Migrating the Enterprise from Windows to Linux

Linux in Exile has a new post on the approach an organization may take in moving to Linux. The post is here, and I recommend reading it.

I have a few comments on JH's commentary. I won't quote what he says (you should just go there and look at it), but my responses correspond to his numbering system.

1. I disagree, despite JH's experience, that video and wireless cards are a problem in migrating to Linux. If one is looking for pure FOSS approach, then yes, but otherwise, there really should not be much difficulty. But yes, an inventory of these issues is worthwhile, and when doing so, remember that replacing a video or network card is actually not that expensive .

Most likely, the lower-end computers that Microsoft is putting out of business can be resurrected with a Linux move, which will offset other costs of migration.

2. I agree with everything JH says about file formats, but I would add this: Linux does have significantly expanded, deployed, and very cool text processing capabilities that are hard to replicate in Windows (without, essentially, installing the GNU operating system on the Windows machines). It is possible to write a document that has headings, text formatting (bold, italics, etc.) that can be exported to HTML (for the web) or PDF (for printing and sending as attachments) and all that, using outlines, and numerous other features in a run of the mill Linux text editor. The text editor starts up in seconds (much faster than any word processor) and the files are plain text, and thus not mess-upable. The text files are also a fraction of the size even compared to text-based XML type files which are full of formatting codes.

Given that so much processing actually happens on web sites anyway, and users producing text are actually using a simple text editor to fill in the boxes, teaching users to use stand alone text editors would be easy and simple.

So, in considering step 2, I'd also consider the benefits of training employees to open a quick and dirty text editor for most of their day to day tasks. They already are, in email and web applications. Extend that to the interoffice memo, personal notes and to-do lists, etc.

3. All I have to say about this is, yeah. For the most part, you have two choices. Use a 200 dollar operating system on hardware that must be up to date to run that operating system to open the web page, or use a free operating system on a free computer (because you can use your older hardware) to open the web page. Duh. (That duh was not for JH. It was for all those corporate execs who still don't get this!!!)

4. Excellent idea. Actually, all enterprises should do this even if they are not moving to Linux.

5. No comment.

6. Yes, and indeed, diversify. Nothing wrong with having a windows box (preferably virtualized on a Linux server) and a couple of Mac's laying around, when all is said and done, depending on what the enterprise is. From the end user point of view, this often makes more sense than it does from the IT manager point of view, but the IT manager is not the reason the enterprise exists.

More like this

Many years ago, when I was still an MVS SysProg instead of an out-of-work UNIX SysAd, I remember seeing the early IBM thin-client products. Even then, it seemed that using a Web Browser as the main application interface was an excellent idea. It frees one up from worrying about the actual client platform. For an external client it's not appropriate to dictate to them what platform they should use, (although all too many companies and even public bodies don't recognise this). But even for an internal client, it's good to have flexibility.
There are excellent reasons for restricting the number of platforms in use: trying to support umpteen different operating systems, different types of hardware, different versions is a real pain. But being too rigid about it can be a bad idea. Maybe the graphic designeres would really benefit from using Apple. Maybe somebody else really needs an application which really is MS only. Maybe your sales people would benfit from using a mobile device.
Exceptions need to be controlled, but if you make the browser the interface to your key applications - email, time and attendence, whatever - you can have that flexibility.

Simon, I think we basically agree but for one point: The idea that exceptions must be controlled. That, in my view, is a myth perpetuated by lazy IT directors. Diversity is better. Diversity is harder.

(Within reason)

To stick up for the IT director, she is constantly under pressure to reduce costs reduce costs reduce costs. So uniformity does reduce costs if nothing else by minimizing the training required. To go the other way the business types must be educated in the value add that diversity provides, and then indicate that they are willing to pay for it. Once you understand that business types regard computing as a necessary evil so they want it to cost the least possible.
Diversity being harder implies more overall costs.

You are correct.

But the corporate culture has to change. People at meetings where these things are discussed, when someone says "It's better to have one system on every machine" need to do something other than what they do now. First, the person saying it needs to say it differently. They need to say "it's lower cost in some ways to have the same exact system on every machine, but there are costs as well. You maybe giving up important funcitonality"

And the other people at the table have to say "why? What are we losing by doing that"

I've sat on two different IT committees for major enterprises, and I've never seen those questions asked. The idea that one system deployed across all machines is pure fetish at this point.

The basic question is still unanswered:

WHY SHOULD A RANDOM ENTERPRISE MOVE TO LINUX AT ALL?

@Roman:

Did you read the article? This was discussed early in the first section.

"Most importantly, understand why you are moving to Linux. For many organizations, cost is the #1 driver. It's expensive to run a Windows environment, moreso if you also run Microsoft applications on the back-office (think MS Exchange and MS SQL Server.) Why are you making the move? If you don't have a good answer to this, you're going to have an uphill struggle the rest of the way."

Roman, I'm not sure if "random" means what you think it means. Perhaps you meant "arbitrary."

One possible answer: They don't like the taste of Kool Aid.