One factor likely at play here is the low cost of licensing MS products under the enterprise/volume licensing agreements. Where I'm at it's costing well under $100/person/year to license Windows, Office, Exchange, SharePoint and a few other odds & ends. IIRC, for just under $50/person/year, we get Windows & the basic Office suite, add another $10/person/year to get client access licenses for Exchange & SharePoint; add another $10 for Visio, Project...
I've never had a good sense of why people are so desperate for Linux to take Windows' desktop market share. Linux already clobbered Windows on the server, and probably will in perpetuity because of its momentum, price, and developer support. It's a good place for Linux. No one can legitimately call open source unviable at this point.
I can understand it from an ideological perspective, but many of the people I see talking about it aren't ideologues.
Mainly because as Sys admins/Users we don't want to be forced into using Windows. It's selfish but seriously for those who know Linux, they are generally much more productive on it. But this often is not the case for people in more administrative type jobs. So the question becomes should it be easier to administrate and use for us (sys admins and power users) or for the average user in the office.
I'm lucky to work in a place where windows doesn't exist except for one of our project managers who like windows. No one complains about it and she is capable of supporting herself. (As a research lab, our tech support team is really small, desktop users usually have to be self sufficient.)
> I've never had a good sense of why people are so desperate for Linux to take Windows' desktop market share.
Why?
Linux is a platform, and therefore depends on network effects: The more people use it as a desktop, the more software (apps, games) will be available. The more software is available, the more people will use Linux as a desktop.
Thus, market share seems to matter -- especially for those who aren't ideologues.
Unfortunately, the theory only holds if there's money to be made by developing desktop applications and games. This is still more complicated than it should be.
I"m less concerned about Linux taking Windows market share, but I do not want Windows to be considered the only viable option, as so many IT departments do consider it.
Even just a 10% drop in market share, with Macs and Linux splitting the gains, would be awfully nice.
I call bullshit on IDC's figures. If you take public-facing websites -- something that can be checked -- MS has a 20% market share[1]. Some of these are multiple sites on one server, and some are multiple servers on one site, which probably cancel each other out so c.286M sites is roughly the same number of servers.
By contrast, IDC's figures are for a total of less than 2M servers sold per quarter. If each server lasts for 5 years, that means there are a total of 40M in use worldwide, which is way too low.
I don't see Linux share on that? I ask because we use Apache on WS08. I'm sure we're not the common case, but it would be useful to know if Apache==99% Linux or 75% Linux.
Taking a double digit percentage of a proprietary juggernaut's market with something free is what I'd count as clobbering.
That report also seems to cover shipments, as in prefab servers with the OS already installed. Don't most larger operations use homemade distributions and configurations? That's on top of any number of spawned instances of Linux in a virtualized environment. I wonder how many Linux instances are running on EC2 at any given time.
I don't see anything in the IDC report about overall OS share. Just shipments. A computerworld blog referencing a zdnet blog (which also cites the IDC report) for marketshare numbers seems fishy, as though they were fudging facts to support a headline.
Well, once consideration mentioned by Linux adherents in the 1990s was that if Windows is all the people who make decisions about enterprise computing know, they will tend to block the use of Linux when Linux is appropriate.
Most of the highest traffic sites on the web use linux/unix with the exception of MS owned sites and a scattering of other sites. Smaller sites tend to gravitate toward linux as a matter of course due to lower operating costs and wider availability of hosting.
Oh, there was no mention of this statement being just web servers.
I'd think that across the server market as a whole that Windows had won that one.
Every organization I come across seems to have a small percentage of Linux servers, but Windows dominates. I'd be pleasantly surprised if the facts had it the other way around.
> It's hard to carry the FOSS banner at that price.
Are they also counting the extra downtime, anti-malware packages and so on?
Because a Linux desktop usually just works.
And, BTW, in order to Sharepoint be as cost-effective as the worst open-source alternative (as in "groupware for masochists") Microsoft would have to pay you a truckload of money.
With Windows 7, there should be no more (or less) downtime than any other operating system.
"anti-malware packages and so on?"
Anti-virus for a large enterprise can be as cheap as $1/desktop/year.
Automated enterprise patch management is expensive, but when added to MS licensing, you still should be under $100/person/year.
Keep in mind that if you put an OS X or Linux desktop on my enterprise network, I will make you install some form of enterprise grade automated patch management on your desktop. I.E - not only do I need you to have automated patch management, I need to know that you are patched, when you last patched, what you patched, etc; which implies an enterprise class solution.
"in order to Sharepoint be as cost-effective"
I'm curious, do you have any reasonably objective data to back up that statement?
I ran enterprise class document management and collaboration with FOSS tools. At $10/person/year, SharePoint is a steal.
So here's a question. Suppose a 1000 consumers get together, and petition MSFT to buy a copy of the Windows du jour.
Will MSFT give them the same price as a 1000-user company?
And note that this story is about government. If they're being suckered into cut rate deals that disadvantage small businesses, education and consumers with discounts then there's a failure somewhere.
Word and Excel. For non technical users, you need at least these two. If Open Source is serious about the desktop, interoperability with these two applications is job one.
EDIT: Some company should create a bounty for developing and implementing a process for achieving and maintaining 100% Microsoft Word compatibility for year. You'd have to put up a million dollars for something like that, but it would be worth it.
Part of the problem is the open source movement's increasing antipathy towards Microsoft. This is an understandable attitude to have, of course, but it doesn't win you fans when you try and acquire users outside the movement itself. For most people, computing isn't a political stance or a way of life; it's just another thing they need to use to do their job.
Integrating with Active Directory to some extent, for example, shouldn't be incredibly difficult - and you would have thought that full integration with the new Office XML-based formats should be possible. But I suspect there isn't much appetite to actually do the work without substantial commercial incentive. I agree that a million dollar bounty would do the job, but then you'd have to do it again the next time a version of Office came out, and again and again. When it's so difficult to actually make money through an open source project (see LibreOffice's recent plea for 50,000 euros), that's difficult to justify.
Despite what many people who run projects say, the open source business model problem is anything but solved. I know from experience that running an open source project as a business can occasionally be a soul-destroying experience (although, particularly if you're socially orientated, the good times far outweigh that). When there's a reliable way to make payroll and grow companies through producing open sourcing software, as opposed to packaging it or selling customization services, all of these problems will go away. Until then, we're - as ever - reliant on developers and teams within companies scratching an itch.
"Integrating with Active Directory to some extent, for example, shouldn't be incredibly difficult "
I get confused on this one. I am trying to implement some simple software at work, and the software (not Open Source) is supposed to be LDAP compliant. The Unix admin in charge of the LDAP Directory insists that Active Directory itself isn't LDAP compliant, and that when people gear their software towards Active Directory they "break LDAP."
So far, his comment is correct. The product keeps trying to pull in windows AD objects just because the server is running Windows. Frustrating.
Sorry, I guess my point is, that for some of us, it gets really difficult to figure out who is breaking what in terms of compatibility.
"For most people, computing isn't a political stance or a way of life; it's just another thing they need to use to do their job."
I agree that a million dollar bounty would do the job, but then you'd have to do it again the next time a version of Office came out, and again and again.
Way ahead of you. Note I said the bounty would be for establishing a process and demonstrating currency for a year. The goal would be to take the project over and keep it going (for much less than a million a year).
Don't get me wrong, I actually think it's a great idea. But I suspect it's not above Microsoft to deliberately change things so radically that this becomes untenable in the long run.
Then if the compatibility project was tenacious enough, it could get Microsoft to increase it's rate of inconvenience to customers, to the point where they might want to switch, or realize how bankrupt their lock-in strategy is.
Don't forget Outlook/Exchange - the scheduling capabilities are awesome. I could figure out when the people I wanted to meet with were free, find out what rooms were free, invite the people I wanted to meet with (and then see whether they RSVP'd), and reserve the meeting room for a desired time slot, all through Outlook. I don't really know of an open source solution that does this quite so well.
And spend days trying to explain to people how to not screw invites up on their Blackberries and then have fits that their meeting vanished.
People don't understand that they have to manage their calendars and that meeting requests are a convenience feature. They think there's a central calendar and everything gets updated automatically.
Have fun when the CEO misses an important call or meeting from his calendar. This stuff is like crack, it makes people think it's easy and magical, but it's a support nightmare.
While not open source, there are free(mium) solutions like Tungle (or doodle, etc) do this quite well. Tungle even allows you to do this with people that aren't on the same exchange server
What about a proprietary solution that can integrate with an open-source stack? I don't mean a service like Google Calendar, but something people run on their own servers.
Interesting. 10 years ago, printing and scanning from Linux was just agonizing. But today, I actually find Ubuntu 10.10 almost as easy as the MacOS: I ask it to look for new printers, it gives me list of printers on the network, and I choose the one I want. Everything just works. Similarly, the Gnome scanning application is a lovely way to make PDFs.
This usually surprises the Windows users, who were telling me, "Oh, you're going to have figure out where the CD is, and install some software, and you may have problems with <X>..."
Now, I'm sure that there are other printers and network configurations where Ubuntu will fail. But for the most part, they've done a great job.
On the other hand, I do need to keep MS Office around for marking up contracts. Open Office doesn't support Microsoft's Track Changes feature, which is a dealbreaker.
The Foreign Office started migrating its servers to Linux in 2001 and since 2005 has also used open source software such as Firefox, Thunderbird and OpenOffice on its desktop systems.
Even in 2005, configuring printers in Linux was painful. And a big government bureaucracy cannot be deploying every distro-du-jour to keep up with improvements... most comparable Windows shops are still running XP.
One thing that I think a lot of tech folks don't appreciate is how much business generally craves reports and printing. Its just reality, utopian dreams of a "paperless office" notwithstanding.
I worked at a company that built a web-based system for a government agency. One thing we did, with EVERY release and EVERY new feature was include new or updated reporting (Crystal Reports) that paralleled the new features. Developers hated working with Crystal and hated doing report work in general. But the customer LOVED it.
100% ? There's a lot of weird stuff in those, I don't see it happening in a year - there could a lot of improvement probably. I guess some sort of automated testing that loads documents with tests in both and compares the rendering would a great step forward.
Interesting idea. It should be relatively easy to make various versions of Word and Open/LibreOffice load documents and print them and then compare the bitmaps for meaningful discrepancies.
You have to wonder, though, in a big bureaucracy, how much of the pain is due to "not 100%" compatibility, and how much is due to "this Linux thing is not what I'm used to".
Having worked with people trying to encourage adoption of open office I can say the main problem is compatibility. Pretty much everyone can adjust to the slightly different interface.
The issue is that if you save a doc in open office and email to a customer who uses word if you've used anything beyond the basic text features there is a good chance when they open it there will be something wrong. The reverse applies too.
Almost everyone I've tried open office with has eventually given up and migrated back to word primarily because they couldn't open files other people were sending them.
We're on Word here and still have that problem. If you have a newer version of Word that does .docx files, you have to save as the older .doc files, because some recipients will complain that they can't open the newer ones. And if you're on the older one yourself, you have trouble opening stuff people send you. Whenever Word files go out to a distribution of more than two or so people including some external collaborators, I've always had problems. Afaict, Word on the Mac also behaves differently from Word on Windows, at least in terms of who reports back to me that they can't open a file I sent, and god help you if you try to embed media or use the reference manager.
Honestly OpenOffice users give me the fewest problems, because they seem to be able to at least open anything I send them, even if the formatting is a bit mangled, while a significant percentage of Word users report back with "I couldn't open your file" or "your file seemed corrupt" or something.
"Honestly OpenOffice users give me the fewest problems"
Just a guess, but perhaps that's because the OO users that have actually stuck with it are the more technical ones who look for their own solutions first. I use OO primarily at home and when I get files I can't open I try Google docs, Office live apps, or the Word viewer.
We're on Word here and still have that problem. If you have a newer version of Word that does .docx files, you have to save as the older .doc files, because some recipients will complain that they can't open the newer ones.
This tells me that Microsoft is actually in a vulnerable position here. If they start ramping up their rate of change to maintain lock-in they're going to inconvenience their customers even more. OpenOffice will start looking like the better alternative.
Honestly OpenOffice users give me the fewest problems, because they seem to be able to at least open anything I send them, even if the formatting is a bit mangled, while a significant percentage of Word users report back with "I couldn't open your file" or "your file seemed corrupt" or something.
It's not good to get egg on your face because the formatting is messed up. That looks like your fault. A corrupt file is seen as something like traffic. "Computers are tricky and stuff happens" is the usual reaction I see.
Sorry, but even Microsoft software itself cannot provide 100% backwards-compatibility between versions. And forward-compatibility is almost non-existent.
1. I have a feeling, that even if MSFT will open-source all their SW products - people still will be complaining ;)
2. Office is de-facto industry standard. The only way to beat it is to innovate into new paradigms, i.e. something like Prezi vs. PowerPoint.
Or if somebody will find new better way to do electronic tables - they will beat Excel, etc.
3. How can one expect to make open-source desktop better, when most open-source hackers using proprietary Macs with OSX?
I'm a huge fan of Linux and want to see it succeed but it still has a ways to go for the enterprise desktop market.
The one biggest problem is Microsoft Office. I use OpenOffice on Linux and iWorks/NeoOffice on Mac often but I still find myself using Microsoft Office or Visio when I have a lot of work to do. There are just so many things that can be done quickly in Word or more advanced in Excel that I have yet to find in the other options despite all my googling. Regards of how you feel about Windows, IMHO Microsoft still has the edge when it comes to Office Suites.
It's funny, Excel was the reason I switched to OpenOffice.
1) Closing one Excel windows closes ALL of them. Often I (or other people I was supporting) would be working on an important Excel file, and then open another one temporarily, close the 2nd one, and then unthinkingly say "no" to the "Do You Want to Save" dialog that you THINK is for the 2nd file, but is actually talking about the very important original file. Oh, and Word (which also has a separate window per document) doesn't act this way at all. Oh, "save more often" you might say? Well....
2) Excel doesn't let you Undo past the last save. This is annoying for a lot of reasons, but is completely terrible when you opened a file with the intent of saving your changes as a copy. Oops, you "saved" instead of "saved as"? Your original document is gone, and pretty much unrecoverable.
OpenOffice just works.
(I don't know if the latest versions of Office addressed either of these issues... and I don't really care)
No it doesn't. At least not for the past two versions of Excel (I can't remember Excel 2003). But there is a button to close the current window or specific window.
Excel doesn't let you Undo past the last save.
Yes it does. At least it has for the past two versions.
I don't know if the latest versions of Office addressed either of these issues... and I don't really care
Honestly, did you really ever? Those were the only two reasons you switched to OpenOffice?
I never worked at a place willing to put up with supporting the 2007+ versions of Office.
Those were certainly the "I can't take this anymore" reasons, yes. OpenOffice also read the new XML-based file formats better than the older Offices with the converters.
Since then I've also switched to Linux, which simplifies my decision.
I disagree. I have tried using lots of other office software like OO, more lightweight text writing apps, and of course online apps like Google Docs. In the end I'm back to using MS Office (both on my primary Macbook and the secondary WinXP), even without the interoperability issues it still easily beats all other solutions. Also, as another commenter noted, if they can get them $10-100/yr, there's really not much to think about.
The truth is, Office software is hard, and even though the MS product has a lot of suck, the others are much worse. Finally, the new 2010 version of MS Office actually really good.
I'd say Microsoft Office, everything else Microsoft which integrates seamlessly and most of all Active Directory. Getting anything equivalent to Active Directory up on Linux is an absolute nightmare, or at least so it was last time I tried. Making something as basic as making a Linux machine join a domain and authenticate against the DC work cost me a full weekend, and yet in the end I just had to give up.
That's really not a user- nor computer-management strategy that scales.
> Getting anything equivalent to Active Directory up on Linux is an absolute nightmare, or at least so it was last time I tried. Making something as basic as making a Linux machine join a domain and authenticate against the DC work cost me a full weekend, and yet in the end I just had to give up.
I feel that your account of making a Linux machine play nicely with AD deserves a 'YMMV' addendum. In my experience it is not common that this should be so difficult to implement with a reasonably up-to-date Linux system. It also doesn't hurt to use a distribution that has undergone testing for this usage scenario, which is definitely true for the "enterprise" distros - both RHEL and SLES should give you very little trouble when it comes to joining a domain and authenticating against AD. It is indeed mostly a "just works" type of thing in the majority of cases.
Maybe that was because you were trying to apply AD ideology to GNU/Linux, i.e. do things in "Active Directory-way"?
I don't know about AD, but implementing a single sign-on authentication on GNU/Linux is simple. Just configure PAM and roll the configuration to all machines (with rsync or whatever you prefer). I've never tried authenticating against a LDAP database, though. However, I believe if I'll see a Windows machine and be asked to set up a simple SSO configuration, it would take ages to understand (not just step-by-step copy some tutorial from the net).
I agree. I actually meant to say the biggest hang up for the enterprise to switch over to Linux is finding a good replacement for Office. However, you absolutely correct about how easy it is to get a Microsoft AD up and running.
Though I am sure there is a way to do it in Linux if you have all the experience, sometimes when a System Admin has to get an environment up and running it is easy to get it done with Windows and AD.
They have specialised scanners that allow them to scan passport booklets and other special formfactors without the need of manualy opening and lying them down on the scanner.
The german Federal Commissioner for the Stasi Archives has special setups that allow them to enter bags of torn apart papers to scan... so i wouldn't wonder if they really need to write special drivers for their hardware.
I have a common canon multifunction laser printer (bought 4/5 years ago) and don't have any driver for it...
I suppose they could have even older printers.
"Savings were limited" but they decline to say how much. Also "Users complained of compatibility problems" -- all the more reason to switch; a proof of the MS lock in.
I wonder if they write something like "We were only able to save $5K/year/person, but we didn't want to inconvenience our workers for such a paltry sum", what the response would have been. I can assure you it is very far from the $100/year licensing that people mentioned in other threads. If that were the case, no one would switch on one hand, and Microsoft wouldn't be the behemoth that it is today on the other.
It's $100/year/person for their desktop software, + tens to hundreds of thousands for the server software that goes with it (to microsoft), plus a $2000/year/person for lost time due to viruses, spam, and being able to play WoW on their work computer, plus $2000/year/person when you look at the administration cost -- windows shops everywhere I've seen have a much lower admin/user ratio (e.g. 1 in 10 or 1 in 20) compared to unix/linux (1 in 50 or 1 in 100), and even more so when the end users are doing just word processing+spreadsheet work.
Also, what the hell were they doing writing printer drivers? I've been using linux on the desktop since 2005, and didn't have to install a printer driver once.
It's hard to carry the FOSS banner at that price.