This article does nothing to explain why software engineering is more tricky for toyota, other than saying that they're forced to release models faster (industry-wide problem), and a set of scare quotes around the word engineer when referring to software guys...
I'm not in the states right now (and the economist is mostly, but not entirely, written in the UK, not sure about economist online) but my sense is that there is a lot of hype in the states, and now I assume the UK, about the Toyota problems without a lot of substance to the reporting about what the actual issues are.
If anyone has found a good website/blog/mainstream article that explains the issues clearly, i'd love to read them.
I didn't think that the article was saying that software engineering is more difficult for Toyota than it is for other care makers, but rather that software engineering doesn't bring with it the 60 years of QA knowledge and process that mechanical engineering has in the Japanese auto industry, and that the lessons of those 50 years that make mechanical engineering so well refined are not readily transferred over to software engineering.
Having said that, I didn't think that Toyota's problems were strictly software related.
That's funny since the Toyota Production System is the inspiration for today's Lean Development Practices. Apparently, someone else extended those 60 years of learning into software development but didn't tell Toyota about it.
The article provided an interesting history of the Japanese auto industry, went into a fair level of detail on statistical control, then completely choked with a couple of lines of opinion on "software," as though it's an abstract substance that we can blame all our problems on (like witches, republicans, or Al Gore). It's like the author hit his or her word limit with all the history and didn't have time or space to coherently tie things together.
The author should have done some research into software quality control techniques to complement the SQC/6sigma background, such as MISRA-C (the safer subset of C used in automotive systems) and unit testing. Another page of text and a bit less whining about the good old days of the metal bashers and this could be a decent article.
P.S. Thank-you pg for creating HN. The comments on here are, without fail, better than anything on the sites referenced. The overall tone and intelligence of comments on a site are a reflection on a publication's average reader, and from the looks of this article, The Economist has nothing on Hacker News.
I would take it as such that Toyota just happened to be the first to get hit by this. The others are no less vulnerable.
As software propagates into the design of previously mechanical products, we'll get more of these. There must be lots of bugs in other car models and makes that could surface anytime.
Avionics software does keep planes in air but if you applied the same level of rigor to making automotive software nobody could afford the cars.
I'm not in the states right now (and the economist is mostly, but not entirely, written in the UK, not sure about economist online) but my sense is that there is a lot of hype in the states, and now I assume the UK, about the Toyota problems without a lot of substance to the reporting about what the actual issues are.
If anyone has found a good website/blog/mainstream article that explains the issues clearly, i'd love to read them.