I took his point to be that things like "feels readable" and "feels fast" and to do with feelings rather than actual quantifiable things that could reasonably be compared between languages.
So it isn't that those things aren't important; just that they are subjective
I'd argue that ease of use and readability are perfectly quantifiable - there are thousands of people whose job is precisely optimising usability for many products. Sure, dealing with people is harder and less clear cut than measuring cycles in a processor - there's a lot of variability, context based on previous experience, etc. But that's no reason to ignore the human factor of programming and gather some metrics/statistics, then seeing where you can improve.
What those features are, rather than subjective, it's contextual, and I feel that behind the author's phrase might lie the usual complain by functional programmers that non-functional languages just feel foreign because people are usually taught other styles during their learning process.
As an explanation, that might be partly true, but it being true doesn't really save you from the fact that if you want people to use your tool, you have to make it easily usable, and usable for real world people rather than for spherical cows.
> I'd argue that ease of use and readability are perfectly quantifiable
Do we have any objective metrics? The only thing I can think of is all the (inconclusive) studies about error density. Unfortunately error density is a very far cry from measuring the overall quality of a software product, or the mental effort that was required to produce it, or the mental effort required to maintain it.
> there are thousands of people whose job is precisely optimising usability for many products
Yeah but programming languages are more abstract; they're tools of thought (just like mathematical objects), and optimizing their use is nothing like observing a user press the buttons on a machine to get themselves a cup of coffee.
Controlled studies are very difficult to run, for obvious reasons. (Unlike for the coffee machine)
kace91|5 years ago
What those features are, rather than subjective, it's contextual, and I feel that behind the author's phrase might lie the usual complain by functional programmers that non-functional languages just feel foreign because people are usually taught other styles during their learning process.
As an explanation, that might be partly true, but it being true doesn't really save you from the fact that if you want people to use your tool, you have to make it easily usable, and usable for real world people rather than for spherical cows.
clarry|5 years ago
Do we have any objective metrics? The only thing I can think of is all the (inconclusive) studies about error density. Unfortunately error density is a very far cry from measuring the overall quality of a software product, or the mental effort that was required to produce it, or the mental effort required to maintain it.
> there are thousands of people whose job is precisely optimising usability for many products
Yeah but programming languages are more abstract; they're tools of thought (just like mathematical objects), and optimizing their use is nothing like observing a user press the buttons on a machine to get themselves a cup of coffee.
Controlled studies are very difficult to run, for obvious reasons. (Unlike for the coffee machine)
dkubb|5 years ago