As a widely published author, I’m plugged into a number of author networks and always enjoy talking with authors of interesting books to learn more about what they write about and how they produce their material. Recently, I bumped into Maria Winslow and was intriguied by her new book, The Practical Manager’s Guide to Open Source. What’s so interesting about this book is that the focus is to explore how Linux and open source software lets you switch from an expensive Microsoft Windows based environment to a considerably less expensive Linux platform, a topic that I’m also quite interested in exploring.
Anyway, Maria and I went back and forth via email to produce the Q&A included herein. Do check out her book and if you have your own views on Windows versus Linux, please feel free to add your two cents here too!
Maria, what was your motivation for writing this book?
I advise mid-sized organizations on how to assess their IT environments to discover those areas where they can take best advantage of open source software. During the course of this work, I’ve talked with a lot of IT directors and technical managers, and I’ve found that while most believe that open source can save them money, they don’t really have a good idea of how much. And they are generally unaware of the rich selection of open source applications – not necessarily just for Linux – that are production-ready and well-supported. This book grew out of that experience, and was an effort to create a resource for companies working to adopt a practical open source strategy.
Your emphasis throughout the book is on taking a practical approach. Can
you talk a little more about this?
There is a lot of misleading information out there about open source software. The Yankee Group did a survey last year in which they asked this question, “If you were to migrate your entire environment to Linux, would it cost you money?”. The answer will obviously be yes in most cases, but it’s a useless question. The truth is that most organizations are going to find that a mix of operating systems makes more sense. Anything you might consider basic computing infrastructure – file servers, DNS servers, web proxies, etc – can be a good choice for a migration because since they are common across industries, you’re more likely to find stable, supported open source options.
But you are more likely to stay with the status quo for functions that are specific to your industry, since a smaller market makes it less likely that an open source project has developed. At the time I started writing this book, there was not a lot of information out there to help people sort this out. IT directors are naturally a practical group, and they are seeking practical advice.
So this is a book for managers, then?
Actually, I wrote it with the concerns of technical managers in mind, but I’ve found that it can be a help to the technical people who are trying to convice management to adopt more open source. I talk with a lot of people in Linux user groups around the country, and many of them would prefer to use more Linux at work. But they’re having trouble convincing their bosses. I recommend preparing a simple cost-savings analysis for their migration projects to help make their case. There is nothing like proof of savings to get management’s attention.
But yes, this book was written for managers. I step through how to take an assessment of their IT environment to look for ways to use open source software, and how to evaluate the ROI of those potential migrations. I include listings of recommended software based on hours of interviews and research – something most managers don’t have time to do. And I included eleven case studies to highlight how some of the recommended software is being used in production environments.
Why have you put such a focus on “return on investment”?
One thing I’ve discovered as a consultant is that many people think that Linux will be cheaper for them, but they aren’t sure exactly how to go about quantifying that to upper management. Bloor Research did a study a while back finding that while most organizations track IT expenditures, very few track the ROI of those expenditures. As a result, technology executives tend not to be practiced in any method. So I wanted to provide a simple methodology for quantifying what open source might mean to an organization in terms of the budget. I focus on the factors that are easily measured, like licensing fees, and keep it simple but useful. The book steps through spreadsheets that you can download and fill in to save time calculating your ROI. The formulas are already in there.
What was your selection process for the software you recommended?
I researched the most commonly used software, and only selected software that I found to be widely deployed in production settings. Commercial support is important, too, so I list where to find it for each project. I also interviewed people responsible for the code, and found out histories and future plans. I actually dropped a project from the book when I discovered that even though it was widely deployed, development had stopped, so I knew it would be a dead end. Listings are divided into three categories – server-side software, client-side software, and application development tools. It’s not all open source, but a practical mix of open source and proprietary that represents a good value for the typical IT environment.
Great stuff, thanks for sharing with us, Maria!
The Practical Manager’s Guide to Open Source by Maria Winslow
The real beauty of the original PC revolution was that end users and individual departments began to have computing options that *they* could choose from rather than only what central IT and its vendors offered them on “the big iron” and the 3270 screens and “dumb terminals”. So where are we now as far as user control over applications?
Does open source open up new opportunities for user-oriented applications, or is the emphasis on enabling IT to further tighten its control over the enterprise?
Are we at the stage where end-user application innovation is effectively dead and we’re simply discussing how to landscape the cemetary?
Do open source end-user applications offer a more appealing innovation path than sticking with Microsoft?
One “beauty” of Microsoft is that users will eventually get tons of new features regardless of whether IT planned for them or not, but with current open source promotional efforts focusing almost exclusively on “cost”, what incentive is there for significant investment in innovation? I’m thinking of Sun and OpenOffice, in particular.
Sure, there’s a debatable level of innovation in the areas of browesers and email clients, but without Microsoft pushing so hard, where would the incentive for further innovation across the broad spectrum of user-oriented applications come from absent significant financial incentives that simply aren’t there in a cost-oriented IT model.
Or, maybe corporate users are already overwhelmed with featues and more worried about losing their own jobs to “cost cutting” and the point is actually moot.
I would also note that people tend to be paid based on the degree of technical complexity that they have to cope with as well as the number of people who report to them and the size of the budget they manage, so there is a financial incentive for IT and even not-IT departments and users to want to see their environments get *more* complex rather than simpler. Maybe that’s one of the factors that makes open source and “mixed” environments so appealing.
— Jack Krupansky
Here we are discussing the “epicenter” of the “battle for the future”, a critical “fork in the road”. Open source is about “freedom” because we know for certain that code will “control” the options available to all of us in the future. If the code is proprietary, then our options will be limited by a “them” over whom we have zero influence. If the future of most code is open source, then at least the “logic” (within the code) behind the list of available options will be “readable” by people who can read code. We will have an equivalent of the “freedom of information” act that enables citizens to read documents that affect them.
Cost is best viewed over the long term. The Base Realignment and Closure (BRAC) committed is trying to save money over the long term, even though it may cost money now. Businesses will need to think more long-term before they will embrace OpenSource.
Good government is essentially “open source” — everything must be visible to the citizens. Can the concept of openness also be beneficial for business? I think so…
I’ve seen this book in Amazon listings and always wondered about it, so thanks for a good interview. I’ll have to check it out.
While I can’t speak much to the big-IT world, I have some perspective from the embedded systems side, where open source is also making a dent. In the automated tape libraries I work on, we have a mix of Linux and Windows CE. WinCE was in use before I joined the company in 2001, largely because there weren’t many options at the time for embedded OS’s that had a touchscreen GUI, and partly because management looks to Microsoft first for solutions.
When it came time to do a new big enterprise-class library, even the most pro-Microsoft of managers was leery of trusting WinCE for total control of the library. I was one of the chief proponents of Linux, and because I was doing one of the most critical parts of the software, I was told to go ahead. We used two identical controllers in the library, one Linux for critical stuff (robotics, host interfacing) and the other for the color touchscreen interface. The library will continue to function even if the WinCE controller is removed.
The experience has taught me a few things. First, Linux is not perfect. Especially at the lower levels like device drivers, not having fixed APIs creates headaches with getting the kernel and kernel modules talking right. We had a big headache trying to get GDB working with multi-threaded applications. The 2.4 kernel we’re using has some bizarre thread starvation problems, which should be resolved in 2.6 (all new scheduler), but we don’t want to jack around with a working system because we’ve already worked around it in our application code.
But, unlike Windows CE, problems get fixed. We’ve been through three major versions of CE (2, 3, and 4), and there are still totally ridiculous problems that have not been resolved, e.g. basic things in the compiler that just don’t work right. The scheduler is terrible. Debugging is painful. The tools frequently crash or hang. Even getting the OS to consistently build right is a hassle. And if you don’t like it… tough. There’s nothing you can do to change it.
Another great thing about using Linux in embedded applications is that you can still use “big boy” tools and frameworks when you need to. Want an embedded web server? Use Apache. Just strip out the parts you don’t need. Many open source projects can be built in minimal configurations that will easily run on modern single-board computers. Plus you can use tools like NFS during development that greatly speed up your process, e.g. telnet into the embedded controller, have the controller NFS mount your workstation’s development folder, and run your app straight from there.
Since our project in 2001, significant changes have occurred in the open source world that make it even more appealing. The 2.6 Linux kernel looks fantastic. Qt/Embedded brings a world-class GUI (and general OS) framework to embedded devices. There are now good books on using Linux in embedded devices. (Sure would have saved us some head-scratching.) Where Microsoft has been stagnant, at least as far as things I care about, Linux is making huge strides, far surpassing WinCE.
I think a significant problem is that some (many? most?) open source projects are terrible at marketing. That is, they don’t answer the basic questions (like this book does) about how they’ll benefit the enterprise and address the cost ramifications. For example, I asked Trolltech for competitive analysis information of using Qt/Embedded versus WinCE that I could take to my management. They said they didn’t have any such thing. WHAT? They have nothing I can take to my management, explaining why they’re better than their primary competitor (from a business point of view), and the comparative costs? That’s ridiculous, and they’re a commercial company charging big bucks for their tools.
That’s where this book does a great service to the open source community, by addressing the questions that management really cares about. I hope it’s a great success!