"Linux Gazette...making Linux just a little more fun!"

Creating a Linux certification and training program

Certification. The very word spawns Usenet flame wars and conjures up images of inexperienced novices with a piece of paper winning employment over more experienced candidates who lack the appropriate acronyms. It reeks of centralized control and of forcing people to take exams to demonstrate their knowledge. It creates a swirling alphabet soup of acronyms - MCSE, CNE, MCP, CCIE, CLP, ATEC, LAEC, CNA, CSA, EIEIO...

Yet if Linux is to be widely accepted into the computing mainstream, I believe a certification program is essential. First, though, before I am flamed, let me state clearly that I do not believe a certification program should ever replace experience in the hiring process. A certification program is ultimately a tool, primarily for marketing and recruitment. In this article, I want to begin a discussion of why Linux needs a certification program and how such a program should be focused.

Why Linux needs a certification program

The need is out there... here is what I believe a certification program can do for Linux:

• Create industry recognition - Microsoft, Novell, Lotus and others have spent millions upon millions of dollars convincing the IT industry of the value of certification. People, especially managers, perceive that there is value and importance in certification. A Linux certification program will allow those who value certification to see that Linux "has emerged as a viable option."
• Provide an organizational path for students - Students who want to learn about Linux may not even begin to know where to start. Yes, they can read an intro book on Linux... but where do they go from there. Some people can learn well on their own with a book or CD, while others find greater benefit in instructor-led classes. But what books? What classes? Students who want to learn about NT have a nice, easy path - just follow the MCSE track - through either classes, CDs, or books. The certification program takes away the pain of "where do I begin?"
• Provide an organizational mechanism for training centers - Training centers love Microsoft's MCSE program. Instead of getting students in for one class here and there, training centers can fashion a program that entices students to sign up for a block of 6 classes at a time. Typically, an MCSE program can generate $8-10,000 in revenue for a training center. If training centers want to teach Linux, how do they begin offering classes? A certification program allows a center to provide a path for training that can generate significant revenue. A common training/certification program also allows training centers to compete with each other.
• Enhance marketing - Every training center and every book that focuses on the path to certification creates more marketing about the product. A training center in the northeast will spend $1.4 million dollars this year promoting Microsoft certification. Every advertisement for this training center beats the drum for Microsoft and helps promote NT and the value of certification. When you go to a book store, how many books do you see on the shelves for Microsoft or Novell certification? Each book creates more marketing opportunities for publishers, bookstores, etc. A Linux certification program will help promote the overall growth of the Linux operating system.
• Counter the "no-support" argument - Opponents are quick to slam Linux for a perceived "lack of support." We in the Linux community know the truth... about the support from newsgroups and mailing lists... but corporate management is looking for some way to support IT products. New programs from Red Hat and others that allow corporations to purchase support contracts are certainly one great step in this direction. But the existence of "certified" individuals is another. Microsoft is quick to point out that there are now hundreds of thousands of MCSEs out there who can support Windows NT. This argument provides a certain degree of security to a manager about to plunge into NT - he or she knows that there are people out there who can support the product which is about to be used. A pool of Linux-certified professionals would counter this argument.
• Turn students into advocates - If students learn all about Linux and how to install, configure and use the operating system, they will then become advocates for Linux as they move within the IT industry. This is both due to the knowledge learned during the course of preparing for certification... and also due to the fact that since a certification candidate has invested a serious amount of time, energy and money into the product, they want to use that actual product. An MCSE candidate may spend $10,000 and six months studying to become an MCSE. At the end of that time, they have a vested interest in working with NT. People recommend what they know. We need them to know Linux!
• Provide other means of employment for Linux-skilled individuals - A certification program creates many opportunities for employment for people knowledgeable about Linux. Books can be written. Courses can be developed and taught. There is a whole industry out there teaching and developing certification courses and books for Microsoft, Novell, Lotus and others that is looking for new products. Each person who can be employed writing or teaching about Linux becomes yet another advocate, potentially full-time, for Linux.
• Recruit new Linux users - How many people have responded to ads in the paper along the lines of "Want to earn $70,000 per year? Come to xxx training and join our MCSE program."? How many of those individuals might never have entered the computer industry? Now they are employed... through marketing and by providing a training path. The development and acceptance of such a certification program for Linux will keep expanding the user base of Linux users.
• Assist in the hiring process - This is the controversial part. Flame wars will rage about people who were passed over from employment/promotions/projects because they didn't have the appropriate certifications. Other arguments will ensue about "certified" individuals who passed exams but had little practical experience. Novell's program at one point became so diluted that you could find "paper CNEs" who had become a CNE through reading books without ever touching a NetWare server! While Microsoft has tried to make their exams tough enough that you must really use the systems to pass the exams, there are certainly large enough quantities of study materials out there that someone could try to get by purely on studying.
  But certifications can assist in the hiring process. When I evaluate candidates for a position, I do not decide solely based on what acronyms they may tag after their name. Yet knowing that a candidate has passed a certification program allows me to know that they have some base level of product knowledge. It also shows a candidates persistence and determination to complete a project (the certification process). Both factors help me evaluate what a candidate is all about and how they can work.
  I have had people who want to use Linux ask me, "how do I find someone who I can hire to support this Linux server? If I don't really know much about Linux myself, how can I know if this technical person I'm hiring does, in fact, know anything about Linux?" A certification program can assist in helping managers know that a person has at least a certain level of Linux product knowledge and capabilities.

Creating a certification program

So if a certification program can benefit the Linux community, what are the characteristics of a program that will work?

• It must be exam-based. - Most vendors' certification ch as Linux International are not (yet) set up to handle such a task. LI would be an obvious choice to run such a program. Perhaps a group of Linux distributors could contribute financially to LI to set up and staff an education division.... or perhaps the funds come from training centers...
• The program must have multiple levels. - Training centers love Microsoft's program because there are several different paths for students. After one exam, a candidate becomes a "Microsoft Certified Professional" (MCP). After a certain three exams, you can become an MCP+Internet. After six exams (4 core exams and 2 electives chosen out of a pool of a dozen or so exams), you can become a Microsoft Certified System Engineer. Another three (for a total of nine) brings you to an MCSE+Internet. Students find this useful in that they achieve certain certifications as they progress. Training centers love the program because they can sell lots of classes. Publishers love the program because they can create and sell lots of books. For a Linux certification program to work, it, too, must have a series of exams.
• The exams must be difficult. - If it's easy to become a Certified Linux Professional (or whatever title evolves), than there is no value. It must be challenging to obtain the certification or the title will become meaningless. One of the toughest credentials currently is the Cisco Certified Internetworking Engineer. The CCIE involves a single Sylvan Prometric test, but then a 2-day hands-on exam at Cisco's site where you are put in a room with Cisco equipment and told to build a network. At the end of the day they send in an engineer to break what you built and your second day is spending finding the errors. You have only your knowledge and the product documentation. The problem of making it too difficult, however, is that it limits the number of people who can obtain the certification and therefore the certification does not become widely recognized. Linux needs to strike a balance.
• Training centers must become "certified" to offer Linux certification. - In a very competitive market, training centers are seeking to differentiate themselves from their competition. Becoming a Linux Certified Training Center (or some such wording) would be a marketing point for training centers. However, someone (that central certifying authority again!) must "authorize" those training centers... if anyone can be a LCTC (my example acronym) that why should a center do it? As an example, to become a Microsoft Certified Technical Education Center (CTEC, formerly "ATEC"), a training center must pay Microsoft several thousand dollars (per site), agree to stringent hardware requirements, and maintain at least two MCSEs directly affiliated with the center. Additionally, Microsoft approves which centers can become CTECs and does limit the number within a region. Fees from Linux certified centers could go to fund the central certifying authority.
• Instructors need to be certified. - If someone has the knowledge on a particular subject, does that necessarily mean they can teach it? Most certification programs have some mechanism to specify who can teach certification classes at authorized training centers. Microsoft has the Microsoft Certified Trainer (MCT). Novell has the Certified Novell Instructor (CNI). The Information Technology Training Association (ITTA - an organization representing companies in the computer training industry) has been promoting the Certified Technical Trainer (CTT) credential. In all these cases, candidates must demonstrate abilities to effectively communicate information and some knowledge of adult learning techniques. The CTT even involves a videotape of the candidate being reviewed by other senior instructors.
• Courseware must be developed. - Certainly if a certification program is to be developed, some entity out there needs to write some courseware that training centers can use. Publishers must also create books that candidates can use for self-study. Does this courseware need to be "authorized" as it is in other certification programs? I'm not sure that it does...

Moving forward

If you agree that a certification program can be beneficial for the growth of Linux, how do we as a community go about addressing the points I made above about creating a certification program? Do we create another mailing-list or newsgroup? (Does such a group or list already exist? If so, I have so far failed to find it.) Do we meet at a conference?

Is a Linux certification program even possible? I have felt the UNIX/Linux community (of which I count myself a member) has historically been extremely resistant to anything remotely resembling certification... are we still? Such a program by its nature involves a certain degree of central coordination that has not necessarily been a part of our ethic? Can we create such a program?

Who should be players in this project? Which vendors are interested? (Would Caldera be interested in joining with others to create a common program?) What is the role of Linux International? What about other efforts within SAGE or USENIX?

I don't necessarily have the answers - but I would like to participate in the discussion. If someone can suggest the appropriate forum in which this discussion should take place (or is currently taking place!), please let me know. Also to that end, I'd like to propose creating a "Linux Training Alliance" consisting of other training centers and/or instructors who are interested in creating a Linux training and certification program. Please email me at if you believe such a program would have value and would like to participate.

I believe that a certification program can accelerate the growth of Linux into the computing mainstream, while creating new advocates for Linux, expanding marketing and countering arguments about the lack of support. Can we build a certification program to do this? What do you think?