Slashdot Mirror
10 Years of Pushing For Linux — and Giving Up
boyko.at.netqos writes "Jim Sampson at Network Performance Daily writes about his attempts over a decade to get Linux working in a business/enterprise environment, but each time, he says, something critical just didn't work, and eventually, he just gave up. The article caps with his attempts to use Ubuntu Edgy Eft — only to find a bug that still prevented him from doing work." Quoting: "For the next ten years, I would go off and on back to this thought: I wanted to support the Open Source community, and to use Linux, but every time, the reality was that Linux just was not ready... Over the last six years, I've tried periodically to get Linux working in the enterprise, thinking, logically, that things must have improved. But every time, something — sometimes something very basic — prevented me from doing what I needed to do in Linux."
They don't want to migrate of Microsoft...Hell, that's the root of the whole problem. They want to not have to use Windows, and Microsoft has a huge amount of money riding on people not being able to use Office or Exchange in a Linux environment.
Being a veteran of many different Linux migrations, some successful, others dismal failures, it always comes down to a few applications:
Office: StarOffice/OpenOffice is not as good.
Exchange: Goddamn managers and their shared calendars.
Unsupported Widget: Every goddamn company has an Unsupported Widget written by a savant who was killed by a bolt of lightning. The Widget is always absolutely critical to their business, and ALWAYS runs on some piece of hardware that doesn't exist anywhere else in the world, and only talks to certain versions of Windows.
Every one of these things will come up, and even if you're successful in talking them into going over to OpenOffice and Lotus, and you manage to slay or replace the widget, it's going to take longer and cost more than you would have thought.
In the end, it's always about the damn tool. Use the right tool for the job. Don't try to force Linux in where you know there are going to be problems. The jackass in the article was subcontracting for DELL, the king of the Windows shops, and he thinks he's going to be able to get by on a pure Linux environment? He's a fool.
ad logicam Claiming a proposition is false because it was presented as the conclusion of a fallacious argument.
I am TheRaven on Soylent News
I use Thunderbird to access my email at work, and I'm assuming that's on an Exchange server. Sunbird can do calendar sharing, just not with Exchange (and I haven't tried with Evolution lately) -- plus, there are web-based solutions. So, the email itself is a known and solved problem, if we have decent IMAP support. The calendar/scheduling stuff may require a different infrastructure -- but keep in mind, this is a lot like having the open office suite (which took a LOT of work) -- Microsoft hasn't given us any specs, therefore we can't really do this. And we'd much rather do it in a better way anyway.
Also, what do you mean by "Push email" and "mobile operators"?
Don't thank God, thank a doctor!
It's called GroupWise and it runs on Suse Linux Enterprise Edition. It does everything Exchange does, except run random vbscripts!
Hmm. I rolled out Linux Corporate wide. The only problem I had was convincing the administration that a service contract is NOT Neccesary and that I'm better off with the Whitebox Kickstart server I built rather than DAMN REDHAT! I built Linux
database/application clusters, a sendmail server that outperformed Exchange. I substituted LDAP for the addressbook and wrote
my own damn fuzzy search so if some idiot can't spell a name correctly hell I can still list a user that has a name that sounds like what was typed. The problem with coporate deployments is because too many managing bean counters listen to FUD. All I say is wait until Vist get's crammed down everyone's throughs. I even refuse to upgrade my Mac G5 with dual a 970MP to OS/X 10.5 leapard.
I'm gonna leave it at 10.4.6. Why cause DRM shit is gonna lock up leapord up eventually. I've got code I wrote for my Linux and
G5 so I can mange my mpeg files on my IPOD. I can copy them off the ipod with no problem. That's gonna get locked on leapard.
I've got my Linux and OS/x hosts doing what I want. Now getting back to coporate. I've built clusters with Linux connected to
EMC storage Arrays that can keep up with IBM P5 systems. How do I know, my company that I worked for had an IBM Mainframe, An AS400, An 8 processor P5, and an Linux cluster that consisted of 6 Dell 6650's dor database, 12 2550's for application servers,
6 Concurrent managers, 2 configured in a cluster as a mail hub running GFS, 2 servers in a load balancer to offer IMAP, and POP
connections. The system screamed.
So, if any of you need consulting on the side, hell that's how you earn the big $$$ with Linux, drop me an email, maybe I can design a KILLER system on the side. I'm happy where I am now so I'm not moving. I get to play with P5 hardware, turning the
company on to Linux Cluster, educating to what DRM is actually all about.
So heck... Now you know why I call this guy a Mother Fudder, he don't know jack shit and it sounds like FUD to me.
What ignorance: As for Linux groupware packages lets start with the best known: Lotus Domino and Novell Groupwise - Both run on Linux Then there is the open source crowd including Zimbra, Hula, OpenGroupware.org, egroupware, phpgroupware and a host of others. As for push email, funambol, aka sync4j, will sync and push to a wider variety of devices than any proprietary variant out there. As a matter of fact one of the largest wireless carriers is using it for their 22 million handsets, Fortune 100 companies are using it, and even phone oems are including the client software. L
...We've just transitioned to using Scalix for our email and calendaring. It's got public folders, calendaring, email, push capability, nifty webmail client, Outlook Integation, Evolution Integration, and pretty soon CalDAV support so Mozilla Sunbird/Lightning and Apple iCal (in Mac OS X 10.5).
It's based on what used to be HP OpenMail, so has roughly a 20 year history. It's mature and well tested and not that expensive (compared to Exchange etc.).
I am NaN
Zimbra is indeed promising. For those that said I haven't looked hard enough, it's not me that's ignorant.
My post was not to say there's nothing developing, rather that after a decade or more we haven't seen a standard develop.
Almsot everything that's suggested does about 80% of what Exchange does, never quite filling the need entirely. Many have proprietary add-ins to work with Outlook as we don't have a client to replace it (bar the attempt by Evolution which doesn't run on windows yet. Sunbird has promise but it's very early days).
What frustrates is that from the outside it seems lots of folk are trying to be the next Exchange with their own formats and techniques, rather than us seeing an open deployable standard with interoperating clients and servers.
Too bad it's not open source. If you want Outlook/MAPI compatibility, you have to get the "Network Professional Edition" which has a per-user license fee.
"You never know when some crazed rodent with cold feet might be running loose in your pants."
-Calvin
In the finder, you can get the very same dialog twice, and one time you have to click to activate and then click the button, when the other time the dialog comes up in the foreground as it should (neither time was I typing anything when it popped up) and you just click the button. Some context menus let you click submenus to open them; some of them you have to hover, because clicking the submenu option closes the context menu. The amazing expanding and contracting dock is just a stupid idea. Icons can appear underneath the dock and will do so. The desktop updates on its schedule, and when creating a PDF (for example) on the desktop, I have to force the desktop to update by clicking in another window, then clicking on the desktop, or it will literally never update (this is 10.3; supposedly this got better in 10.4 but we don't get a patch or anything. Even Microsoft is kinder about backporting fixes.) Applications bundled with the OS use three different widget sets to this day.
Apple talks a lot about interface consistency and uniformity, but they are full of shit. Apple has also done any number of nefarious things to boot, but that's a topic for another day.
"You're right," Fisheye says. "I should have set it on 'whip' or 'chop.'"
This blows, because then I can't use my distro's package manager to keep up on updates.
Also, iirc it requires a pretty hefty machine.
Nope...
:)
To add a computer to a domain you must be logged in as an administrator onto that PC. Assuming all network settings are right (DNS specifically), you can opt to join a domain if one exists on the connected network. Once found OK, a network administrator must authorise the computer onto the domain, and thus, two separate administrator passwords are required for this 'rootkit' to install.
By joining a domain of course, the original 'owner' of the PC is agreeing to hand over complete control of the network administrator(s), and therefore be subject to all network policies pre-defined. By entering a network admin password, the network admin accepts they want the machine on the network.
Clear? Marvellous
throw new NoSignatureException();
I have to agree with you, but I think I think changing to linux has more to do with the costs of training and hiring or educating an IT staff on how to manage linux instead of MS. Red Hat DOES offer support, but people still don't buy it because the people there just know how to use the MS software already.
Think about your average desk monkey (secretary, VP, manager...), how much of their time is it worth to teach them how to use something new over simply paying a fee for upgrading what they're already used to? My company employs tens of thousands of people, and any time you say "well, 10,000 Office licenses are EXPENSIVE!!!!" you have to figure out what the cost of reteaching 10,000 people is, as well as the lost productivety during the transition.
Moreover, when you have 50 or 100 IT people who only know Windows, then WTF are you gonna do when you switch?
So there are studies, which I wholeheartedly believe, that show that a small shop can switch, or a new shop can use Linux, and save tons of cash. Unfortunately, large companies would take such a huge hit it simply isn't worth it. I know it's a one time hit, and in the long run companies can save, but the initial cost is so much more significant that it takes many more years to pay off than people give it credit for.
As a home user, I have to agree with the gist of this article. I've been using Linux almost exlusively for the past year and a half (off and on before that). FC 4 finally hit on a combination good enough for me to use it full time for work.
So I installed FC4 on my desktop at home, and laptop, too. So I was getting the itch recently, and thought about upgrading to FC6. For the most part, it went swimgingly. At work, I had a few configuration problems, some minor bumps (why on earth would an upgrade overwrite my networking config files? The hardware hasn't changed! My network hasn't changed! I'm just upgrading the OS).
Then I tried installing it on my laptop, and that's where the real problems began. FC 4 installed just fine, and I got my wireless networking up and running in no time. FC 6 was like going back to the stone age of wireless. If I didn't have a desktop PC to download stuff on (and a USB drive to transfer files), I would have been screwed.
So it might be a minor point, but do you think a 2000 to XP upgrade, or an XP to Vista upgrade is going to give you FEWER wireless options?
It's time we face the facts that developers develop for themselves when they are donating their time to OSS. The programmers working at MS are giving the people what they want, regardless of how painful or annoying it is to write the software. Back in the day, OSS programmers would complain about having to write even a simple UI for their tools because the command line gave you all the options you needed. I see the point - the UI often takes like 95% of the work for development, and often 99% of the resources used to run the program. Many programmers (not all, by any stretch of the imagination) have gotten over that, and there are tools to help with making UI's for your command line tools, but it's still not keeping pace with all the work MS (and Apple) put into making the experience easy.
If anything, if Apple could get some development along these lines, they would be the ones poised to take some market share. But we're already in the toilet, spiraling down towards oblivion - things have been like this for so long, that there's simply too many Windows only legacy applications that tie companies to Windows. Forget about exchange, what about programs like ScheduAll (a broadcast resource management system), TaxWorks (accounting software, obviously, only available on Windows), and even IE (we have internal sites that only work with IE!).
Working in broadcasting, even if we could use Photoshop, Maya, XSI, VizRT, and FinalCut Pro (to name a few) on any platform, we STILL couldn't switch! We buy and rely on Maya plugins, for example, that ONLY work on Windows.
Unless you c
Stupid sexy Flanders.
I looked at Zimbra. The problem was that it wouldn't intergrate with our existing LDAP/Kerberos environment, or our mail for that matter.
I can appreciate that I might need to migrate mail over, and that's not a big deal, but I'm not ripping out a tested and running LDAP and Kerberos deployment that runs across three continents and a dozen servers just for some calendering.
As long as the business demands compatibility with MS format documents
there is nothing you can do. If you start out from scratch in a start up company
and had to build the IT structure from the ground up you could do it.
Open source software fine for running a business as long as you are not
locked into some vendor already for something. I'm sure somewhere out there,
there is a company that has gone this route and was running Linux from day one
(or maybe gave MS the iron boot, bit the bullet and started over from scratch).
Or you could run Ubuntu Server and ask it for a LAMP-out-of-the-box.
Don't thank God, thank a doctor!
I don't think this guy is a professional. I really don't. His writing sounds like he's more interested in trolling Linux users than actually imparting wisdom.
So I'll bite.
A professor of mine once said, "I use operating systems for what they're good at, not what they're bad at..." This guy could use that advice. At the time, the college was a mixture of Windows NT and Linux machines - the Linux boxes were used for file and print sharing, and the NT boxes for Exchange.
Complaining that Linux doesn't support Exchange is like complaining that Windows can't read your ext3 formatted floppy, or that it can't see your NFS shares. Windows wasn't built to use UNIX filesystems; Linux wasn't built to use Exchange.
So why don't we turn the argument around: Microsoft failed to build software that interoperated with UNIX. After, their web site says it does. I think the real failure here is Microsoft's: Office doesn't support OO.org file formats. And they don't support using the UNIX mail command, either. I mean, clearly, this is all Microsoft's fault because their software doesn't do what it wasn't designed to do, right?
I don't have problems using Linux and Windows, mostly because I've come to know the strengths and shortcomings of each. I'm not going to bang my head against a wall because Windows doesn't support OO.org file formats, or because Linux doesn't support Exchange.
Instead, I'm going to use the right tool for the job.
The society for a thought-free internet welcomes you.
The saddest thing is not that the world is imperfect but that there are far too many people that have such low expectations.
A Pirate and a Puritan look the same on a balance sheet.
I'm the editor of the article, and many of these comments are mean-spirited and shortsighted.
Where, exactly, in the article, is there any indication that either Jim (or I) have any disdain or hatred for Linux. The article is essentially saying "I've tried using Linux in the workplace for ten years and it hasn't worked yet, partially because business needs to work with Exchange, and most workers do not have any say in what backend they are forced to work with. It's enough to make you give up on it, but I've got such a love for Open Source I keep going back to it."
This is not Microsoft astroturfing (I'm actually working on a freelance article talking about how Ubuntu Linux works in the home - I prefer Ubuntu Linux to Windows for day to day home productivity, personally,) and I'm insulted by the insinuation that it is.
This details a real problem towards Linux adoption in the workplace. If you support Linux, you can complain about it being FUD or you can get off your bottoms and start coding to solve the problem.
-- Brian Boyko
-- Editor, Network Performance Daily.
-- Questions about the article will be addressed at brian dot boyko at netqos dot com.
I used to work for NetQoS. I no longer do, but want to keep the excellent karma attached to this account.
Exhibit A
Occam's razor is the blind faith in the natural selection of least resistance and in universal oversimplification. -- EF
Your not really getting that much.
I can plug in a completely blank system into my network (no OS at all) and have it fully installed with linux and configured in less then twenty minute. The only interaction required is entering the machines MAC address in a config file on the dhcp server.
I also get daily change reports on all my machines, thanks to tripwire. A consolidated and condensed log reports from every machine in my network. Plus automatic pushes of new configurations, based on groups, thanks to cfengine.
The only place I see linux lacking today is in groupware. In terms of administration a single admin can handle 10x the linux desktops as windows desktops.
Besides, most of the time when I use rdesktop, it's to manage windows servers. So it matters little whether I run Linux or Windows in that case. Actually, I do prefer rdesktop over the Windows terminal client. It's much more configurable, and consequently I've written a script that will dynamically size a new rdesktop window to fill in my current Linux desktop, minus a strip on the left hand side to view/access Linux desktop objects. I just can't do that with the Windows client.