Hacker News Reader

Comment by don_albrecht | Hacker News Reader

This advice should be universal in coding. When I first started programming, I had a manager that hated 1-2 character variables. But they make sense for loop iterators.

View on HN

For loop iterations it's fine, but I think Go code often takes this too far.

It often takes longer to read a single character than a word, because I have to mentally map the character to the word anyway.

Reminds me of when people go nuts aliasing table names in SQL queries, which IMO makes it harder to read as well.

I agree. Plus, with IDEs and auto-completion, why not use longer names?

One reason so use short names is to keep lines short and reduce line wrapping. This helps to read code, especially when you do 3-way merges.

- Longer names mean longer lines, and lines that are longer than your viewport are terrible. Less of a problem with modern ultrawide screens, but passed ~100 characters, it starts becoming a problem in some cases (ex: diffing). Personally, I have a soft limit at 80, hard limit at 120, because that's what works best for me.

- With such rules, variable length is a hint of its scope. For example, I tend to use "i" when the loop body is small, "idx" or "index" when it is a bit larger, and a more descriptive name when it exceeds one screen. This way, just by looking at a single line, I already have a hint about its context.

- Just because a variable name is short doesn't mean it is meaningless. For example "i", "j", "k", are loop indices, "x", "y", "z" are point coordinates and "dx", "dy", and "dz" are differences, "a" and "b" are both sides of a comparison function, "t" can be a temporary variable or a time depending on context, etc... The corollary is to make sure you use your single letter variables consistently. If I see an "x" in a place where a coordinate can be used and it does not refer to a coordinate, or an "i" that is not a loop index, I will be confused.

If you have to use that name a bunch of times close together, it bloats line lengths which also affects readability.

There's also a bit of Fitt's Law in here as well - I find it a lot harder to select, or put my cursor in, a single letter variable. So if I want to rename `i` to something else, I must carefully select `i`, whereas with a longer variable I find it easier to put my cursor in the middle of `index` and hit F2.

If you are using a mouse for things like this you already lost though.

Most developers use their mouse quite often.

I started using ii as my iterator variable name for similar reasons. Easier to highlight and Ctrl-F for. Also removing ambiguity when using i as the imaginary constant.

The constancy in Go makes this better. I have come to expect `r` to be an io.Reader or http.Request depending on context.

There are a few interfaces in Go that are used heavily and I don't mind that people often use a single character for them.

It's the same thing as everyone using `i` for iterators.

I just use reader//req. Code is instantly easier to read for me personally and no time has been wasted.

It's all preference, of course.

Sometimes you run into packages called reader which makes for a less fun time using variable names that are also words.

This is much nicer for other people coming into this. It's the same problem with spoken language and slang, slang is better if you know it, worse if you don't.

Sorry to be pedantic (well, not really it's M.O. for a hacker news commentor after all, isn't it?) , but as someone who knows a few languages, slang is not an issue with spoken languages at all. You learn a language by full immersion with other speakers, who impart that knowledge unto you. Unless all you are doing is conversing with grammaticians and newscasters, you will pick up on regional and local differences, euphemisms, and slang.

I don't believe so.

Imagine you're new to Go and you haven't mentally mapped all of these abbreviations. What would you rather read, `r` or `reader`?

Designing an API that "stutters" is a very common mistake that many programmers make. reader.Read() is pretty jarring.

There is also nothing wrong with naming variables after what they're for instead of what they are. Consider io.Copy(dst, src). That's nicer than io.Copy(reader1, reader2).

But not necessarily nicer than io.Copy(destination, source).

Sure, so call your reader `file` or `network` or `stream`. Don't call it `r`. Nobody is charging you by the character.

It isn't a huge deal when the code is simple, e.g. you create a reader named `r` and in the next line you use it. That's easy to understand. The problem is that the pattern propagates. A few more lines of code are added, the reader is still named `r`. It's name `r` here, so we start putting it in a struct as `r`. Soon enough the codebase is littered with instances of `r`. Not great.

This makes me wonder: what if there was a language where variable names are determined according to the type, with the option of overriding with a custom name. So a variable of type http.Request would automatically be named “req”, the next one in scope would be “req2”, etc.

If you think about it, when you solve a physics problem, for instance, you call every mass “m1”, “m2”, etc. Maybe this would be another step in Go’s direction of conforming style to make code more standard and readable.

This is interesting.

The idea of having a standard library of types related to what are effectively program keywords... could be really good. Or hideous like global vars.

I wonder if any language has tried this?

https://en.wikipedia.org/wiki/Hungarian_notation

Oh cool, thanks for that.

I agree that most types come with a natural variable name.

However, in many cases a more descriptive name is way appropriate. On top of my mind:

- Multiple variables of the same type. How are you supposed to distinguish between req and req2? Compare it to something like "apiReq" and "cdnReq"

- Primitive types, that does not inherently carry a domain value. An integer called "seconds" or "max_offset" has a lot more meaning that one named "num"

Forcing such a notation into the language would make impossibile to represent all these cases

The real question from this example is: why are you creating two http requests in the same scope before using them?

I've been writing Go cloud stuff for the better part of a decade and "req" for a request has never been ambiguous because I go ahead and send the request and process the response before sending another one, at which point I can just reassign the variable and let the first one fall out of scope.

The request type is just an example, of course, building up from the example in the comment I was replying to.

You’re right. Going back to my physics example, all the variables in a physics problem are actually of the same type, float. So you’d need more specific types, but that would be infeasible-the whole point of types in a general purpose language is that they’re general enough for any use case.

If you were just coding physics problems you could have types restricted to the physics domain, that is, physical units. Which it looks like MATLAB does now support: https://www.mathworks.com/help/symbolic/units-of-measurement...

Werrrrlllllll... It all depends on what yo umean by "type". In physics problems, I find taht quantities may be measured in "floats", but have the types "mass", "amperage", "distance per second per second" and the like.

In Go, I would probably model this as:

    type mass float64
    type speed float64
    type acceleration float64

That way, they'd all have float64 as the "storage type", but it would be harder to accidentally pass an acceleration when I wanted a mass.

> Going back to my physics example, all the variables in a physics problem are actually of the same type, float

This is why some people recommend a practice I've forgotten the name of, where they defined a new type that maps to language type. So type Mass could just be a float, but you know better how to treat it.

What if I introduce a new http.Request between 1 and 2? Now all the references are broken.

Terraform has an extreme approach to that, every time you reference a variable you have to spell out the complete TYPE.NAME.ATTRIBUTE, eg resource.aws_s3_bucket.my_bucket.website_endpoint. There is no way to access my_bucket without the whole prefix, even after assigning it to local variable you must prefix the reads as locals.my_bucket.

For sensitive infrastructure this makes sense, i surely wouldn't want to write other code like this.

That could be easily done by the editor in a statically typed language like Go, so there is no point bloating the language with it. Just turn on an overlay that displays the types of all names. I believe LSP adds this to many languages/editors already.

Or in languages that you can import a resource but give it a custom name. So whenever you are reading someone else's code you have to double check to understand what is going on. For instance, import DiskUtils as DU, and now all the code references DU.diskSpace()

Exactly.

I think a better heuristic rather than tying it to scope is to sort of semantically huffman encode. Using an iterator is probably the most used variable name that starts with an 'i', so it gets just plain 'i'. As a _concept_ gets more specific, then it gets a longer name.

Yeah i for for loops (in python) x for comprehensions, etc, make sense. But in general, single character variables make code really hard to read.

Yep, you should just use the best judgement, for example

    for distance, time in driving_metrics:
      speed = distance / time

is a lot clearer than

    for d, t in driving_metrics:
      s = d / t

even if it's used in a tiny scope once. But some things don't have a valuable meaning, or the meaning can trivially be inferred.

I meant i as in the iterator variable for a counter like when using enumerate() or a c style for loop where you manually specify the list index.

When I started coding I'd type

   10 FOR F = 1 TO 10 
   20   PRINT F, F*F
   30 NEXT F

Because "FOR" and "F" shared the same key, on my ZX Spectrum keyboard.

And for those that don't know - entering Basic code on the Spectrum (well the original one at least) was ALL single/shifted key shortcuts for keywords rather than typing them out.

eg https://www.old-computers.com/museum/photos/sinclair_zx-spec...

The low cost and quirky masochism was mainly why we loved it.

Same thing for the ZX-80 and ZX-81. If you were careful with your text prompts, you could save several bytes by injecting a keyword instead of typing it out, but anything that was a "start of line only" could be tricky injecting into a string.

Reply on news.ycombinator.com