Hacker News Reader

Hacker News

Comment by tyingq | Hacker News Reader

Full thread

tyingq

I imagine our pointy-haired bosses would appreciate an "unsuck it" for tech jargon, with stuff like:

orthagonal -> unrelated

immutable -> unchangeable

idempotent -> (is there a very short phrase that works here?)

View on HN

You could replace orthogonal and immutable perhaps, but idempotency has a rather precise technical definition that doesn't really have any synonyms. It's a useful word too, because idempotency is generally a very desirable property to have in many systems.

If you do any kind of system automation, it would get quite cumbersome if you had to keep talking about operations which only effect change when they actually need to.

Got me.

But I like to think that using precise technical terms such as "idempotent" is perfectly fine. "Immutable", works as well, since you have "mutable", and "mutation" to go with it. Both also carry a lot of contextual information that make them hard to replace.

(There has to be a proper linguistic term to describe these properties)

"Touch base" is none of those things. Ever.

I like that orthagonal suggests that they definitely intersect, but go in two different directions. Unrelated is too harsh in most of the situations you'd use orthagonal.

Ah, yeah, that's true. Though some developers use it without that context to really just mean "unrelated".

Well, orthogonal lines don't have to intersect in in geometry (you just need more than 2 dimensions), and "intersect" is meaningless in algebra. So, I don't think that's really implied.

Criticizing jargon in the field of programming because it's not used in exactly the same sense as when it's jargon in mathematics is.. weird.

Hm. "Safely repeatable"? Idempotence is a pretty specific concept.

Idempotent -> singularly consistent

It's not a perfect definition, but I don't think execs really need to know what idempotent is.

Those aren't even jargon, they're just words a well-educated person would know.

It's not about education or vocabulary. It's about using words to make yourself sound important, when you're not.

The problem is that people who don't have a good vocabulary abuse these words and change their meaning for self-serving purposes, and not to communicate.

Perhaps that's just projections of your own insecurities and those words are just simply what come to mind for them or find the most descriptive.

Perhaps that's just projections of your own insecurities and those words are just simply what come to mind for them or find the most descriptive.

No, it's because I have a degree is Communication, and understand the importance of language in the workplace.

Any "insecurity" is just defensiveness on your part.

Or, to put it in language you might understand: "Troll harder, n00b."

Ooooh the guy with the degree in communications has decided what is useful and what is over the top in regards to mid tier vocabulary usage.

I think you’ve missed the point of this whole website my friend.

I'm merely saying that if someone using "orthogonal" in a sentence seems pretentious to you, then perhaps you're the issue.

Idempotent is definitely jargon. It was coined by a mathematician and the word has no use outside of a math/cs context.

Sure, but if you're using it in the context of math/cs, then it's not really unnecessary jargon.

idempotent -> side-effect-free? safely-repeatable?

Ah, that's helpful. "Safely repeatable, with no side effects" will be my new answer if anyone asks.

Well, per others' responses the "no side effects" part is not quite accurate. The definition in plain English is actually pretty simple: idempotency = "has the same effect, no matter how many times it's run". Like a dedicated "Power On" button (vs a "power toggle" button).

> side-effect-free

Hum... No, that's not the meaning at all.

> safely-repeatable

And now you'll get stuck explaining exactly what "safe" means.

> side-effect-free?

No, those are pure.

A pure, idempotent function would be one where f(f(x)) = f(x).

An idempotent function could have side-effects. Imagine turn_on() where if it's already on, then nothing happens and it just stays on.

I think there's two different readings of "side effect free" here. The fact that nothing happens on the second+ invocation is, for me, the example of no unwanted side effects. An unwanted side effect might be sending more voltage down the line on each invocation...120, 240, 360.

Side-effect is a very well-established term, though[1]:

> In computer science, an operation, function or expression is said to have a side effect if it modifies some state variable value(s) outside its local environment, that is to say has an observable effect besides returning a value (the main effect) to the invoker of the operation.

> The fact that nothing happens on the second+ invocation is, for me, the example of no unwanted side effects.

I don't think you'll ever find a definition that says something about second invocations.

[1] https://en.wikipedia.org/wiki/Side_effect_(computer_science)

Quite right. How about "same-result-returning"?

"same-result-returning" sounds like pure functions which return the same result when given the same arguments.

I prefer entropicdrifter's "one-and-done".

idempotent -> one-and-done ?

It's kinda hard to summarize

idempotent -> same same, but different

performant -> fast enough

is already there :)

transparent -> opaque

"Hey, you're an idempotent kind of person: all I need to know about you I learned on day 1, and nothing's changed since then." /s

Does this work too?

"Hey, you're an idempotent kind of person: all you think you need to know about me you learned on day 1, and nothing's changed since then."

Reply on news.ycombinator.com