Slashdot Mirror
Bjarne Stroustrup On Educating Software Developers
jammag writes "Bjarne Stroustrup, creator of C++ and a professor at Texas A&M, weighs in on the problems in today's CS programs. In particular, Java (there's too much of it), the quality of graduates (companies aren't happy), and the need to balance the theoretical and the practical (long overdue). Not pulling punches, Stroustrup even talks about high schools — 'High schools could teach students to work hard at something (just about anything), to search out information as needed, and learn to express their ideas in writing and orally.' He finishes by giving advice to working developers: 'Serious programming is a team sport, brush up on your social skills. The sloppy fat geek computer genius semi-buried in a pile of pizza boxes and cola cans is a mythical creature, best buried deep, never to be seen again.'" Read on for more choice quotes from the quotable professor.
I have even had questions from strangers in airplanes: "You're a professor? In software? Have you got any students? Here's my card."
The US industry could absorb more good developers than there are currently students enrolled in IT-related programs — but not all of those programs and all of those students would qualify as "good" in this context.
The companies are complaining because they are hurting. They can't produce quality products as cheaply, as reliably, and as quickly as they would like. They correctly see a shortage of good developers as a part of the problem. What they generally don't see is that inserting a good developer into a culture designed to constrain semi-skilled programmers from doing harm is pointless because the rules/culture will constrain the new developer from doing anything significantly new and better.
The contemporary Math, Physics, and Biology books I have seen are far, far more conceptually challenging than what we present to CS and engineering students in the area of programming.
I think the ultimate aim is to make programming more of an engineering discipline, more mathematical or scientific; "craft" and "art" are both needed, but there ought to be a scientifically based core on which people can base their craft and art. Software design and implementation is more than a craft; there is more math, science, and engineering to know and apply than is customary for fields we call "crafts." Incidentally, I find it appalling that you can become a programmer with less training than it takes to become a plumber.
I have even had questions from strangers in airplanes: "You're a professor? In software? Have you got any students? Here's my card."
The US industry could absorb more good developers than there are currently students enrolled in IT-related programs — but not all of those programs and all of those students would qualify as "good" in this context.
The companies are complaining because they are hurting. They can't produce quality products as cheaply, as reliably, and as quickly as they would like. They correctly see a shortage of good developers as a part of the problem. What they generally don't see is that inserting a good developer into a culture designed to constrain semi-skilled programmers from doing harm is pointless because the rules/culture will constrain the new developer from doing anything significantly new and better.
The contemporary Math, Physics, and Biology books I have seen are far, far more conceptually challenging than what we present to CS and engineering students in the area of programming.
I think the ultimate aim is to make programming more of an engineering discipline, more mathematical or scientific; "craft" and "art" are both needed, but there ought to be a scientifically based core on which people can base their craft and art. Software design and implementation is more than a craft; there is more math, science, and engineering to know and apply than is customary for fields we call "crafts." Incidentally, I find it appalling that you can become a programmer with less training than it takes to become a plumber.
Bjarne Stroustrup, creator of C++ [...] weighs in on [...] Java (there's too much of it)
Oh, gee, now THAT is a surprise!
Coding can be a hobby, in which case sure, you can do it alone (though for significant project, it gets rough, with all of the cross cutting concerns... someone who's extremely good at algorythms may totally suck at designing a public API... fairly common in CS-heavy companies that do a lot of backend stuff).
I'm guessing though, they were referring to coding in the real world (for a living), in which case, for anything significant (no, the ecommerce website someone makes as a freelance isn't significant), you'll need to be a good teamplayer. People who are good at making functional GUIs often suck at backend programming, and vice sera. People good in architectures often miss the details. Computer scientists often cannot understand project managers and architects. Thus, it becomes a team sport (like you'd have goalers, offense, defense, etc). Being able to work as part of that team is almost as possible (sometimes moreso) than just being good at your specialty.
Yeah, but what Stroustrup is trying to say is that the formal process has to be taught, so that the latter process can work more efficiently. Now matter how much of a flash of insight I get, I'll never even code Hello World if I don't know how to program properly.
I've never known this to be the case. When I managed developers, I paid market rates which I am sure you would say are low. I never had a single offer turned down for salary considerations. I had a range of programmers from good to excellent. Expectation of higher salaries was not correlated with skills or performance at all.
You are right: programming _is_ hard. But salary is only useful as a motivator up to a point. Beyond that, what good programmers want is respect, appreciation, and freedom to do great things. They want to work on stuff that they feel good about. They also want to work with other smart people who they can learn from and build great stuff with. Those are decent starting points anyway.
So your general idea is right: that complaining companies are full of it, but I don't think it's because of money. It's because they have lousy uninspiring development environments.
Cheers.
Most of these topics can easily be researched and learned without a university.
Yes, they can be researched - but, let's face it: most people either think it's beyond them, think they'll never need to know it, or are just too lazy. The ones who would research it themselves are likely to be the ones who are good anyway.
The problem is, you're lumping all development into some generic activity called "coding." That ignores the simple fact that being an expert in C++ does not make you expert at UI design. Being an expert at UI design doesn't make you a SQL whiz. Being a SQL whiz does not make you an expert at designing flexible & scalable application systems with many moving parts.
See the pattern? To make any practically-useful system today, you MUST work with other people who are expert in areas you're very likely not an expert in. You can create the best data model in existence for your company's payroll system, but if you don't have a clear understanding of the business rules, a good architecture, and a solid UI (and I've seen very few people in software engineering who are experts at all of the pieces that go into creating a good piece of software), the resulting product will be an unusable piece of crap that doesn't meet requirements.
Stroustrop's point of the single genius programmer is valid. The model you describe supposes that a single person can possibly learn and be expert in all of the disparate knowledge areas that are required to make a working piece of software, and in any group, the people who are capable of that level of insight & knowledge will *always* be the statistical outliers. Instead of punishing the majority of people for not being born with genius-level intellect, teach them how to be good at their chosen area of expertise, and how to relate to the people they have to work with to get things done. You don't have to engage in XP or some other "pair programming" methodology to realize value from being able to communicate clearly and work with your architect, QA / Reliability engineers, your customers, your project manager, and other developers who are working on separate subsystems.
This is all well and good, and there's no doubt that an engineering/logical approach is very important in engineering, but there is something of an art about engineering as well. I can slog all day designing/building/coding/prototyping away, working from diagrams, notes and even flowcharts (though not formalized ones, I'll admit, like days gone by), but sometimes my best and most productive work are those creative flashes I get, when any kind of formal process is tossed out the door. Sure, I have to go back later and document my work so that even I can understand it, but there will always be that creative aspect to engineering, and that ain't necessarily bad.
The companies that complain about the lack of "good" developers are the ones paying their developers crap and looking to offshore/cut costs. If you want good developers you're going to have to pay them.
I work at a company that pays at least 10% more than the going rate at other companies in the state and interviews 2-3 / week. It takes 3 weeks on average to find a developer they're willing to make an offer to. Two of the companies that are complaining are Google and Microsoft, both of whom give good salaries and great benefits when compared with the average job.
I've worked at several companies that paid decent amounts of money, and between 50% and 75% of the developers are crap. In my experience, good developers are hard to find at any price.
Lack of funding? Young man, I learned Pascal on a TRS-80 Model II. The cellphone that is likely in your pocket is probably more powerful in just about every aspect. The book we learned from was probably close to a decade old at the time and was sad in the shadow of what you can get at Borders for 25 bucks. I learned programming just fine.
You don't need the latest and greatest to learn how to code. Infact, the more I see of computers and teenagers the more I think I had it better being limited to machines that didn't have a 1/10th of a meg of RAM and no Internet. Today there are just far too many distractions on a PC to get kids away from the task at hand. To be honest, if I were teaching coding I would go with the 2004 text about whatever language that normally can be found in the bargin bin for 5 bucks, a bunch of P-II 350s, a 10 dollar flash drive for each student and whatever freeware IDE I could find for the chosen language. No other software on the PC and no network connections either, thankyouverymuch.
This would keep the students mind on the task in front of him. When I first got my Vic 20 with no media storage and no game cartridges I had to sit down and learn how to make the machine do what I want it to. I found that it was a very gratifying thing to do. If I would have had MySpace and EverQuest I doubt I would have ever have gotten as far.
BTW: Get off my lawn!
Dedicated Cthulhu Cultist since 4523 BC.
I know about extreme programming, and I've really enjoyed some team programming sessions, but when it comes down to it, I think one reason some people hold a much deeper level of knowledge than their peers is that they spend extreme amounts of time alone.
...
People who are fascinated with math and coding tend to have fewer peers who can understand what they are doing. Is this a bad thing? ... I guess the challenge is to guide young people to seek out their peers (those who are fascinated with the same things), and to make friends without making everything into a competition. ... There is a place for ambition, but that instinct can be a hinderance.
I think you've misunderstood was Bjarne was talking about. Most software projects have a dozen or more developers requiring close collaboration and the ability to cope with egos and positive/negative emotions. Also, just the ability to communicate your information to someone who needs it. Damn straight modern programming is a team sport! Are you working on tiny apps in a small closet hidden somewhere inside a small company? 'Cause for the rest of us, our day is filled with interacting with people. I'm very introverted, but I do need social skills to deal with the people I work with every day -- which causes me extra stress, but I digress. Software development in big companies and FOSS projects is a team effort, and for successful projects, the humans working on those teams need to work together smoothly.
Eventually, you start talking about what Bjarne was talking about at AT&T: social skills doesn't mean your hosting dinner parties and chatting up girls. It means being able to share and communicate your idea to another very smart person, and then being able to listen and learn from that other very smart person. I think you shifted into that point at the end of your post: the idea that introverted smart people need to find other smart people they can talk with.
Social skills in this context doesn't mean being the life of the party or tossing a football with non-geeks. Social skills means being able to tell the prima donna that his code has a bug in it. It means being able to explain to others why a design or architecture is bad. It's coping with hurt feelings: yours and theirs. The basic social skills of working in a modern business environment.
Programming has evolved waaaay beyond the lone programmer in his little corner office working his own hours. A programmer is expected to talk to end users and tease out their real requirements; it requires them to be able to stand up and argue for something in a room full of managers; it requires having the ability to tell another programmer his coding style is awful, but you want to help him to get better.
That's what Bjarne meant by social skills and team programming.
The bitter lessons of a veteran coder: http://bitterprogrammer.blogspot.com
I know it can still be used that way, but it seems to be more and more difficult to only code that way using C++, because you're going to have to use libraries at some point or another.
Yeah, those damned templated libraries with their: performance, type safety, design patterns, generic programming principles. I can't imagine why Java wasn't happy with containers of Objects and added Generics. C# definitely shouldn't have followed suit. All you need is void*, size_t, and int (*)(void *, void *), right?
It's amusing seeing Strostrup whining that schools are teaching Java instead of his C++. The problem with C++ is Strostrup. He's in denial about many of the fundamental problems of C++. He's publicly stated that there's nothing major wrong with C++. If that was the case, we wouldn't need Java and C#, which are, after all, attempts to improve on C++.
Down at the bottom, the fundamental problem with C and C++ is the "pointer=array" concept. That was OK for 1978, but it didn't scale well. It's the cause of most of the buffer overflows in the world. C and C++ don't even have syntax for properly talking about the size of an array parameter. That's just broken. C and C++ need something like conformant array syntax for parameters and in other contexts where one needs to talk about the size of an array.
The second fundamental problem with C and C++ is that the programmer must obsess on "who owns what", and the language not only doesn't provide help with this, it doesn't even have syntax for talking about it. There's no distinction between a pointer that "owns" an object and one that just "uses" it. Yet if the programmer doesn't carefully make that distinction, the program will have either memory leaks or dangling pointers.
The trend in C++, since templates went in, is to try to wallpaper over the problems with the underlying model. It never quite works; the mold always seeps through the wallpaper. Container classes almost, but not quite, succeed at encapsulation. There's almost always some place where a raw pointer has to be allowed to leak out. The standards committee has gone off on a "generic programming" tangent, with emphasis on weird template features used by few and used correctly by fewer. Just because you can abuse the C++ template system as a term-rewriting engine doesn't mean you should do that in production code. This results in a whole new class of incredibly obscure compile-time bugs. The standards committee has been thrashing for over a decade in this area; in the 1990s, the new version was to be "C++9x"; now it's "C++0x", and since we're close to 2009 already, "C++1x" looks like the reality.
C++ is the only major language to have hiding ("abstraction") without memory safety. No previous language had it, and no later language repeats that mistake.
Back around 2001, I made an effort to do something about it, but the political hassle was more than I had time for.