Comment by gizmo | Hacker News Reader

Comment by gizmo | Hacker News Reader

Parent Full thread

Dijkstra writes directly and with and air of authority, but his writing is not great. His style is pretentious and obfuscatory. Even worse, he uses his authoritative style to get away with tons of unsubstantiated assertions. Let's look at a passage from the linked lecture:

Those who want really reliable software will discover [authority!] that they must find means of avoiding the majority of bugs to start with [unsubstantiated!], and as a result the programming process will become cheaper [unsubstantiated!]. If you want more effective programmers, you will discover [authority!] that they should not waste their time debugging [repeated assertion!], they should not introduce the bugs to start with [repeated assertion!].

He could just say:

When creating reliable software debugging takes up much time. If we develop techniques that prevent these bugs from being introduced in the first place then programmers will become much more effective.

Good writing is terse and easy to understand. It should lead the reader from the premises to the conclusions. Good writing shouldn't contain pretentious detours. I don't think you should emulate Dijkstra.

I doubt Djikstra would have been Djikstra is he communicated in such a sterile way. His writing achieved the ends he sought. He's memorable and endlessly quoted which gives life to his ideas beyond what is immediately provable.

You might argue this is improper, but I think that's a bit morose. I dread the day when all people write like that second sentence.

That's not to say someone should emulate him, though. It's not that it's wrong, it's that they're bound to fail and produce something tinny and derivative. That's the thing your revised sentence has going for it: it's as cheap as it is limp.

That's the thing with blowhards. When you take away their rhetoric and literary flourishes they've got nothing left.

If something becomes boring and derivative when stated plainly then there was nothing of substance to start with.

  > Good writing is terse and easy to understand.

Good writing does what it sets out to do. "terse and easy to understand" is a fantastic style for certain objectives, but there's no one "good writing" style.

And you say you don't like Ayn Rand! Terse is just another word for unimaginative, when it comes to prose.

I think you might be blurring the lines between writing something in a socratic dialogue versus writing technical documentation. Dijkstra, obviously often is writing in the former style and so there's plenty of room for his personality in the text and it doesn't take away from the overall quality and cohesiveness of the text.

I don't think I'm blurring; if anything, I'm sharpening! Not all text is the same, and it's important to know what you're trying to accomplish when you set out to write something.

Hemmingway has a terse writing style, but few consider his prose unimaginative.

Not that I dislike your version, but aren't the claims as unsubstantiated there as they are in the original? (And I don't know how they could be anything else when someone is simply sharing their informally gained insights.)

Oh, you're right. I didn't introduce new arguments because I only wanted to change the delivery of what Dijkstra wrote and not the substance.

I also don't mind unsubstantiated claims. It's just obnoxious to phrase it like "any person with half a brain knows [controversial claim]".

> Dijkstra writes directly and with and air of authority, but his writing is not great. His style is pretentious and obfuscatory. Even worse, he uses his authoritative style to get away with tons of unsubstantiated assertions.

This bugs me as well. He's obviously made profound and exceptional contributions to the field but I don't like when he writes in the manner you pointed out where he's making bold statements without evidence. Worse is when people quote something Dijkstra wrote that he didn't back up with evidence as if it's true just because Dijkstra wrote it.

But, if Dijsktra wrote it, it is right, where as if seanwilson wrote it and cited 100 papers in Datamation it might still be wrong.

Thanks! Your statement gives me good insight and I find it helpful, and it seems to complement @jonahx's comments.

Reply on news.ycombinator.com