Are Subject specific programming books a good idea

Question

I have a friend who is starting graduate studies in a psychology. Part of what he is learning is building testing apps for trials. He purchased a book that was supposed to be vb for psychologists after reading though it I found that it was not a very good programming book and espoused what I think of as bad practices. I instructed him to buy a standard learning vb.net book from a known publisher.

So the question is there any good reason why standard methodologies of programming shouldn't be applied to certain disciplines.

Answer 1

Non-programmers don't need heavyweight methods, but they should at least follow the basics. Break down the code into cleanly-separated pieces, understand what all the pieces do, and understand how they combine. Save copies of the versions that work so you can go back to them later. Comment the code so that you understand what you did later. Figure out why a bug happens instead of just randomly changing the code until it passes a trivial test.

I don't think most software people understand just how badly non-software people code. Research today mostly requires computer analysis. The quality of the results of the analysis is heavily dependent on the quality of the code that did the analysis, so bad code to do this work is worse than none at all.

Answer 2

So the question is there any good reason why standard methodologies of programming shouldn't be applied to certain disciplines.

Books are one thing. Classes in college are another. In school, minds are open. Students have to do what they are told. After graduation, they don't.

I work with some very capable people in the areas of biostatistics, and as programmers they are very strong. They also have none of the instincts of a software developer.

I have had some luck (not much) educating some of them on the importance of

• readability
• maintainability
• source code control
• modularity

What they are doing is what they were taught. Some of them were taught Fortran, with built-in line length limits and short variable names, causing them to pack code together like raisins in a can. I worked with a well-educated scientist who treated parentheses like they were gold-plated. They view declaration of variables and static type checking as a major misinvention of young whipper-snappers, and are surprised when bugs appear because after all "Why would I put bugs in the code?"

The point is, no matter who you are, your code does not just tell the computer what to do, it tells other people (and yourself) what you did. It should allow them (and you) to know if what you did was correct. (Argument-by-prestige is entirely too prevalent.) When software gets over a couple hundred lines of code, it's going to suffer changes, and that's going to put in bugs, and you have to know how to manage that, no matter what your stated profession is.

Books may or may not be fine, but what we really need to do is teach teachers, at a pragmatic level, the basics of software engineering. After all, programming is a craft. Non-plumbers can do plumbing, but they need to know the basics of the craft.

Answer 3

The big problem here is not that the book uses bad practices (it may be better to learn fast and dive in than to learn good software design principles). The big problem is that it's probably not a good introductory text.

It's difficult for a non-programmer, somebody who lacks the specific talent to be a good programmer, to learn to program. It would be a good idea to pick up an excellent teaching book. As it happens, teaching technical things to non-technical people is hard, and most people aren't all that good at it. Therefore, a specific introductory book like VB for Psychologists or Python for Physicians or Electronics for Dogs is likely to simply be a worse book to learn from than a more general one.

Answer 4

There is actually a good reason to not talk about good practices in those books.

The book is intended to be read by psychologists. Try to talk to a psychologist (or any other person who doesn't care at all about software development) about OOP, about how to indent code properly and about version control, this person will stop listening to you in a few minutes.

It's like in The Big Bang Theory, Series 3 Episode 5, when Howard asks Bernadette if she likes computers, and she answers: "I use them. I don't like them".

People who use computers want their work to be done, with some help from a computer, but they don't want spending their time learning how computers work, how real, professional software development is done, etc. They are not geeks. They are users.

That's why books oriented to specific non-geek audience are better when they tell little or nothing about best practices. It can feel strange for us, who have years of professional experience and who think that it is impossible to work for a company where there are no source control. But that's why those books are not for us, nor must they be read by us.

---

There are lots of publications about the low quality of source code written by scientists. The problem they found is that published code must be easily verifiable by other scientists in order to be able to verify what is said in the publication.

Finally, I think best practices must be enforced for people who are not geeks, but who write software daily. This is a case of most scientists. But this is only the case of some psychologists; most of them will never write programs, or, at least, not very often, and would never care about source control or other stuff like that. For those who will actually dedicate their career to the development of applications related to psychology, I hope they may be a bit more interested in professional software development, and will read general development books which promote best practices.

Answer 5

It's about motivation, interest and getting things done.

If a practitioner in another domain can use a programming language and get their work done without utilizing all the best practices then why not? This does not mean the book should publish errors.

What are the problems we try to avoid with best practices and do they apply? Most of these people will only write programs that they are going to use for their own projects. I know, there are cases when they call one of us to clean up their mess, blah, blah, blah, be thankful you have a job. You can get away with a lot in small projects.

Answer 6

I think that the comments and some of the answers have a biased point of view (coming from those whose full profession software engineering). My background is in Economics and I can say that those whose study is something other than software engineering don't necessarily have the time to learn all the odds and ends that go into professional Software Engineering. Somebody who wants to study Psychology wants to do just that, study psychology, not dedicate a month or two (given that they could pick up relatively quickly, although I would assume most people in the Social Sciences who need to pick up on programming have little to no programming background) to learning best practices of programming/software engineering.

Personally, I can see the worth of specialty books like the one that your friend has as it helps those need to learn programming to help them with their academic focus and it lets them know what they need to know and quickly. If they want to get into more indepth material, that's fine but they can do it on their own time. Not everybody enjoys programming as we all do and I think that is something that comes into play here.

Answer 7

You can find poorly written books anywhere, but niche books are less appealing to expert authors and less likely to have serious technical review, as are small or unknown publishers.

Answer 8

The book was most likely written by a Psychologist who knew a little about programming rather then by a programmer who knew a little about Psychology.

Learning about the subject of your programs is one of the challenges of programming that is often overlooked. Too often programmers write very elegant solutions to the wrong problems.

Also keep in mind that the purpose of the book was to teach programming to a non programmer. The examples must be very simple and clear to understand. That certainly does not prevent you from using proper techniques and methods, but these often have to be introduced slowly once the basic principle of programming have been explained. Too often the author will want to show a "working" example as soon as possible without going through the extra steps to properly code the project.