Hacker News Reader

Comment by skybrian | Hacker News Reader

The comment is confusing, but the idea seems to be that instead of calling this function, you should inline the code - that is, just write the comparisons yourself. You don't need a function call.

I guess this is for stylistic reasons, but I don't know why anyone would feel strongly about doing it one way or the other.

View on HN

I don’t think that’s true. I believe they want you to explicitly define how you want you compare strings. Comparing Unicode strings isn’t as straightforward as comparing individual bytes or code points. For example, there are multiple Unicode strings that will yield the character “ï”. If you use a naïve comparison function, you can’t be sure that it will behave as expected when it encounters the word “naïve”.

You’re looking for Collator.CompareString

https://pkg.go.dev/golang.org/x/text/collate#Collator.Compar...

> I believe they want you to explicitly define how you want you compare strings.

If that were true, they wouldn't have made strings comparable in the first place, and they wouldn't recommend that callers inline the implementation (they'd recommend calling some locale-dependent function instead).

An explicit comparison like "naïve" == "naïve" or "naïve" < "naïve" isn't any more clear about how the comparison is performed than Compare("naïve", "naïve") would be.

Yes, I mentioned that that's basically what it says. But why would you tell people to actively duplicate code? It's like "I don't get what this type of function does and have to look it up all the time, better to have everyone inline this so it's clearer".

To be fair, I agree that seeing bad examples of its use shown by one of the other folks here, some could have been avoided by them being forced to inline, but funnily those examples still exist despite the desires of the function author.

Nothing was achieved here.

I think manual three-way comparisons are very clear. But so is strings.Compare. For better or worse, copying code is a pretty normal thing to do in Go. See "A little copying is better than a little dependency." [0]

[0]: https://www.youtube.com/watch?v=PAAkCSZUG1c&t=9m28s&themeRef...

I even somewhat agree with that principle, but it don’t think it applies to using the standard library, which is (a) not little and (b) not a new dependency you add to the project.

If that’s the justification to keep an implementation bad on purpose (which I’m not sure is the actual intention, but is at least that’s what the referenced comment claims), then I’m not even sure I speak the same language as the people who made that decision.

This is a part of Go's rather odd perspective that I appreciate.

I've worked in codebases with library helpers for everything that mostly served to turn two clear lines into a function call. Once such functions exist, folks feel obligated to use them; after all, why duplicate code? So now, what would've been a couple dozen lines of self-contained straightforward code has 3 imports and 5 functions you need to be familiar with to understand it. It also makes compilation more expensive.

I'm not saying library functions are bad or anything like that, but there's value to keeping the typical vocabulary of code small, of not adding new dependencies to solve trivial problems. Especially in a standard library that is widely used and you expect to be maintained for years. Providing functions that solve problems that are impossible or tricky in the language is important; providing `Plus(a, b int) int` that wraps `+` makes the library worse.

I think there's a good argument for providing Compare() and Abs() and other fairly trivial functions even if it is perhaps less effort to not use them in most cases, but for a stdlib I can appreciate the logic of leaving out what isn't providing clear value.

I would understand that if the method was left out. But in this case, a strings.Compare method is provided but the implementation says not to use it. That's the worst of both worlds. The stdlib still has this extra method and users will feel obligated to use it because it exists but the method is inefficient.

The standard library can't ever remove anything because of the Go compatibility promise, and a certain percentage of it is mistakes. Once some functionality is later realized to be poor or incorrect, and the design prevents fixing it, there is not much that can be done other than telling people not to use it anymore. What you refer to as the worst of both worlds is unfortunately inevitable, but at least it's in service of a greater goal.

edit: clarity

That is what happens when they refuse to adopt a deprecation process, even Java, .NET, C and C++ with their high regard for backwards compatibility do have such processes, and have removed features from standard library and languages.

Go marks features as deprecated, though, both in source code and in the docs. For example, strings.Title() has been deprecated and instead the case package should be used. The backwards guarantee merely means that the feature won't be removed from the library in Go 1.x, but when you write new code you won't use it and the preferred way of doing the same is mentioned in the docs and version notes. I think it's good not to remove deprecated functions within the same major version.

Except Go certainly is never getting another major version, I am waiting for Go 1.10000.0, given how decisions are made on the ecosystem.

What's wrong with Go 2 with tools to automate migrations? Of course that limits the scale of possible changes, but moving methods between packages should be possible with this approach.

Community culture, most likely it will never happen as such.

Change doesn’t have to come at the pace of npm. Let’s be charitable, people. :)

I seem to recall one of go’s creators snapped back at a similar question about code duplication with “what’s the matter, are your fingers broken?”

I think that sums up the philosophy, though maybe not as well as “Nothing was achieved here”, which I would like to translate into Latin and get on some stickers.

Go doesn't have many features of other languages. There are several competing goals, including keeping the language small, not hiding complexity, etc.

Manually writing the three-way comparison fits in with these goals. If slices were comparable then I'd wager that bytes.Compare would not exist.

What sticks out to me is the part saying "the compiler should be changed". It's weird enough that this isn't even a doc comment but a comment inside the function (making it harder for people using the function to notice), but even as someone who thinks the passive voice is often unfairly maligned, the phrasing immediately brings to mind the question "who should change the compiler?" Is the Go standard library not maintained by the same group of people as the compiler, or is this comment just a doubly obfuscated "TODO"?

It's not a TODO yet, it 's a note that says if you think this would be a good idea, do that instead. It's a potential TODO-to-be, waiting for the need.

Is that better? That would mean that they have a method in their API that they don't think anyone should use but haven't deprecated it or documented it that way and have no plans to do anything about it.

strings.Compare is mostly used in trivial demonstration programs, and is there because bytes.Compare is there. It makes https://pkg.go.dev/sort#Find documentation simpler. Deprecating strings.Compare would make sort documentation worse.

Real code tends to not be that trivial, and that's why real code most likely shouldn't be using strings.Compare.

Even if the compiler had good optimization of three way compares, this would be useful for passing to a higher order function. Say a function that allows you to specify how strings should be ordered.

Reply on news.ycombinator.com