctance to change is not just for Business – everyone have it. Its called social inertia.
Interoperability of Desktop Apps. It is essential that businesses have applications that are 100% interoperable with fellow businesses and government departments. You might say “but what about Open Office?”. This is not 100% interoperable – sometimes because of the fonts (causing different page layouts), sometimes because some government department geek guy has created a Word “form” with a bunch of macros, sometimes for other reasons. What you see in Microsoft Word is not exactly what you get in OpenOffice. This makes document exchange a royal pain.
If OpenOffice and Microsoft Office were 100% interoperable, and other Linux apps improvements were made (e.g. a groupware app that matches Outlook and interoperable with regard to meeting invitations, etc), I think you would see a wholesale movement to Linux.
Businesses don’t care about operating systems – they just want a no fuss, will work, will interoperate solution. And they are happy to pay for it. What’s the point in saving a few hundred dollars if I can’t use the Tax Office forms and on-line e-Tax apps? Or if every document I get from other businesses has to be converted and re-formatted to make it look the same, or if I can’t just email a document to a contractor and have them see what I see?
Interoperatibility at application level. If this is achievable, businesses would not care if the OS is Windows, Linux, MacOS, or a bucket of monkeys.
IMHO.
The “commercial software bad” mantra adopted by most Linux and OSS types is hard if not impossible to swallow by any business manager, regardless of age. Business is about making money, not about idealism.
I think businesses will open themselves more to Linux when the Linux community opens itself to commercial, non-OSS applications.
Software is certainly not alone where management tends to toss new ideas into their mental Oubliettes. I recently told 2 managers how we could save a conservative $4000/year on fuel costs, and reduce the man-hours in handling fuel by 60% by investing less than $1000 in a single piece of equipment.
Manager #1 response: “My people don’t have enough time to make it work.”
What part of reducing m/h by 60% did he not understand?
Manager #2 response: “We don’t have the money in the budget to buy the equipment in your proposal.”
Despite the fact that it would pay for itself 4x in the first year of use?
No wonder their eyes glaze over when I attempt to explain to them the benefits of migrating to Linux! Remember that as far as they’re concerned, a surplus in this year’s budget usually means their budget for next year will proportionately cut, and NO manager wants LESS money to work with.
My guess is that the main reason is not laizyness, but information lock in. If you switch to another OS you need to make sure that your existing information remains accessible.
Sure there is OpenOffice and some other apps that do a decent job of MS word and excel, but what about MS Access, what about VB script and othr in house applications.
This article should be titled “Why Linux Doesn’t “get” Business”. The issue is that an enterprise has hundreds or thousands of line of business applications that must work for their business to run. All these apps would need to be tested and packaged for Linux and it is a GOOD bet that a large percentage will not work. Thus alternatives will need to be found, there may be recoding involved and the costs start to mount up. This is exactly the reason why Vista is meeting so much resistance in the enterprise.
Add to this the question of how the desktops will be managed, apps deployed, etc. There are entire systems developed for this such as System Center, Altiris, etc. that are very advanced systems for this. They go a long way beyond “Ghost”.
These are two obvious reasons, there are many more and are the same issues that Macs face. Pinning this on bad management, laziness, etc. is just plain ignorance of business and is the real reason why Linux continues to fail as a desktop system within the enterprise.
There is more to this issue than anti-virus, fear of changes, etc. There are serious interoperation issues.
How often do you need to do an upgrade on a Linux desktop to keep up with Firefox, Evolution, and other GTK/Gnome programs? I can install a Windows Vista desktop and run it for 3 years with just patching without worrying that Microsoft Office 2010 or Firefox 4 or a large amount of other GUI programs not working.
You don’t get that with Linux. Take a 3 year old Linux desktop and try using any newer software without replacing large chunks of your operating system.
How much value is it to a business to rebuild Linux desktops every year in an attempt to ensure that newer software will work?
I don’t mean to sound nasty about this. I’m an UNIX Desktop/Server Admin. I love my job, and I wouldn’t want to give it up. When Linux can go three years with the same level support that I can provide a Mac/Windows box. Then I’ll state it is ready for normal business use.
Granted, I work in an environment where most Linux desktops are for software developers and testing folks, and most of those folks are savvy and are worse than most Windows power users. 🙂
– Ben
I don’t think you can pin the failure (to date) of serious Linux adoption on the desktop to any one thing, but there is certainly more than a grain of truth in what you are writing. I know the same frustration at how people are stuck in their ways and know it and don’t care.
I work in IT at a municipal government; I have run the Linux migration scenario in my mind 1000 times, and every time I come to the same conclusion: it can be done, but it would take everyone being on board — IT, department heads, elected officials, down to the rank-and-file users. A lot of custom-written software would have to be phased out, and a whole lot of people would have to adapt to a new office suite.
The key is to know in your own mind “why?” Why is it worth it? Why does bob the spreadsheet geek have to give up Excel? Why does Karen the power user have to relearn an OS? We know there are sound business reasons that go well beyond “Linux is cool” or even “Linux is free (as in beer)”. How about having a platform that is not tied to a single vendor? How about lowering the administrative overhead of tracking licenses and auditing software (a major chunk of my job)? Support for older hardware… more flexible configuration possibilities… the list goes on.
I’ve read the bullet-points of a boatload of naysayers, and they all demand the same thing: a drop-in replacement for what they already have. But why? Open source offers so many more possibilities! It’s like saying “We don’t want to build with bricks, because we only know how to build with wood. When brick acts just like wood, we’ll build with it.”
Take a 3 year old Linux desktop and try using any newer software without replacing large chunks of your operating system.
Answer: Are you really a Unix Sys Admin, come on be serious now? Here is my reply to your stupid comment. Take a 3 year old system and try running Vista or even XP. At least old software on Linux doesn’t degrade/deteriorate performance. As far as the Desktop is concerned there are several package managers that automatically upgrade your packages for you and everything related to it, its called “resolving dependencies” genius and on several leading distros it is now automatic.
How much value is it to a business to rebuild Linux desktops every year in an attempt to ensure that newer software will work?
Answer: Thats the whole point of Linux, you don’t have to rebuild it if you are not using any feature supported by newer kernels, just keep patching it. There are absolutely no associated costs. With the exception of security patches the higher numbered kernels have more hardware support. If you use hardware that can be supported by older kernels you can stick with the same old kernel for years. The newer kernels basically have better hardware support thats all. Why use it when your hardware doesn’t need it. I am still using the 2.4 kernel on my Slackware machine.
When Linux can go three years with the same level support that I can provide a Mac/Windows box. Then I’ll state it is ready for normal business use.
Answer: Linux can go way more than 3 years with the same level support as a Mac. As far as Windoze is concerned don’t kid yourself, MS gets more support calls than you can even begin to fathom. I am surprised you call yourself a Sys Admin. My old grandma can use Linux without any support from me and it has been the case for the past 2.5 years. She had more problems with XP. She uses Ubuntu (since we are talking of Desktops), we already know Linux reigns supreme in the Server market.
You ask why.
There is a simple answer.
Early adopters go with new things; whatever the new thing is.
Then comes the pragmatic adopter.
Once they have adopted, the others with follow.
Windows falls into the last group. Its momentum reached the ‘follower’ group.
Linux is still in early adopter state, maybe starting with the pragmatic adopter.
Again, people, business are not stupid and they certainly understand money. I am certain that if you were able to show, say, a 15,000 seat company an ROI analysis that demonstrated a positive value in moving to Linux they would jump at the project if the analysis was believable and demonstrated hard dollar cost savings.
You would need to take into account probably 12,000 desktops and maybe 3,000 laptops, probably at least 600 applications, tens of thousands of documents (maybe hundreds of thousands), training for end users, training for administrative staff. You’ll probably need to recode some custom applications and more than one macro. Better take some logitics into account as well in terms of how that new Linux image and new apps will be deployed. I think an IRR of 18% is reasonable for use in your calculations but you should check with your finance department for what you use internally.
Now, this is just barely scratching the surface of what would need to be considered in a Windows to Linux migration. I would peg the cost between 4-6 million dollars as a blatant SWAG. And you’re saving, what? 1.5-2 million dollars a year in license costs? Not good my friend.
Note, if you don’t know the acronym IRR without looking it up then this, again, demonstrates ignorance of business and what businesses really care about when they consider technology projects.
Believe me Seth…I’ve BEEN one of those managers (for a Fortune 500 company with over 20 thousand employees). I had to glaze my own eyes over or I would have been fired…it’s one of the reasons why I left…I couldn’t stand to see profits that could have been reaped pissed away.
I was responsible for around 12,500 desktops and close to 300 laptops. We ran Windows 2000 even though it was 2005 and had to certify all the software and my department did all of the image prep and software/hardware installation.
I know the cost of migration and it wouldn’t be pretty. The thing I do understand though is you don’t have to migrate the entire shop. Allow me to explain…
At any given location of the business I worked for above, there was 60-200 PC’s. If I cut those PC’s down to 2-3 in each section of business, I would reduce the # of PC’s using Windows to around 15-20 per store. Now, the other PCs would be used to check email and access the web, perhaps open a webform/image file. That was the extent of what those PCs were being used for…yet we had to have antivirus/antispam/antimalware software running on them. We also bought Office XP for each computer.
If we had run Linux on all but those 15-20 PC’s (we used Samba for shared drives already on HP-UX servers) we would have had major cost savings…there’s no getting around it. Most of our employees checked email, did limited surfing, and saved stuff to their shared drive. That’s it.
So while not everyone who uses Linux “gets” business…I do. I don’t throw out numbers because each and every shop is different (my example above is just one) and each has different size. When I say “business” above, I’m taking into account the large and the small.
Please remember, it’s not just license cost for the OS, but for antivirus, office software that can be replaced, and other software that may or may not have replacements that are freely available. That ups the savings as well.
It’s also the ability to not have to renew licenses every 3-5 years because you’re locked into a single vendor.
It’s obvious to me that you think I am ignorant to how business works or that I am someone (who doesn’t know what IRR is) who has never managed a multi million dollar project before. It’s sad that you approached this article with that in mind…mainly because you’re wrong 🙁 Throwing an acronym in that the general public wouldn’t know doesn’t mean you know what you’re talking 😉
I didn’t approach the article with any particular point of view in mind. After reading the article I formed an opinion based on the fact that no real effort was expended on business financials, just blame, blame, blame on things that make no difference at all.
Certainly every business is different, but if I am in a business office and looking at your proposal, I still need a solid ROI or CDB analysis because what you are proposing is that I am now managing two images instead of one and thus I am exposed to twice as many bugs and security vulnerabilities and thus have twice the patch management issues, driver management issues, application and OS compatibility issues, etc. That’s not even mentioning how I deploy apps remotely to those systems or back those systems up remotely, etc. Again, in that scenario, now I need two of everything and that is expensive, probably more than enough to offset any cost savings.
Rolling out a Linux desktop side-by-side with a Windows desktop is going to be a hard sell. You have doubled all of the management and this is going to be a hard pill to swallow for a modern IT shop that has their image management down to one or two images (a 32-bit and 64-bit let’s say). All of their desktop management, patching, app deployment, desktop engineering (security and other settings), backups, etc. are all geared toward their current platform and it is expensive to move.
Were you able to put together such an ROI analysis and they still didn’t listen to you? I think that if you did one, that would make a very interesting post for discussion.
But, pegging this on bureaucracy, laziness, resistance to change, inability to understand “free” software, etc. is just patently misleading and demonstrates a grossly warped view of business. It doesn’t do anybody any good and leaving out any semblance of actual fiscal considerations makes your article look amateurish.
Sorry.
Honestly Seth, this is my personal blog. I don’t publish ROI studies here. Nor will you find financials for major projects or rollout project dates, conversion rates, migration costs, etc here either. What you will find is what I feel, think, and believe.
If my thoughts and feelings on a subject are amateurish there is nothing I can do :/
[stoobie] >> Remember that as far as they’re concerned, a surplus in this year’s budget usually means their budget for next year will proportionately cut, and NO manager wants LESS money to work with.
Where is the problem? Oh, you forgot to add into the proposal a way to spend the saved funds.
You can go for the merely wasteful. Or you can even give it some thought and find something that can take up the saved hours and dollars that will simultaneously add value to the business or to the manager. Maybe version 2 of your proposal will be successful AND get you a promotion.
[Gregory Smith] >> Interoperatibility at application level. If this is achievable, businesses would not care if the OS is Windows, Linux, MacOS, or a bucket of monkeys.
OK, since Microsoft is not likely to open up their source in the near future, the option presented is to continue along the same path, becoming ever more dependent on a single supplier, requiring even more pain to leave them in the future as the competition continues to get better. Or to start working on turning things around the earlier the better. For example, move to technologies that are also supported on Linux.
Some businesses will react in time and some won’t. That’s life in a capitalist society.
[Seth Moupre] >> These are two obvious reasons, there are many more and are the same issues that Macs face. Pinning this on bad management, laziness, etc. is just plain ignorance of business and is the real reason why Linux continues to fail as a desktop system within the enterprise.
So are the businesses sinking themselves into a greater “escape liability” or are they moving towards technology supported on Linux? Are they aware of their options? Do they care to know?
For example, who in their right mind, wanting to regain leverage on Microsoft and put a stop to the growth of their liabilities, would start using silverlight or OOXML today?
Companies can take steps by converting documents into PDF for interchange for example.
It goes to show the lack of power companies frequently have against Microsoft that Microsoft so thoroughly controls their lock-in and upgrade cycle (“what silverlight? oh, well, if Microsoft says we have to use it then we have to use it”).
Sometimes people know the inevitable, but it’s easier to lie in the gutter a little longer and wait than to start getting up and cleaning yourself up. Humans require a certain amount of pain to put themselves into a situation of change. The problem is that easying into that pain means you can lose sight of what life with a lot less pain can be like, of what gains can be achieved when you are significantly less encumbered.
[Ben Lindstrom] >> You don’t get that with Linux. Take a 3 year old Linux desktop and try using any newer software without replacing large chunks of your operating system.
You should try asking Red Hat to hold your hand. They hold a lot of people’s hands and that’s nothing to be ashamed of. That’s what a service contract is for.
By the way, doesn’t Microsoft replace chunks of their operating system very frequently (and you can’t even tell what they are doing)? Are you telling me that SP1 for any of their operating systems does anything less than a huge replacement of many key system components?
>> Now, this is just barely scratching the surface of what would need to be considered in a Windows to Linux migration. I would peg the cost between 4-6 million dollars as a blatant SWAG. And you’re saving, what? 1.5-2 million dollars a year in license costs? Not good my friend.
There are many costs not included in your calculation as you stated (depend on too many factors to generalize very much).
Whatever your IRR has been for the last few years, it likely will go down in the near future (at least this is probably going to be the case on average if we go through a broad economic down swing).
There are many things that cannot be quantified or known prior to doing pilot tests. With very little that is static in an enterprise, opportunities to introduce Linux abound.
>> Note, if you don’t know the acronym IRR without looking it up then this, again, demonstrates ignorance of business and what businesses really care about when they consider technology projects.
I had to look it up, but I am more than aware of the concept. That you think a person has to be aware of the jargon of a particular field of study to understand its motivations and math shows a little short-sightedness on your part. The same concepts and formulas make their way from mathematics to science to economics/finance.
So with IRR in mind (ie, the time value of money), why would you not put a premium on periodic Linux pilots? If the savings at some point in time were found to be significant (and there is evidence to suggest this could be the case depending on the particulars), delaying by months and years will be very costly as the potential savings fail to accrue.
And there is also the matter of opportunity costs that are hard to quantify when you are very inexperienced with a particular set of options.
Also, the focus should not be on migrations but on containment. Every new cycle of Microsoft (or any company’s) closed source product upgrades increases the break-away liabilities. Are you ignoring this liability (ie, are you assuming Microsoft will be around forever)? Costs that will be eating up company funds of some companies in the future will be nonissues for others that have moved to a free/open platform and have undergone the training and are excelling having adopted open source practices.
Of course, those that are pressed to move to Linux and are moving usually deal in very competitive and rapidly changing environments where the competition is leveraging Linux already. [Don’t mean to imply Linux is right for everyone in such an environment since there are many variables to consider.]
[Sean Kerns] >> Answer: Are you really a Unix Sys Admin, come on be serious now? Here is my reply to your stupid comment. Take a 3 year old system and try running Vista or even XP.
15+ years doing server and desktop unix adminstration along with software development and deployment. And I have multiple 3 year old Windows XP boxes. Not only at my home, but also a lot of friends of mine. My PC gaming rig built with the idea that it will survive a 3 year gaming cycle with little to no hardware upgrades, and as close to zero OS reinstalls as possible (yes, dead harddrives are not included in this count =). So please, don’t give me bullshit that a “3 year old system can’t run XP well.” I can point you to an OLD Dell Laptop my mom has (going on 4+ years and I’m damn scared the thing is going to die leaving) XP and it has been reinstalled once due to some bad “shareware” software.
My last gaming box (AMD Athlon XP build) runs Windows perfectly fine. We have a lot of Dell 2[67]0 deployed at work that are 2+ years old running Windows and people are happy (no reinstalls). We have a lot of 2[67]0 and 740 Dells and they have been PXE installed at least every 18 months to step up to new OS releases due to the large GNOME changes because large chunk of our user base community want newer GTK and other GNOME libraries just to run newer software that they claim will make them more productive. Something we never hear from our Windows community.
So seriously, think about this. This isn’t coming from someone who “hate linux”. This is coming from real world experience. I’m not complaining about servers here. Most of my life I’ve spent on a 3 year cycle for OS updates (Solaris, OpenBSD, or Linux) excluding hardware replacements.
[Jose_X]>> You should try asking Red Hat to hold your hand. They hold a lot of people’s hands and that’s nothing to be ashamed of. That’s what a service contract is for.
Wow? So you are telling me that Redhat will ensure that the current version of Evolution will run on RHEL2 acting as a desktop? Hot damn.. Where do I’ll sign up.
Come on. I’m not talking about “Linux doesn’t support X” or “My print deamon is unstable.” I’m talking about end-users that are tech-savvy. They run Linux at home or they are power Windows users at home. They are willing to reinstall their computer at a drop of a hat to get the latest release. Then come to work and whine that they can’t get Firefox 3 beta 4 to compile on their work Linux box, and wants IT to install/compile/etc the right GTK/Gnome bits so they can.
I’ve had Redhat support contracts at past companies. If the application that is shipped with the release doesn’t work. They will fix it, but that’s it. No new version.
[Jose_X] >>By the way, doesn’t Microsoft replace chunks of their operating system very frequently (and you can’t even tell what they are doing)? Are you telling me that SP1 for any of their operating systems does anything less than a huge replacement of many key system components?
Yes, however rarely does Microsoft decide to throw away even a single WinAPI without giving their developers decades to ensure their apps don’t use it. I’m assured that as long as Adobe decides that WinXP support is profitable I can get any Adobe product. I can’t say the same about Linux.
The point here isn’t that “Linux sucks”.. The point is there isn’t enough stability in programming API for gnome and other libraries that are HEAVILY used to develop desktop software.
Can you image if GNU Libc with every point release changed memory application API so it took different params or worse yet decide to dump the “*alloc()” naming style for “*GetMem()” out of the blue with no backwards compatibility. This happens with every release of GNOME.
Again, this isn’t a Server discussion for me. I still have Suse 8 servers running happily with security patches. This about making end-users happy with their desktop experience while not killing your admins. Something you can do well with Windows and Mac, but that isn’t as easy under Linux.
[Ben Lindstrom] >> Wow? So you are telling me that Redhat will ensure … They run Linux at home or they are power Windows users at home. They are willing to reinstall their computer at a drop of a hat to get the latest release. Then come to work and whine that they can’t get Firefox 3 beta 4 to compile on their work Linux box
I have a better idea now of what set of circumstances you are describing, but I think you are exaggerating a bit or at least picked a bad example. There also seem to be some inconsistencies.
Why would you complain of users not being able to easily run “Firefox 3 beta 4” on their work machines when they can easily run something *just* less bleeding edge (and maybe even that one but I haven’t checked)? Where can you get the equivalent bleeding IE app? You can’t because it isn’t open source. Compare apples to apples. A firefox that is on par with what a commercial company like Microsoft releases works fine. Tell your users they will have to occasionally settle for less than the most bleeding edge open source stuff. They’d get no better on Windows (tell them that).
Of course, if you want to run a very old desktop system or one with weak backing from repositories I am not sure what to say. Why would you not keep the system updated/modern if it is not a critical system? I mean, if you are willing to let people run “Firefox 3 beta 4″ on it, you can’t be that worried about keeping it up to date (and risk slight short term instability every once in a while.. though I’d still test a little before pushing the updates through to everyone).
I think part of the problem is that you are picking on Red Hat, which doesn’t specialize in bleeding edge (or desktop) software support (you took my choice but I didn’t know what you were hinting at). If you want to allow bleeding edge Firefox, which is already a risk, then risk going with a distro that is set up to handle that even if you find you can’t get anything except free community support (but I’d bet that Canonical would be willing help you with support for a modern desktop Ubuntu). Desktops are inherently a bit less stable than servers no matter who is the vendor. That seems to be acceptable.
>> Yes, however rarely does Microsoft decide to throw away even a single WinAPI without giving their developers decades to ensure their apps don’t use it.
My earlier response was to your mentioning of how Linux system requires heavy changes sometimes in order to install new apps. Again, I did not realize what you were getting at. It seemed to me you were talking about how new apps might require new functionality. That is no different than with Windows except that, with Linux and automatic package updates, adding a new application or app version updates the entire system (bringing it completely up to date if you want). It would be like having Windows 98 turn into Windows XP for no charge and transparently to the user.
However, now that I see you are talking about losing old API, I have to say that it must have been the bugs (Vista) or perhaps the targetted strikes aimed at key competitors and other similar anomalies that led me to think that Microsoft does in fact toss API out the door with upgrades. Surely you have heard of how many applications failed to work in Vista.
To get back to Linux, chaging API is not much of an issue when the system is upgraded for you for free. Microsoft doesn’t provide those free upgrades as Linux does (eg, from Windows 2K to WinXP) so they better keep backwards compatibility at least to some extent (to the extent they actually want your old apps to continue working.. which they sometimes don’t seem to want to happen as it forces many users to upgrade and pay again).
Maintaining systems up to date is a strong point of Linux.
>> The point here isn’t that “Linux sucks”.. The point is there isn’t enough stability in programming API for gnome and other libraries that are HEAVILY used to develop desktop software.
Closed source apps have to deal with this issue to some extent, but as mentioned above, distro packaging of open source software makes updating a distro painless (use Debian for example). And closed source vendors to manage this. Look at Nvidia with their closed source drivers that work across many versions of Linux because they recompile the glue binding while keeping the binary blob mostly in tact (or moving forward).
By the way, you can run simultaneously old and new libraries. This is done all the time. Firefox binaries come with many libraries in its self contained binary package just in case you don’t have them on the system. Certainly, they don’t come with everything, but they do include a lot specifically to allow users to try beta versions of software without affecting the existing system very much at all.
So to recap, Windows users keep old systems because they only have a license for the old version yet want newer apps in cases where these happen to be compatible with their older system (many apps are not compatible with very old Windows systems, as they sometimes require new functionality not available in some old systems and not provided with the app). Linux users, OTOH, can forever maintain a modern up to date system if they so choose. If they want to keep an old version of their system to run an app that just doesn’t work on new systems (this is necessary sometimes and is what you were refering to) then Linux is free and disk space is cheap. It is easy to give this to the users as needed.
Have you considered thin client? The user logging in can pick what platform they want (or done automatically based on what app they need). It’s easier and safer to maintain files centralized than on every machine scattered throughout the company. For the occasional app that requires an old state, virtualization can solve that problem, too. In fact, I sometimes run multiple Linux distros at once through virtualization. It’s an easy way to get access to something that just doesn’t work on your current setup for whatever reason (it’s easy on users once admins set things up, but since Linux is licensed no charge, it’s not a problem to just keep lying around all necessary snapshots).
Not sure why my reply to you yesterday did not go through. I’ll reread it later to see if I said something I should not have said.
The main points I had made were:
— Linux can be maintained up to date usually easily if you rely on a good desktop distro (ie, w/ good repos). Red Hat’s strength is on the servers and stability so you may want to opt for something closer to Ubuntu for desktops (currently I use PCLOS). Relying on debian repositories gives you access to almost anything you would want. And you can have many different versions of libraries without conflicts. Firefox binary comes with many libraries it uses just in case you don’t have them. They unpack to their own directory iirc.
— Windows users don’t get free system upgrades. You dont’ have Win98 evolve into WinXP; thus, Microsoft has to go to more pains to keep old garbage around. The Linux world trims fat as it goes because it’s open source so the community (eg, debian/ubuntu/fedora..) allows you to use new software and keep the entire platform up to date (for no charge). Ie, OldLinux becomes NewLinux so it doesn’t matter if API’s are a bit more unstable. It costs you nothing to upgrade. [See next point if you find you need to keep an unmaintained app around.]
— With virtualization, thin clients, $0 licensing, etc, you can easily maintain many older or just different snapshots of Linux in case you need an old setup to run a particular app.
— Microsoft breaks compatibility unlike what you claimed. They may call it bugs, but companies have gone out of business because of MS shananigans. And look at what a horrible job Vista did maintaining the old API. Yeah, Microsoft may still “officially” support something, but when the app breaks, who cares about the “official” lie. Call it a bug. They use their control to their advantage well. It’s all part of the upgrade treadmill and their monopoly abuse (documented in court extensively). Linux upgrades cost nothing and happen transparently and frequently. Windows upgrades (eg 98->XP) cost money and require manual reinstalls of many items.
You might be using a nonideal distribution/repo if you are having problems with gnome or whatever. Also, consider KDE/Xfce/etc for the desktop. And thin client computing can simplify your life while cutting down on walking hours. This means you can do more interesting things with your time.
OK, too many thoughts going on to keep track of but if you have never heard of IRR then you have apparently never put together a business case for a technology project. It is difficult to avoid tripping over that acronym at some point when talking to CFO’s, CIO’s and the business office in general.
Microsoft has been around for over 30 years. Linux has been around for less than 20 and how many flavors of Linux have come and gone during that time? I’m not following your logic there. So, because I fear that something might go away, it makes more sense to go with the less proven and more unstable system? Huh? Are you assuming Linux will be around forever?
I don’t get the feeling that you have a true handle on the total costs of ownership related to desktop technologies and are viewing things from a very narrow, single desktop perspective. In addition to all the other issues that have been cited, Linux resources are scarcer and cost more, there is more support involved, etc. If you have to have a third-party “hold your hand” then you are already paying a significant portion of the same cost of Microsoft systems. For example, the cost of a A Red Hat server license and a Microsoft server license are not all that different.
People said the same thing you are saying about vendor lock-in with IBM 30 years ago. It wasn’t true then, isn’t true today, it is just FUD.
You, of course, have citations for your claims of who is moving to Linux?
"For example the cost of a Red Hat Server License and a Microsoft Server License are not that different"
Actually, they are….Red Hat Licenses are free…you pay for the support licenses/updates. That's the GPL at work for you.
As for a less proven and more unstable system…I'm afraid you're completely off base there. Linux IS stable. There is a reason that science runs a Linux based grid system for important research projects (lookup Nimbus). The LHC runs Linux…all of these choose it because of the open source nature of it AND because it is stable.
As for citations, I don't need them…as I said earlier, this is my personal blog…my feelings and opinions on things. But if you want to see a list of linux conversions…the only ones I know of are tracked at Lxer.com: http://lxer.com/module/db/viewby.php?uid=108&…
BTW, I have heard of IRR…remember, I was a project manager (1 of 3) for a fortune 500 company…all I dealt with were desktops…software and images. When something new came down the pipeline…it was me that tested and deployed it. It was me that created the budget and timeline for it. I feel uniquely qualified among my peers in Linux to voice an opinion on subjects like this. Sorry if I didn't cover enough areas for you to have a nice tidy package you want on TCO and ROI…like I said, my own opinion and blog…not a white paper or research paper.
There’s another reason, and this one is far more scathing and vitriolic towards many people in IT. I come across it everyday. But let me explain.
The last company i was at almost fired me for writing a blog on my own time, using my own resources, and never mentioning who i worked for publicly, all because i criticized some pointless litigation that microsoft was threatening linux companies with. This company was also a Fortune 500. I brought in a Linux Journal once to read during my breaks, and I got looked at is I somehow managed to kill a resurrected Jesus Christ while drawing cartoons of Mohamed. Linux there, or negative opinions about microsoft from a consumer standpoint (which i never aired in the workplace mind you), it wasn’t about bureaucracy, it wasn’t about fear of change, it was straight up politics. You sold the Microsoft line. And every employee there was subtley encouraged to sing the praises of Microsoft. Which hell, MS puts out some damn fine products. I just think there’s better tools for certain jobs. I’m a “best tool for the job” kind of guy, in terms of quality, price, security, etc.. And sometimes Microsoft wins, but hey.. they didn’t want to hear that. Microsoft is god over there. Try almost losing your job over a personal opinion in your private life and you can see how deep MS’s claws are in some companies.
Now i’m at a smaller, private company looking to go public. We utilize all kinds of technology here. Unix, Solaris, Linux (ubuntu, gentoo, debian), FreeBSD, XP, Vista, OS X, AS400, you name it, we got it.
And while the opinions about technology and “the best tool for the job” is around and it’s something i adore, there is an older generation here, not too terribly older than me, but old enough. And they’ve built their careers on knowing the Microsoft infrastructure inside and out. They have certifications, they have worked their way up to manage teams that are windows centric.. and their livelihood depends on it. If i was in their positions, I’d be going on and on about IE7 and Windows Server 2008, and Exchange 2007 all day too. Learning updated versions of familiar tools is far easier than learning another platform and another way of doing things all together. And these guy influence their managers who, as they grow in their careers up the food chain, become more and more detached from the technology coming out today. So they push MS and they lean towards MS, b/c that’s what they built their careers on, that’s what they know, and the food on their table depends on their skill set being valuable. SO of course the MS guys at the place i’m at now aren’t going to give “too” much credit to Linux. That being said, a talk with an MS guy here is far different than a talk with an MS guy at the previous company. They’ll give things a chance at least and they’ll entertain open discussion. At least in this atmosphere, they’re human.
I’ve worked at 2 enterprise class companies where desktop Linux is valued and deployed: Red Hat and now Google. My job involved (and continues to involve) the support of desktop Linux to lots of employees.
It’s a thought provoking article. Fundamentally the issues raised can be extrapolated to companies with management structures that allow their managers to operate in insular, truly unimaginative and old-fashioned ways. This is a generic problem. If they do not see the value in open source and Linux then they are likely to also not see the value of inspired, modern technical development, business and management practices.
A book I’m reading about “wikinomics” spells this out perfectly…
Great article – I’ll prioritise this RSS feed more now 😉