Hacker News Reader

Comment by LandR | Hacker News Reader

> I'm guessing that you two are having a communication problem as it is highly unlikely that a "senior colleague" can't understand simple logic.

I've seen code written by a senior developer who writes

   if some-predicate
      // left blank.
   else
     // do something
   end if

Because it's apparently too hard to work out how to invert logic. It's a pattern I've seen in our code base enough that I've started calling it 'if-meh-else'

View on HN

> Because it's apparently too hard to work out how to invert logic.

This is almost never the case. Please don't assume people don't know what they're doing, especially to such an extent. I would actually recommend attributing expertise and wisdom to their decisions until proven otherwise.

This particular construct is not a lack of understanding but often used to improve readability, transition code that may be in that block in the past or future, denote the normal vs exception branch, and several other scenarios. Did you ever ask why this person wrote it this way?

I've written logic like to draw attention to the fact that the ELSE block is the exception.

Semantically you're not saying "if X is true" you're saying "if X doesn't happen". It's a subtle difference and this format draws attention to it for the Next Developer.

That how I would read it too.

Do you really think they don't know how to do one of the most basic logical operations in all of computing?

It could be a convention to maintain semantic consistency in the predicate so you don't have to randomly flip your brain each time you encounter a new conditional statement.

Ask around. Could at least be a funny story behind it. The great not not of 2014.

You already have ample examples of how this could make sense, so here's a mental model that will serve you well in life.

https://en.wikipedia.org/wiki/Wikipedia:Chesterton%27s_fence

>> 'In the matter of reforming things, as distinct from deforming them, there is one plain and simple principle; a principle which will probably be called a paradox. There exists in such a case a certain institution or law; let us say, for the sake of simplicity, a fence or gate erected across a road. The more modern type of reformer goes gaily up to it and says, "I don't see the use of this; let us clear it away." To which the more intelligent type of reformer will do well to answer: "If you don't see the use of it, I certainly won't let you clear it away. Go away and think. Then, when you can come back and tell me that you do see the use of it, I may allow you to destroy it."'

I don't know if you have read Code Complete. They kinda explain there why such a code would make sense. I've written constructs resembling this many times and got them through rigorous PRs with quite experienced and senior people.

One must remember that code is not a paint and a canvas for you to feel like an artist but a documentation for for Next Developer to comprehend. One must write it like a book or some sort of navigation manual, drawing attention of the reader to the important parts and guiding them through intricacies of the requirements and logic.

When I have an if clause that is too complicated to work out how to correctly invert it (because while I understand de Morgan's just fine, it can get very detailed), I find it far preferable to start factoring out the expression, e.g. if I need to invert,

    if ((a and b) or (a and c and not d)) and not e)

should be something more like

    orderSubmitted = a and b
    orderValid = a and c and not d
    orderNotYetFulfilled = (orderSubmitted or orderValid) and not e
    if not orderNotYetFulfilled {
         # code here
    }

Ideally, a through e are also named something reasonable enough that the variable declarations read cleanly.

The raw de Morgan's transform is only necessary if it's a performance optimization, and in 2019 the conditions for that to be an issue are very, very precise.

I assume the compiler can easily optimize something like that similar to constant folding. Isn't that the case?

In most cases I'd expect so. The rare cases I was thinking of was on the order of "I'm on an embedded system using a crappy compiler and I can't change it and I literally can't afford extra bytes for boolean values the compiler won't optimize away", and on the other end, "I'm in the cloud and my code has confused the optimizer and I have to code this inner loop incredibly carefully because it's going to be called several billion times per second". Maybe the odd other corner case here or there.

There shouldn't be much in-between in 2019 where writing the clauses out this way will have a materially negative impact on one's code, and it should have a generally very positive one.

In this case I'd probably write an inline function which would accept all those booleans and then

return (x && y) || (a && b) && !(c != d) //pseudo example of complex condition

Or write a bunch of macros given that if you're on embedded 99.999% probability you're writing C.

Often when I use this, it's to make it explicit that the normal/common case is met and the other clause is truly exceptional.

It's also a stub where you can add logging or similar along the normal path without touching the conditional and risking introducing bugs by accident.

Depending on the language or runtime the empty clause will likely be factored out by the optimizer so there's no cost for extra readability and flexibility.

Precisely. This construct is used to indicate likely vs unlikely branches in code. It's a pretty common pattern.

Personally, for readability purpose, this sounds a like a good idea. It is indeed (at least for me) harder to read a condition that has negation:

if(!some-predicate)

Maybe it's because you have to do the negation mentally.

But more ideally, it should be done with polymorphism or similar rather than using if..else.

That could be for readability or some weird convention to make the code more streamlined? I've also seen it done that way so there could be a reason behind it.

something I can think of is that at first the if-branch did do something, then later it got commented out. And later again someone deleted the commented-out code instead of rewriting it.

That'd be something pretty silly to do. But sadly I've seen enough bad code that this would not surprise me.

Sometimes I do that when negating would be unintutive or more verbose with regards to the else and if else statement.

How about if(expression) foo = true else foo = false

Reply on news.ycombinator.com