Several months ago, I had the idea to contact as many Linux consultants as possible to find out how they are using Linux for their clients. Drawing on my own experiences, I came up with a list of 12 pointed but open-ended questions that would collect opinions about and uses of Linux from people who work with it on a daily basis. I constructed a Web site with the 12 questions set up in an on-line interview format. The responses were submitted directly to a database for statistical analysis and convenient storage.

Anyone who has worked with Linux for a few years probably will not be surprised by most of the results of this on-line interview. For the questions that are objective or statistically useful, summary data is given below. For the subjective/opinion type questions, I offer summaries of the majority and minority opinions, along with a few interesting quotes from both sides.

The On-Line Interview

1. How long, in years, have you worked with Linux?

2. What other OSes do you work with?

3. What is your opinion of Linux as a Server OS?

4. What is your opinion of Linux as a Desktop OS?

5. What is your opinion of Linux as a Data Center OS?

6. What, in your opinion, are the major strengths of Linux?

7. What, in your opinion, are the weaknesses of Linux?

8. Of all other OSes with which you work, or have experience, how does Linux really compare?

9. Would you on behalf of your clients, or your clients themselves, use Linux again for new services and/or infrastructure?

10. Have you or your clients had instances where Linux was not a good fit?

11. Which OS did you recommend in its place and why?

12. What types of applications do you use Linux for in your work? For example: public web site or other Internet service, intranet site or internal web-based application, other internal application, file/print services, directory services, network infrastructure, etc.

To be perfectly honest, I was shocked by some of the responses I received to these questions. I have my own prejudices about Linux and open-source programs, as do many people in the Linux consultant realm. I tend to believe that Linux should be used wherever and whenever possible. So far, in my 10 years of experience with Linux, I have found little that it cannot do.

There were 104 total respondents to the on-line interview. The data that can be represented statistically is given below.

Question 1, years of experience with Linux: Minimum, 2; Maximum, 15; Average, 8. I am impressed with the amount of Linux experience claimed by the consultants. Seventy-five per cent of them claim to have five or more years of experience with Linux, and 25% claim 10 or more. If you consider that 10 years ago, Windows 95 and Windows NT 4.0 were new and making daily headlines, those of us who were working with Linux were seen as renegades.

In my own experience, I was told that Linux was not the corporate standard, and I must get my server off of the network. That was in the fall of 1995. I didn't remove it. My response was that corporate standards change over time, and I thought we should consider this new OS. I was labeled as a renegade. That was and is okay with me. Later that year, I started the local Linux User's Group (LUG) with the assistance of the staff at SSC, the parent company of Linux Journal. Like many of you, I never have looked back in regret at my plunge into the world of Linux.

Question 2, other OSes with which the consultants works (number of respondents is given after the OS):

In the world of consulting, it is difficult if not impossible to limit yourself to one operating system or platform. The following two questions were oriented toward this reality of the business. Would you on behalf of your clients, or your clients themselves, use Linux again for new services and/or infrastructure? (Question 9): 100% of the respondents said yes.

It is not surprising that everyone would, given the opportunity, deploy more Linux in a client's network. The next question was the true test of the consultant's honesty and desire to give the client what is best for them. Have you or your clients had instances where Linux was not a good fit? In response, 85% said yes to this question. I am a little surprised at the 15% who said no. As much as I personally hate to admit to it, I have had instances where using Linux did not make sense for a client. We, as consultants, must make an effort to do what is best for the client, and we have an obligation to do so even if it means deploying something other than Linux.

As a follow up, question 11 asked "Which OS did you recommend in its place and why?" Of the responses, 61% said Windows, 13% said Solaris, 11% said Mac OSX and the rest mention some other type of UNIX or UNIX variant.

Here is the breakdown of countries represented in the interview, followed by the number of respondents from each:

Questions 3, 4 and 5 have to do with where Linux is placed in the corporate network. I asked for opinions on Linux as a server OS, a desktop OS and as a data center OS. It was not feasible for me to go through each response by hand to count the positive versus negative responses, so I summarize them for you here. Question 3 asked for opinions of Linux as a server OS, and 100% of the respondents gave Linux very high marks.

There is, however, a call for more support from vendors to provide drivers and support for Linux. Linux has proven itself and is no longer a niche OS. Almost every respondent mentioned stability and lack of required reboots as a major point in Linux's favor. The only negative comment about Linux as a server OS stated it is not as good as Windows for file and print services. My only comment on that point is I have set up many Linux file/print servers and have had no negative issues. In fact, at one location, the Linux server is set up as a primary domain controller and provides log-on scripts, file and print services, mail and DNS. If Samba is configured properly, one should have no problems surpassing the stability and speed of Windows file/print services.

I put in Question 4 in hopes of sparking a bit of debate amongst the consultants responding to the interview. Sadly, I got none. Question 4 asked for opinions concerning Linux as a desktop OS. Most respondents give desktop Linux high marks but admit that it is not ready for prime time or the average user. Many cite lack of desktop-oriented applications and vendor support as the Linux desktop's biggest obstacles to widespread adoption. Some also mention that Linux doesn't play a lot of commercial games as well as Windows does.

I completely agree that Linux isn't ready for the desktop, and to me that isn't a problem. Who says that Linux has to be a desktop OS at all? I am perfectly satisfied with Linux as a server and data center OS. At some point in the near future, desktop computers probably will have only an embedded OS that will be inconsequential, as all applications will reside remotely.

This brings me to the next question, question 5, regarding Linux as a data center OS to support such things as major user loads and mission-critical applications. Many respondents did not have appropriate experience to speak about Linux in the data center, and I appreciate their honesty. Of those who were able to respond to the question, all were positive. Many said that they run large implementations of Linux for mission-critical applications or for ISPs. Someone even mentioned Google in their answer as an example of Linux in the data center. My employer supports a client with a major implementation of Linux running MySQL in the data center and is very successful in doing so. Their story can be found on the MySQL Web site as a success story in lowering total cost of ownership (TCO) by 80%.

Questions 6 and 7 refer to the strengths and weaknesses of Linux. For the respondents, reliability, stability, low cost, flexibility and security top the list of strengths. On the other side, lack of applications, lack of vendor support, training issues and its reputation as difficult to use mark Linux's weaknesses. Many consultants and enthusiasts alike want Linux to be all things to all people, and I simply don't think this is either necessary or possible. I do, however, wish that hardware vendors would supply Linux drivers for their products and that software vendors would expand their horizons into applications for Linux.

Question 8 asks, "Of all other OSes with which you work or have experience, how does Linux really compare?" All respondents give Linux very favorable marks against other operating systems. I feel that everyone was honest here, as they did list the oft-quoted shortcomings of marketing and vendor support in their responses.

Finally, Question 12 asks "What types of applications do you use Linux for in your work?" Most of the respondents cited some sort of Internet services, such as Web, DNS and mail. Many other named CRM, database and file/print services as their major applications for Linux.

The responses I received to this on-line survey confirm my opinions that Linux is well suited for Internet and intranet type services, such as those previously mentioned. Linux excels in providing a multiuser environment for shell, FTP, Web, application and file/print service users. As do most of the respondents, I feel that Linux falls short on the desktop, due to lack of vendor support for hardware drivers and end-user applications. Overall, I am impressed by the amount of experience claimed by the respondents and their long-term dedication to this once-criticized orphan operating system.

We must take a strong look at where we, as a community, want Linux to go and what we want it to do for us. In the future, data centers will play the key role in computing. Terminal services and end-user applications as well as Web applications will run from the data center. The desktop and the desktop wars will become a memory. In my opinion, control of the data center, not the desktop, is what we should be seeking as a worldwide Linux community.

I would like to thank everyone who responded to the on-line interview. Thanks also to Linux Journal for posting and reposting the announcement of the on-line interview for this article.