The "linage" implies that they aren't aware of it, or at least they weren't when designing the language. Rob Pike is more of a systems guy than a type theorist.
Like most systems guys, they know of how C handles types and how C++ handles types. The concept of Parametric Polymorphism or sum types outside of OOP is most likely something Rob Pike was not familiar with as it's hard to see why sum types were not included in the language.
Having IO functions return a tuple of error and value rather then an Option type is not simplicity it's complexity arising from lack of a type primitive. The feature is ugly and very much looks like it was implemented by someone who wasn't aware of a type that can be a Value OR an Error. So instead he implemented a type that can be an Error AND a Value and left it to manual run time checks for the programmer to figure out if an error even occurred.
The other thing is that this "tuple" type of error AND value looks like a hack. Tuple types can't be saved to a variable in GO and can only be returned by functions and immediately unrolled into its separate primitive values. It's like Rob knew something was off in this area so he created a temporary concept of a tuple in the return value of a function to make up for it. A consistently designed language wouldn't have a Tuple only returnable by a function call. It seems strange and inconsistent.
Additionally, the fact that, in GO, some types can have Nils and other types default to a zero value implies that Rob Knew nulls were bad but didn't know how to completely get rid of the null.
I'm thinking that Robs initial notions of sum types and parametric polymorphism is that they can only be implemented via hierarchies of inheritance which in itself has many problems. It makes sense because this is what systems programmers are exposed to (C, C++) as opposed to typed lambda calculus or Haskell. So it's easy to see that Go is the result of Robs awareness of problems with OOP but lack of awareness of the theoretical alternative.
He was at the premiere computer science research organization for decades. And you presume that he's not familiar with sum types, because if he was, he would have included it, and therefore he can't have been familiar with it, because he didn't include it? That's the most amazingly arrogant thing I've heard in a while.
The fact is that Rob almost certainly knows more than you, rather than less. And he still made the choices he made. That should make you ask questions, not about Rob's knowledge, but about yours.
Read that quote by Rob and the responses. IN the quote rob describes what he believes Generic Types are at face value... he literally takes it into a tirade about inheritance and hierarchies of classes... something that is not part of type theory at all.
Honestly, it feels like he didn't know about Algebraic Data types. I'm not the only one who thinks this as shown from the responses to his quotation.
One of the responses:
"Or perhaps Rob Pike just hasn't explored the relevant literature in enough depth. At one point he admitted that he didn't know that structural typing had already been invented previously! This isn't to criticize Rob, I find his talks fascinating, I think he's awesome, he's a friend of my boss, etc. But he's hardly the first hard-core hacker to be ignorant of the degree to which type theory has seen dramatic advances since the 1980s."
Honestly this is really compelling evidence that Pike doesn't know much about type theory. That isn't terribly surprising, and the other early collaborators on the language that I know of also came from more of a systems background. I think its entirely likely that Go's crippled type system is partly an accident, and not entirely a design choice. It would be helpful if - with the benefit of hindsight - they would admit it, rather than invent post-hoc justifications for the way things are.
I don't know man. I feel like there are places I know more than Rob Pike almost certainly. Like, I don't know, most of functional programming. I seriously doubt he knows what indexed monads are better than me.
So, at the point they were creating Go, I think it's perfectly reasonable they had even less exposure to fp, and didn't actually know about these better solutions.
Well... those Bell Labs types were polyglots. They tried a lot of things in a lot of languages. Does that mean that Rob Pike knew about sum types? Not necessarily, no. But it gives you two possibilities.
1. Rob Pike spent all that time at Bell Labs, with all these CS experts, read and wrote all those papers, and never heard about sum types. That's... possible. It's not the way I would bet, but it's possible.
2. Rob Pike knew perfectly well what sum types were, and left them out of Go, because he thought they didn't fit with what he was trying to do.
To me, the second is both more charitable, and more in line with what I think Rob Pike's background and experience would have exposed him to. crimsonalucard obviously disagrees. He seems to think that sum types are so obviously the right thing that Pike could not have possibly not put them in Go had he known about them, and therefore he could not have known. And that is in fact possible.
But it seems to me to better fit with Pike's background, as well as with the principle of charity, to assume that he knew. And still he chose to leave them out.
Now, he could still be wrong. And we can discuss whether sum types are really a good fit for what Go is trying to do. But the assumption that he couldn't have known, or he would have done it the way someone else thinks he should have, is what grates on me.
Please read my initial assumption. In no place did I say he COULDN'T have known. Read it. I literally started the statement with "I think the designers probably didn't know" rather than "I know they COULDN'T have known." There is nothing to "grate" you here. I simply had an opinion and a guess, and you disagreed with it and decided to insult me.
What grates me is the assumption that I said it's 100% true that Rob Pike didn't know what a sum type was. I think it's very likely he didn't know. If he did know then I am wrong. That's all.
That FAQ should have been presented earlier in a cordial and civil way. If you did I would have admitted that my hypothesis was incorrect. Science logic and evidence rule the day and I try to not invest any emotion into any of my opinions. It's hard but I follow this rule. If the FAQ says he knows about it then he does and I am wrong. Instead you chose not to present this evidence and call me arrogant.
There was no need to call me "Arrogant." It disgusts me to hear people talk like this. Either way the GO the language feels awkward in the way it uses product types and does indeed feel like Rob didn't know about them because the sum types certainly do feel more fitting then having a function return a tuple out of nowhere.
I also disagree with the FAQ. Plenty of languages have constraint types that are placed on the subtypes of the sum type. There's no confusion imo. Also note that the previous sentence was just an opinion. Please don't call me arrogant because I have one.
Well, I didn't have the FAQ earlier. I was guessing then.
And I don't see the FAQ as necessarily total vindication of my position. The language team considered sum types; it doesn't mean that Rob Pike did in the initial design. It could be that, after it was kind of mostly formed, they thought about sum types and couldn't find a sensible way to make them fit. Or it could mean that he considered them and rejected them from the beginning. The FAQ isn't specific enough to say.
As for calling you arrogant: You are not the first person who has said, here on HN, that Pike "looked like he didn't know"/"must not have known"/"couldn't have known". Those conversations kind of run together in my mind. As a result, I was hard on you at least in part because others went too far. That's not fair to you, and I apologize.
I also cannot call you arrogant for having an opinion. I also have one - you may have noticed this. ;-)
However, I feel that I should say (and say as gently as I can) that you often sound very harsh on HN. A harsh tone causes many to read your content with less charity than the ideas might deserve. (I am not here trying to defend my interaction in this thread.) And this is not very helpful of me, because if you ask for advise on what, specifically, to change, I'm not sure I can give any. I mention it because you may be unaware of it, and awareness may help.
I can easily see how the previous paragraph could offend you. I am not trying to do so. Forgive me if it causes offense.
> To me, the second is both more charitable, and more in line with what I think Rob Pike's background and experience would have exposed him to.
More charitable to Rob Pike, rather than to the person you're in the the middle of a conversation with.
Anyway "appeal to authority" is not an argument, it's a religion. Our lord and savior, Rob, knows so much that his design decisions are beyond question.
Not only is it likely that, as you say, Rob Pike was aware of sum types, but he did not create Go on his own but it was created by a small team. Someone like Robert Griesemer, who studied under Wirth, would have known about them, if the others hadn't.
I have been using the Wirth languages a lot (Pascal, Modula), and one of the big appeals of Go to me is, that it brings back a lot from those languages to the modern times. The Wirth languages are far to underrated in programming today.
The irony being that most Wirth languages are more expressive than Go will ever be, with the exception of the first release of Pascal and Oberon versions, and the follow up on minimalist design approach with Oberon-07.
When Go came out, I though it could follow Oberon, starting small and eventually reach Active Oberon/Zonnon expressiveness, but alas that is not how they see it.
Having the value returned together with the error is convenient for a couple of reasons.
First, it's often possible for the function to return a meaningful value even in an error case (e.g., number of bytes read before the error occurred).
Second, it's often possible to return a sensible 'null' value together with an error which can be handled correctly without checking the error value. (A map lookup is the obvious example of this.) This simplifies logic in some places.
Using sum types for errors in Go wouldn't actually work very well unless you fundamentally changed other aspects of the language. You'd need pattern matching, a whole bunch of generic higher order functions for manipulating option/result types, etc. etc.
>First, it's often possible for the function to return a meaningful value even in an error case
Create a type that explicitly stores this information. The return type can hold the (error message and a value) OR (just a value). This type expression is a more accurate description of what's really going on. Whichever way you want to represent the product type it's not isomorphic to the actual intended result that the function should return. A sum type can represent the return value of the sentence below while GO cannot:
"A function that returns (a value) OR an (error with a value)"
This is the true intention of the function you described.
>Second, it's often possible to return a sensible 'null' value together with an error which can be handled correctly without checking the error value. (A map lookup is the obvious example of this.) This simplifies logic in some places.
But opens up the possibility of a runtime error if you fail to check for it. Historically there are tons of functions in C, C++ or javascript that use null to represent an error and it is self quoted to be the greatest mistake ever made by the creator of null. No language needs a null.
>Using sum types for errors in Go wouldn't actually work very well unless you fundamentally changed other aspects of the language. You'd need pattern matching....
Using product types to represent errors has already changed the nature of GO in a very hacky way.
Only Functions in GO can return tuples and the concept of the tuple can never be used anywhere else. You cannot save a variable as a tuple, you cannot pass a tuple as an argument. You can only return a tuple then instantly unroll it. It's an arbitrary hacky feature obviously made to support error values.
It would be better to have an arbitrary pattern matching feature... this makes more sense then arbitrary tuples returned from functions.
>a whole bunch of generic higher order functions for manipulating option/result types, etc. etc.
Actually no you don't. The fact that go functions return tuples with errors, does this mean that higher order functions need to handle tuples? No! not at all. In fact go explicitly eliminates support for this... The tuples in GO need to be unrolled into their constituent types before they can be used in any other function. The same concept can be applied to Option types. You have to unroll the value and explicitly handle either individual type. You do not ever need a higher order function that accepts the Option type as a parameter.
Like all languages that have the Option Type/Maybe Monad etc... Any function that returns this type represents a function that is impure that needs to be unrolled first before passing the value down to the functions that do closed and pure calculations. A function that takes an Option type as a parameter is a function that says "I am a function that can only take values from IO" It's very rare for functions to be implemented like this even in languages that have first class support for the sum type and Monads. In haskell I can't recall ever seeing a function that takes the IO monad as a parameter. In haskell and in Rust these values need to be unrolled into their constituent types before they can be used.
Please note I am not advocating the inclusion of Monads into GO. Just talking about sum types.
To me, Go is "masturbation prevention" language. Meaning that certain deliberate design choices were made to prevent precisely the types of endless unproductive masturbation you see in some other languages. I.e. type masturbation in Haskell or OOP/IoC masturbation in Java (particularly egregious, Java is not a bad language otherwise), or metaprogramming mastrurbation in C++. The omission of these features is not a bug. It's a feature in itself.
Indeed. More than once, I have seen program projects ruined by creating an overly elaborate class hierarchy, sometimes 10 layers deep. Just to express any theoretical aspect of the domain in the structure of the class hierarchy. Java programs often suffer from this. Which is especially sad, as Java has interfaces, which I think are the right way of representing abstract types for APIs for example. Interfaces don't force you to into a type inheritance just to fulfil a contract. But unfortunately, they are way to rarely used.
If anything in Java they're overused. You often see only one class implementing an interface where the programmer can reasonably expect there will never be another implementation, and where it's not exposed outside the API boundary, so the interface is gratuitous.
