Hacker News Reader

Comment by de_watcher | Hacker News Reader

It is practically impossible to teach good programming to students that have had a prior exposure to BASIC: as potential programmers they are mentally mutilated beyond hope of regeneration. - Dijkstra, "How do we tell truths that might hurt?" (1975)

The author, Traister, has came from BASIC. We have the explanation.

View on HN

Well that is a dumb and inaccurate statement if ever I've read one. I mean I'm sure there might have been some nugget of truth back in the 70s in that it might have been frustrating to teach C or FORTRAN to BASIC programmers but it has as much relevance these days as any of the other memes people like to quite from yesteryear. I certainly managed the transition from BASIC to Pascal with ease and definitely follow good programming idioms. I also know a lot of others followed similar path of learning to myself.

I teach programming, and I find people rarely have any special trouble in switching to a different style or paradigm. If I was teaching in the 80's and wanted to transition someone from old school BASIC to something like C or Pascal with structured programming and blocks, I'd show them an "if" with a code block and how it's different than a conditional GOTO, yet similar in a way. Gradually the mental model evolves to incorporate the new features, and when they look back, they can understand both ways, even though on paper it's "two different paradigms."

Similarly, if I'm teaching JavaScript and want to show the student functional programming, I'll show them how a "for" loop could be converted to use map or filter, or recursion. Far from "corrupting their mind" so they can't understand other ways, the imperative style serves as a foundation to learn other styles. The paradigm shift occurs gradually on some subconscious level, with the previous paradigm serving as a mental model to build up the new one.

I think there’s some truth lurking in there somewhere. I was an Oracle DBA for years, and recently started working on a NoSQL serverless project. I think I’ve been doing a pretty good job of it, but it took a huge amount of effort to get my mind around the new principles. The reason being just about everything I’d learned about RDBMS was a total footgun when it came to NoSQL design. My low level knowledge of how RDBMS operate helped me get there in the end, but the initial designs I came up with were packed full of despair and sadness.

The fact you got there in the end disproves your point. You'd expect any first attempt at a new language, let alone paradigm, to be less than perfect. You can't expect people to be experts before they've even studied a subject. But to argue that any one particular language damages beyond repair ones ability to learn another is simply elitest hyperbole.

I think it was Alan Kay who said something like "arrogance in computing is measured in nano-Dijkstras", so I agree that there's a nugget of truth packed with a large amount of elitist hyperbole.

The nugget of truth is that the paradigm you start with (and the longer you stick with it) influences your problem-solving process, and shapes your way of thinking in ways we don't necessarily notice. If you have a bunch of items and want all of them to have some operation done on them, do you think "I will map the operation over the items" or do you think "for each item I will do this operation on item[i]" or do you think "return the set of items which has this thing done"? Very different paths to the same result, and which path is best is extremely context-sensitive.

The mind isn't so inflexible that it can only be shaped by the first language you learn. There are plenty of games developers who started out on BASIC. I've written file systems and I started out on BASIC. Granted some people. Might not have progressed much beyond BASIC but can we not judge those individuals on their own technical competence rather than blaming a language for corrupting them? You wouldn't say "web development destroys ones ability to write back end systems" despite HTML / CSS et al being even further removed from systems languages than BASIC. However equally there are some people who feel more comfortable in markup than writing Java (for example).

It's very easy to blame the language when often it's the teacher or the individual (or a combination of the two) who struggles rather than the paradigm.

I said there was some truth to it. Its obviously an absurd overstatement, but my point was my RDBMS experience made a fair few things more difficult. I initially designed myself into a few corners by applying my experience to solve problems in ways that ended up being hilariously wrong.

Well, we're talking about the early nineties.

Contemporary VBA developers (or whatever it is about) can point their pitchforks vertically, ok?

I also thought of that quote while reading the article, yet I disagree with Dijkstra.

I think an appropriately skilled developer would be able to write a useful application in BASIC in the same way that the same programmer would be able to write something useful in assembler.

I wouldn't want to write a large program in a language like that though, but I do think in the 80s programmers had to make do with what they had

Heh, I've started with machine code, it's not unlike BASIC. Command addresses instead of line numbers, gotos aplenty, global variables.

Reply on news.ycombinator.com