The problem exists in all areas of electronics, from EDA through manufacturing to system developers, and in most geographical markets. It regularly crops up in conversations and seems to be part of a general trend away from Science/Engineering/Technology in Western countries.
A quick browse around some recent publications shows that a lot of work is going into analysing the problem, but nobody has yet come up with any serious suggestions as to how it can be solved.
For example, IEEE USA’s today’s engineer ran a long piece in February about the need for embedded engineers by Mike Anderson, with the sub-title Why the United States is losing its edge in embedded systems…
His argument is that a lack of training for embedded engineers is forcing them to learn while doing, resulting in poor quality products. There are few degree courses in embedded engineering. Electrical /electronics and computer engineering graduates have the understanding of computer architecture, and he feels that they can be brought up to speed on software in a matter of months. However Anderson says that computer science courses are now more like information technology. The focus on databases, web design and a smattering of java prepares the students for jobs that are commonly off-shored. There is no understanding of architecture or assembly code, and the approach seems to teach that memory and processing cycles are infinite so everything can be executed in virtual machines. (And even these imperfect courses are failing to deliver. In June the Press Association reported that “The Computing Research Association’s annual survey of universities with Ph.D.- granting programs found a 20 percent drop this year in students completing bachelors degrees in professional IT fields, continuing a trend seen for several years. Enrollment in undergraduate degree programs in computer sciences is more than 50 percent lower than it was five years ago, the group said. Between 2005-2006 and 2006-2007 the number of new students declaring computer sciences as a major fell 43 percent to 8,021.”)
Anderson’s research suggests that, as well as misunderstanding the market, colleges do not understand how an embedded lab can be quickly and cheaply established and claim that there are few, if any, appropriate textbooks. He calls on the relevant professional bodies (IEEE and ACM) and industry leaders, such as Intel, FreeScale, TI, etc, to work with the educational community to develop courses on embedded engineering. He also feels that there needs to be more work to encourage youngsters to enter engineering courses generally.
In Europe, the European Union has established a project (COSINE – Coordinating Strategies for Embedded Systems in the European Research Area. Snappy title.) to investigate strategies for research in embedded engineering. The project recently had a workshop on embedded systems engineering and training, with the aim of developing a positioning paper on embedded systems education in Europe. (A further workshop, Strategies for improving European ES education and training is planned for April 2009). The 2008 workshop reviewed education for embedded systems across several European countries and found similar problems to those identified in the US. Some countries had a few masters courses, while Israel appears to have a bachelors. Sweden has recently completed a program like the one that Anderson suggested for the US, where industry and universities cooperated on identifying the needs for embedded education, but it was not clear whether this has translated into actual courses yet.
In Britain, Feabhas, a specialist company in embedded training, has analyzed the degree courses available. There are only two universities in the UK offering dedicated embedded systems undergraduate courses. A small number offer embedded systems technologies as an optional module for final year students, and a larger (but still tiny) number of universities offer combined electronics/software courses.
While embedded development tool manufacturers offer training in using their products, both software and hardware, it is only training, and if it is going to be useful, the engineer requires a wider understanding of the subject to make best use of the tools. To bridge the gap, Feabhas itself offers a family of training courses specifically geared to build the foundation and advanced skills that make the tool-use valuable.
It seems that there is a distinct lag between changes in the way that embedded systems are developed, with software assuming a more and more important role, and perceptions within universities. In part, this can be blamed on the fractured nature of university departments. Hardware/electronics and software/IT/computer science are normally in separate silos, and even though individuals may bridge the gulf between them, a new course can be difficult to set up.
Rob Williams, of the University of the West Of England, in Bristol, has been discussing the problems of recruitment for real-time and embedded computing, and they have established a financial sponsorship scheme for students. But he says, “So far, the promise of lots of money does not seem to shift kids’ gaze away from fluffy degree titles (Games Tech., Creative Whatever, Music Systems, Multi-Media Stuff). During Open Days [for prospective students to learn about courses] we easily convince the parents, but their sulky brood continue to scan the floor for easy options and a rapid way out. In truth, we scare them with the promise of hard work and difficult concepts. To us, they lack personal ambition, technical curiosity and self-confidence.”
In an attempt to reverse this situation, Williams and the University have been engaging with schools. They have been running twice annual CST Hands-on Workshops for Schools. These involve a range of practical sessions in programming, electronics and systems build & boot. So far, they have “processed” about 1000 kids through these events, which are very popular and oversubscribed.
In addition, last year they started the RCX Buggy Programming Road Show, which uses the Lego Mindstorm kit to build and control buggies. “This came about because we assisted with the regional First Lego League competition. It so clearly enthused the kids that we immediately grabbed 10 RCX buggies from different sources and set out to use them in local schools,” said Williams.
In England and Wales, university admission is based on a national set of examinations called A (for advanced) level General Certificate of Secondary Education. There are A level courses in electronics, but only a very small number of students study this. The A level computing courses were recently, unkindly, described as being geared to a mastery of Microsoft products, and even their strongest supporter would accept that that they are mainly IT oriented.
Against this background, Williams says, “It [is] really important to support teachers in their struggle to retain some practical programming activity, which has unfortunately been removed from many CS A-level syllabii. On a different level, the Linux Boot Camp has been a great success with hackers.”
“I now feel certain, however, that the real problem at the university level emerges from their funding model. They [universities] are encouraged to offer attractive degrees for mass recruitment and to ignore the vocational and graduation issues which then inevitably follow. From friends and relatives I have heard stories of their children graduating from good universities with fairly worthless skills. They then have to pay out large sums of money for sensible training in order to gain a job.
“What we need is a change from the top. Until then, we can only attempt to mitigate the unsustainable skills mismatch within our own limited powers.”
It is not as though there are no growth prospects for embedded engineers. Lisa Su, chief technology officer at Freescale Semiconductor, is reported as saying, “Today there are about 150 embedded microprocessors around the home … plus there are another 40 or 50 in your car. We see that trend accelerating, and we predict that there will be over 1,000 embedded devices per person by 2015.” While the silicon will be relatively easy to manufacture, all these devices will need systems to be designed, implemented, tested, integrated and maintained by people who combine understanding of software and hardware – embedded engineers.
Where are these people going to be coming from? What should we be doing now, not only to meet today’s demands but also to build the workforce of tomorrow? Should we be going along the road that Siemens is reported to be traveling in Germany – introducing engineering to kindergarten/pre-school children? Does it matter: perhaps instead companies should be off-shoring development to countries where there appear to be pools of talent? Please leave your suggestions on the comments board.
Footnote: Just as this was being written, a study in Britain looked at A level exams (those sat at around 17 or 18 years old and whose results are used for university entrance). Those for science and math were evaluated as being significantly harder than those for media studies and drama. Schools concerned to make their exam results look good are likely to be steering even able students to the easier subjects. Couple this with a lack of qualified teachers, and there is a developing gulf between the needs of our engineering-based real world and the world inhabited by the educational community.