Let's kick off this forum: open source in the workplace
 

I have always wondered... is Linux still free as in beer; and free as in freedom for profit-oriented corporations? Perhaps someone can tell me what companies do this, and which companies don't.

Linux of course is free as in beer; similarly it is also free as in speech - just look at Sourceforge. The "free as in beer" simply means that the monetary cost is zero dollars; the "free as in speech" simply means that anyone is able to participate (as in creating new apps).

In terms of for-profit companies, both Google and Amazon run Linux exclusively (as far as I'm aware) -- what better examples would anyone need to answer the question "Is Linux ready for prime time?" Most assuredly, YES. Any number of other companies are also using Linux, but not for philosophical reasons, they use it because it works. -- J.W.

PS: if you're interested in checking out what a website is running, go here: http://news.netcraft.com/

I work as a contractor for a very large government agency where I specialize on network security. I manage many boxes that handle monitoring and detection on the network and these machines are almost exclusively Linux.

Before taking this position I worked on a project for the FAA where we were rolling out a Linux version of their ETMS (Enhanced Traffic Management System). There is a lot of Linux being used out there.

Linux is everywhere, like it or not. A lady I work with (the lady that used to have my job) tell me that she would give me a years salary if Linux was ever put to use in my company. We now have 3 and our e-mail server on the upgrade schedule next month.

I still have yet to see that years salary bet though. Unlike Linux, she was just all talk.

Smoothwall
 

A great example of a linux based open source product is Smoothwall (a very nice firewall distro) to be found at smoothwall.org

The company has a number of products from the free opensource version (www.smoothwall.org), to a closed source $$ more feature rich version.

the open source base product is very good (free as in beer), and a number of people who do not work for the company (myself included) have contributed modifications that improve VPN support, viewing of current DHCP leases, better logging, multiple config profiles, etc etc (free as in speech).

The freeware product is used by a number of companies and works very very well. Some of these companies use versions with the user supplied mods, some use the vanilla product.

Cheers
Jo

Nice this forum!!! My question is what possibilities are there for IT companies to make money with linux? For example is it possible to install a server for a customer and manage it for a monthly fee and still be doing things legal?

So not charging for the software but charging for the service of keeping the box up?

Of course thats legal. the firewall I talked about in an earlier post is an example. I manage the config, security updates (ensuring they get done), port forwarding etc for a few companies. These companies aren't large enough to have full time IT staff, but need secure robust systems.

I Am also looking to do linux servers running samba in the same way.

Regards
Jo

Like anything else, if you have the business skills, you can made as much as with any other OS.

You'd be legal as long as you pay attention to the licenses.

There's really only one angle to watch out for regarding using Linux in a for-profit business (or in fact, just distributing any software that you obtained under the GPL):

Are you modifying any GPL'd source code?

If yes: You must also release the source code of your software, in addition to any binary version
If you did not modify GPL'd code (i.e. you added your own extra code as additional software, but you did not modify the GPL'd software in any way) then you're free to do whatever you want.

There is another problem with "Open Source" software, and that is that some of it is not distributed under GPL, but other licenses entirely. Software from The Apache Foundation and XFree86 are two examples that use their own, special licenses. You need to be aware of what the terms of those licenses are if you're including them in your products.

One reason that a lot of larger companies have been hesitant to use GPL'd code as a starting point for products, and prefer to use BSD licensed code is that the BSD license allows you to do anything, without releasing the source. The main requirement of BSD licenses is that you include their copyright notices in your product documentation. Although many companies (even very large companies) do use Linux (and other GPL'd code) in their products, most of them are very small companies the see Linux as a way to create a product with a very small monitary investment, and they don't always do their due dilligence on understanding the license requirements. Companies that use BSD licensed code never have to worry about that.

I'm not try to say one way or the other is better, I'm just explaining so you understand what the differences are in different Open Source licenses and what they mean for a business trying to create and sell products.

After doing some internal benching marking my employer ( a Fortune 500 company ) has recently decided that they will not longer buy Sun or HP UX hardware/software and will standardise on Red Hat Enterprise Linux. Now whenever a vendor calls about a product the question is "What is your Linux strategy". Windows and Solaris are now second tier vendor for us.

In addition, just this week, I found out that executives at the top are asking questions about how well we can leverage LAMP (Linux, Apache, MySQL, Perl/PHP/Python).

Beware of MySQL, it has some pretty serious scalability issues :( If you want to do highly available/high load database transactions, there is pretty much no alternative to Oracle. Yes, I have a lot of experience with applications built on MySQL. For low to moderate load, it's great... when you start putting stress on it, it cracks.

On the bright side, Oracle runs on Linux!

What a great idea for a forum!!
For most non-computer enthusiasts, their greatest experience of computers will be in the workplace. i.e. Word, excel etc. The issue with using open source applications is the costs it passes onto companies in terms of transferable skills versus the costs of proprietary packages (MS) or open packages (Openoffice). I am thinking about non-IT companies, who use IT as part of doing their business.
I don’t have any trouble picking up a different package, but I am thinking of the more non-computer literate workforce, who have gotten used to MS packages, and if they moved to a different package would be lost. i.e. spell checker and other functions are in different places.
This could give trouble in the recruitment market, as most people presently put MS Office in their CVs, if this changes to a raft of different office programs (under Linux) companies may then discriminate people by which package they have used, to ensure that they can hit the ground running once they start their new position.
This type of knowledge issue also extends to IT staff. With most offices employing MS desktop products, workers in this area are conversant with maintaining the MS products, creating a pool of similarly experienced and qualified MS system maintainers. Although the MS products have high capital costs, maintenance costs are reduced by competition between maintainers.
The opposite would appear to be true with Linux products, capital outlay on packages is less (substantially?), but the number of maintainers is far smaller, who could (whilst the pool remains small), demand higher fees for maintaining the Linux infrastructure.
End-user education is one of the most important issues, as in most companies staff are the most expensive part of a company’s operation, and no company would want to adopt an IT policy that seriously increases their staffing costs, or increases the difficulty of obtaining the right staff (i.e. ones with Linux application experience).
I am just wondering if this is an issue that will generally leave Linux working as a Server, with MS on the desktop, leaving public perception of computing as being MS (because that is what they see at work)?

Well, Linux on the desktop basically has what, 2 years (until MS Longhorn is released)? That's pretty much the window (heh) of opportunity. If I recall correctly, MS software licenses are for 3 years, that means WinXP licenses are going to expire well before Longhorn is available. Will enterprises want to spend money on WinXP again, knowing that Longhorn is just around the corner (which they will have to buy totally new licenses for)?

If LotD (Linux on the Desktop) can mature enough before MS starts signing people up for Longhorn, then there's a chance. Otherwise it might not be until the next OS cycle (3 years past Longhorn).

Right now my feeling is that LotD is still not fully baked. The most important thing for a good desktop suite is that all the essentials are provided, and there's no confusion possible (everything about the UI needs to be intuitive). The problem that LotD has today is that most distros bundle everything but the kitchen sink. You have 3 mail clients, 4 text editors, 3 browsers, 2 word processors, etc... Now they do this because Linux zealots whine about being "restricted" if 4 different programs aren't included for each task (everyone has their favorite and won't budge).

On the other hand, the few distros that do only provide one option (such as the KDE suite, Konquerer, Koffice, etc) the quality of the software hasn't quite reached the functionality that MS provides. I know that sounds laughable, but really there's a lot that goes into developing fully-functioning and stable software, and the Open Source alternatives aren't there yet. OpenOffice is pretty nice, but I have problems with it crashing occasionally (both on Windows and Linux). Evolution is nice, but I've also had problems with that crashing (or individual components of it crashing). My experience with GTK apps overall is just very bad. Then you run into the problems of dependencies if you try to install an alternative, etc...

Here's my list of things that need to be fixed (in no particular order) before LotD will become a wide-spread reality (yes, I realize that a few sites have switched already, but I'm talking anything close to even 20%).

• KDE vs. Gnome: Figure out which one, or merge them, or something! It doesn't do anyone any good to not have a standard desktop environment. Splitting development work on something this central to an OS (at least, a desktop OS) is counter-productive.
• Package formats. We have RPM, DEB, MDK, TGZ, etc (probably more that I'm forgetting). That's crazy. Every time someone releases new software they have to do it in each of these package managers? Or the distro has to take the package and rebuild it in their app before their users can download it? If there isn't a package available in your format, then you have to build from source... well that's not going to work for desktop deployments. Yes, some of the tools will convert packages of one format to another, but that's a ridiculous work-around--standardize already!
• Productivity suites, there are at least what, 3 major suites? Again, this is unnecessary. If you want to switch between Linux distros (say you change jobs and the new employer runs a different distro) you don't want to have to learn your software all over again. Pick one and focus development on it. If having lots of different choices was smart, WordPerfect would still be popular.
• Dependencies. This is a major stumbling block, it's one of the worst complaints of just about anyone I've talked to. Particularly dependencies for various graphics and UI libraries. Even distros that ship with with a number of extra packages on a disk some times have dependency problems with there own bundled software! That's bad. I don't see this problem getting better unless the distros exert much more control over what software they include, and in fact do a lot of the software themselves to ensure interoperability. Of course that would lead to worse fragmentation between the distros, so it's kind of a catch 22.
• Laptop support. One of my coworkers is an avid Linux users and developer (he contributes to several projects). He admits that installing Linux on a laptop is more of an art than a science. The last few installs he's done, he said he put the laptop in the refrigerator to keep the temperature down during the initial install, since CPU stepping wasn't active. That's a problem. Enterprises use a lot of laptops, and the trend away from static workstations to laptops is just going to continue. Within a few years I predict that nearly all employer-issued machines will be laptops. They need to be able to install Linux on a laptop, and the OS has to support all the proprietary devices found on laptops. Of course a lot of that is dependent on hardware manufacturer giving up the specs and documentation for their devices so that Open Source drivers can be developed, but also OSS developers need to concentrate on laptops as a primary platform.

So that's my list. If OSS can get those solved in less than 2 years (unlikely) then Linux can take away a large chunk of the desktop market (not a majority, but something noticeable). If all those problems aren't solved, then Linux will continue the slow plodding growth with a few brave adopters here and there until it's really ready for prime-time.

chort, were did you have those problems with? What distro? What versions of KDE / Openoffice.org / Gnome?


Many windows programs have different layouts. Almost all the windows versions have different layouts. Especially Windows XP is a hassle. What about the ever changing 'Personal menus', what about 'hide inactive icons', what about the choice between the classic menu and the new XP start menu. What about the classic view of the folders, control panel, My Computer and other icons missing from the desktop??? What about Office XP hiding items that aren't used often. Seems to me that's not very standard either. If I have customer on the phone, most of them are not to good with computers, most of the time i don't even know where the button is.
Windows has package formats also. What about MSI, MSI 2.0, Nullsoft Installer, InstallShield, WISE installer etc etc. At least if you have RedHat or Debian you can install a package on all machines in your company just with one command!!!
ie. apt-get install openoffice.org
It will take care of the packages for you.
I don't think windows has less suites available for productivity. Is this really an issue? I guess most people didn't bother looking for other office suites. But i can tell you some of our customers are seriously thinking about using openoffice! Office 2003 is just so expensive, and who needs all those features?
I don't think depencies is a OS-related problem. For most games you need DirectX or a new video driver. For most Visual Basic applications you need runtime dll's. For .NET you need .NET framework.

Most programs rely on other programs. That's not a problem. Just a fact.

At least most packaging systems warn you for depencies before hand. And with tools like apt-get and yum you don't even have to worry about downloading it manually, like you would have to do with DirectX and stuff.
I partly agree with you on this one. Although almost all hardware on my pc works on my laptop. You should check if your hardware is supported with linux. I have the following things working (right out of the box):

Wireless networking : Rock Solid! On windows I had problems sometimes with it forgetting my key every once in a while
My mousepad. USB 2.0. Sound. DVD-player. CD-burner. APM suspend. CPU frequency adjustable. NVIDIA 3d opengl stuff.

What would you need more? Just make sure you get the right software.

It feels to me that is a while ago since you tried LotD because all the problems you mention don't seem to affect me.

While I see some hurdles that don't yet make Linux desktop a viable option, I disagree on many of your points. I think what you have mentioned is the very thing that makes Linux as good as it is for the masses. Many of your points should and could be settled on a business by business basis. They can standardize just like anyone else. If "somebody" decided to standard Linux on something I didn't like, they've done the very thing that made me switch to begin with.

The biggest problem I see is that free desktop Linux software, while reliable and simple, just doesn't hold a candle to what is available commercially, yet very little commercial software is available. Other things, like fonts, are a problem too. I prefer KDE, but sometimes the fonts just don't fit the space they are supposed to fit. To me, that is a problem. I could go on,...