Slashdot Mirror
Major Retailer Chooses Linux for its Tills
An anonymous reader writes "ZDNet is running an article on how Matalan has installed several thousand point of sale terminals running Linux rather than Windows. The reason? Reduced cost of ownership. It was a big consultancy that did the work, Capgemini, and IBM on the kit side. Sounds like some people can get Linux to work in an 'enterprise environment' after all."
Point of Sale systems are really not enterprise level software or whatever. Usually the simpler it is it the better. Using linux for Point of Sale systems are just a good idea, first you can make linux very basic without the crap. Having it in a small factor allowing it to run on cheap systems, without the extra junk in the way. But to say this proves the linux is enterprise ready because of these is just silly. Most Point of Sales systems are running on DOS.
If something is so important that you feel the need to post it on the internet... It probably isn't that important.
Of course the fact that IBM was involved in the consultancy had nothing to do with the matter, right?
IBM is only interested in selling its product, guess what their product entails.
Linux is a wonderful solution, but make no mistake about why it was probably recommended in the first place.
I hate to bring down the buzz, but POS teriminals are not the kind of system that you want someone running solitare on. These are single function terminals that are high profile, and like your DVD player or microwave oven, they just have to work, every time, every day, or things get messy.
These types of applications are more suited for embedded closed RTOS software than for any desktop OS. ARM and some of the variants of it support encryption and secure operations, and can easily support the functions required for POS terminals. This would make much more sense than Linux, even if the Linux was stripped down and tuned for security and RT operations. (well, thinking of it, Embedded RTLinux setup might be a good thing too, but desktop Linux distros just are not the thing for this) The article does a good job of explaining why they chose not to use XP embedded.
Think about it, most POS systems just don't have room for a hard drive, so really need to be small and use embedded software, loaded from ROM or FLASH. And Yes, these types of terminals need, and typically do have, remote access functionality. Not often used, but its there, even in those credit card swipe machines. The thing is that its an embedded OS and software, so just is not threatened by hackers and such. That is how it should be. Keeping the OS/applications closed and secure is just one of those "job 1" things for the POS industry.
Seeing a terminal in the local pub running Windows of any kind just makes me cringe! A Treo is capable enough to do the job, just doesn't come with the big fancy touch screen.
Good to see Linux making headway, but worries me none the less that such systems are now becoming off the self, and open to hackers... yikes, well it seems that way with this article.
Support NYCountryLawyer RIAA vs People
No, POS is a completely unimportant step for Linux acceptance. There's almost zero network affects to such deployments, and nobody's going to adopt Linux because that's what their cash register runs. The main competitors are DOS and SCO UNIX -- it's not like a big profit center for Microsoft or anything.
There was a time that 90% of ATMs ran OS/2 -- didn't help OS/2 acceptance one bit.
Simple: Advertising. And, I guess, user perception. But mainly advertising.
I agree with the function/performance argument. The new ATM's do seem slower, especially in transitioning from screen to screen. But people like them better: they're more friendly! And the color screen makes the bank look better. Forget selling you on a loan or something: just the fact that when people walk down the street and see a bank's ATM's, they're bright and cheerful.
People select products based on such factors. And banking is a competitive business, like most any other.
Linux IT Consulting and Domino Development in Michigan
So a system...used by enterprises....is not an enterprise system...
Not true. You do NOT own the software under copyright law. You own the hardware and the disc the software is distributed on. All of the software you run is licensed from third parties, albeit under a very permissive license.
This is the same as with a book. You own the paper it is printed on, but the content is not yours, but you are using it with the copyright owner's permission. In-fact, that was a bad analogy on your part. The GPL is much less restrictive than normal copyright law under which the contents of a book are normally redistributed.
It doesn't matter whether you use software under the most restrictive license in the world or the most permissive, unless you write it yourself or have the copyright assigned to you, you don't own it.
This applies to BSD software too. Ownership of a piece of software implies ownership of the copyrights, which isn't, and couldn't be, granted with the GPL, or any other software license for that matter, since ownership has to be transfered through a contract, not a license.
Point of Sale systems are really not enterprise level software or whatever.
Crikey, first in the ssh story, now here.
If big businesses like Matalan rely on this software, then it is "enterprise-ready" by definition.
Seriously, "enterprise-ready" is a meaningless buzzword that is twisted to mean whatever the speaker wishes it to mean. When the proprietary ssh company was talking about openssh not being "enterprise-ready", they meant "apart from the fact that massive organisations like Cisco etc rely on it". When you are saying that thesse systems are not "enterprise-ready", you mean "apart from the fact that massive organisations like Matalan rely on it".
If there is any meaning whatsoever to the term "enterprise-ready", then these systems fit it. They cannot be simultaneously relied upon by enterprises like Cisco, Matalan, etc, and not be "enterprise-ready".
Surprising....you read Slashdot yet you have never heard of Google.
Until Linux drops the GPL I don't believe there will ever be a company or a group of individuals that are going to devote the resources required to make Linux a true complete contender on the desktop. The GPL doesn't allow for investors to recoup their money. Without a business model it's hard to break specific areas of Linux out from being someone's pet project to being a true complete addition.
.... " and "I helped my grandma setup her Linux machine, she loves it." Or even "There are several companies that have switched to Linux as their main desktops...."
You can reply to this and say, "Oh yeah, I've been running at home for
But the truth of the matter is, for the last 10 years Linux has been 5 years away from being a serious contender on the desktop. Seems kinda' bizarre, but it's true. In about 5 years Linux will be some what as complete as Windows or OSX is today, but the desktop is a moving target. As Linux nears the current target the others have moved well out in front.
There are several huge gaps in the Linux APIs for stuff such as heavy professional audio and video applications. I'm really excited for the switch for Apple from PPC to Intel. The reason for this is BSD is finally going to be getting the limelight it deserves. The BSD license allows for people to maintain their 'forks' of projects that they want to keep completely open while allowing incentives for people to try taking BSD in a totally different direction and allow them to recoup their investment.
The GPL model argues that projects are better if there are more eyes looking at everything. The proprietary model argues that that's all fine and dandy but you can't recoup an investment by giving things away for free. The BSD model allows for the people who want to remain open source may remain open source and the if someone wants to try a proprietary fork of something, go for it.
I would like to see a BSD fork of all the core Linux components and see what someone could make of it. Too bad this would be impossible to really allow happen.
Why would you run any general-purpose operating system on a point-of-sale terminal?
It's much cheaper to use an existing OS than developing a custom OS. There isn't even any reason to develop a custom OS, considering both Windows and Linux work quite well for such applications. Not to mention, you can use cheap off-the-shelf hardware and drivers instead of having to develop your own.
For that matter, why would you use an x86 CPU in a cash register?
Maybe because it's cheap and easy to develop for?
The new Windows ATMs are 3 times slower and 100 times less reliable.
They also don't look fugly, are easier to use, and probably cost less to maintain.
One wonders why the rush to abandon the old software that worked perfectly well.
Legacy custom-developed software is typically a money pit. What if all that crap is coded in Assembler for some obsolete CPU? What if you need support for modern networking protocols?
If they had to reboot, its not really an endorsement of Linux is it?
"Sounds like some people can get Linux to work in an 'enterprise environment' after all."
Who (besides Microsoft and their paid shills) have you heard arguing that Linux doesn't work in the enterprise? Linux is somewhere in the server room at any place big enough to have a server room. What people do argue is that Linux on the desktop isn't ready for the enterprise - but that's not what this is. A cash register isn't a desktop (though it might run on desktop hardware), it's a single purpose machine that's going to run one application only. Linux has been doing well in the embedded market for a long time, and that's essentially what this is.