Background. Recently, two of our stronger graduate students had trouble finding work as software developers. Both students had done masters-level work in operating systems, networking, software engineering, and various "hot" technologies, including Java, C++, and client-server computing. Both students had also worked as programmers before starting grad school. Given these credentials, I was surprised when these students, in effect, had to move closer to prospective employers before their resumes were given serious consideration.

My students’ experiences caused me to conduct an informal survey about the perceived value of a computer science degree. In October 1998, I asked employees from 23 exhibitors at OOPSLA ‘98—software, consulting, and training firms—to assess what their company expected from any applicant with a BS or MS in computer science. What follows are my findings, together with concluding observations as to what these findings might signify.

Wanted: "real" experience, communications skills, and insight. Representatives from 15 firms named "real" experience—experience beyond what college ordinarily provides—as a key requirement for any new hire. For people from seven firms, real experience meant anywhere from two to five years of full-time employment. Others looked for software-related internships, coops, and volunteer work; successful projects implemented on behalf of "real" clients; participation in external student competitions; published papers; honors projects; undergraduate theses; and other "interesting" projects developed outside of the classroom. Several people stressed experience with iterative development. Others stressed demeanor. As one person said, "Lots of times, kids out of college are easily intimidated. Developers don’t like their time wasted; kids get eaten alive."

The two other recurring requirements were communications skills—cited by employees from nine firms–and problem-solving ability, or mental flexibility—cited by employees from eight. Here, communications skills covers a wide range of abilities, including

• strong written and spoken English
• the ability to help customers understand their requirements
• the ability to ask questions—for example, in an interview situation
• the ability to respond to questions quickly and concisely
• the ability to sell oneself
• the ability to create clear metaphors that communicate a system’s purpose

People cited a strong background in operating systems, high grades, and a broad knowledge base that extended well beyond computer science as evidence of mental flexibility.

No consensus regarding "proper" CS training. There was no consensus regarding the "best" sort of training for a prospective employee. Employees from five firms described a need for broad background in computer science and software engineering. Individuals from two of these firms specifically mentioned knowledge of formal methods. One observed that this emphasis on formal methods is far more common in Europe than America.

Employees from eight firms named specific technologies—technologies that, as a rule, mirrored their firms’ marketing needs. The two most common were Java (4 firms) and Smalltalk (3 firms). Three people added that they would consider C++ experience as a negative, given the common practice of using C++ as a non-OO language.

Employees from two firms also described special arrangements with selected universities, where prospective employees were trained in specific products. One firm hired students that completed special projects for their company under the supervision of a particular professor. The other hired students who had taken courses that taught the use of their company’s software.

Special needs. Representatives from three consulting firms stated that a willingness to travel or relocate was a prerequisite for being hired. One, for example, observed that "people [who work for my company] need to understand travel, and know what that means—week after week."

Three people mentioned character as a specific concern. Two observed that applicants should demonstrate persistence—e.g., by sending thank-you letters after interviews. The third stated that his firm, above all, required that applicants demonstrate character and motivation, since these were skills that could not be taught.

Coding skills were, at best, an incidental concern for most of those interviewed: e.g., "The least important skill [for those who join my firm] is knowing how to produce a piece of code." People from four firms, however, did cite coding skills as their key requirement. One said that his firm intended to hire 1,000 new coders from the class of 1999. A second worked for a consulting firm that placed programmers with clients. The third person’s employer needed coders to develop a specific product line. The person from the remaining firm stated that her company routinely hired good people with non-computer science backgrounds, and retrained them as coders.

Criticisms of student applicants and academic programs. Several of those interviewed complained about unrealistic student expectations regarding starting salaries and initial responsibilities. One senior developer, for example, observed that civil architects work for years before attaining partnership—"and why should things be any different at a firm that creates software architectures?"

Others complained about students with weak backgrounds. Specific weaknesses included a failure to appreciate the needs of businesses; an inability to engage in teamwork; and a failure to have mastered basic computer concepts, like fourth normal form and polymorphism. Two individuals from a business with nearby major university noted that roughly five out of every seven computer science graduates of that program—BS and MS candidates—fail to survive a first interview.

For me, the most interesting criticism of academic programs was that class projects, by and large, were "toy" projects that, in one person’s words, were "probably not [even] worth putting on a resume." Exceptions were made for software engineering projects that extended over multiple semesters, and involved work with outside clients.

Conclusions. The observations collected here come from people who work for interesting firms—businesses that offer the kind of work that an aspiring software craftsperson might hope to do. These comments on the need for students to develop a portfolio of experience, a strong command of written and spoken English, and general problem-solving skills need to be communicated to students who wish to do similar work. Having worked in industry for eight years, I have mixed feelings about the notion that industrial experience is inherently superior to "normal" course work. What matters most, however, are the perceptions of those who hire. I also intend to advise students that the market for non-entry-level software professionals appears to be more competitive than the trade publications suggest. Finally, I intend to tell students that different firms have radically different corporate cultures and needs, and that it is good to interview with several firms, if possible, before selecting a first job.

Acknowledgments. This report was made possible, in part, by an educator’s scholarship from ACM SIGPLAN. Thanks are extended to Stan Agerholm (IFAD), Dustin Baxter (Unity Software), Joe Bedits (Sysnetics), Mike Carlo (Quality Systems and Software), Mike Clemente (ETR Technology Center), Mark Foulkrod (SilverMark), Ted Gaunt (Trilogy Development Group), Mike Gizzi (Cayenne Software), Jan Greening (American Management Systems), Ken Greenlee (Kenetiks), Arthur Hu (Raima Corp.), Dan Lesage (Unity Software), Peter Luckey (Rational Software), Scott McBride (Popkin Software), Sheila McIntyre (Synergy Systems), Joanne Martin (Applied Reasoning), Kathi Melendez (ObjectShare), Joaquin Miller (SHL/MCI Systemhouse), Dave Remmel (Cayenne Software), Joe Rubelcava (Select Software Tools), Jason Russ (Software Architects), Mike Silverstein (SilverMark), Michael Taylor (Instantiations), and Paul White (The Object People) for making time to describe their companies’ needs. Finally, thanks are extended to other individuals at Harlequin, Oblog Software, ObjecTime, Quality Systems and Software, and SHL/MCI Systemhouse whose names I have forgotten, but whose observations were remembered.