New in C# 10: Easier Lambda Expressions(dontcodetired.com) |
New in C# 10: Easier Lambda Expressions(dontcodetired.com) |
Try Elm as a simple example (can be done in a weekend), it'll probably blow your mind. You don't have to write type annotations at all, but the compiler complains at build time if the same function is called with two different types in two places.
The reasons are simple: it's promoted by a big company. Same reason why C# and Go are popular.
Won't be back to static typing for a while unless I'll need higher precision and higher reliability module.
The funny .. or sad.. part is that the c#9 explicit ~verbose style would have been the only one accepted before. If you wrote implicitely typed variables people would get angry (I think there are many online articles about how java 9 `var` was bad)
It can be both. Type inference isn't a brand-new idea, but it still takes work to diffuse it into mainstream, practical languages, especially retrofitting it onto existing languages that weren't designed for it. That still counts as innovation in my book.
I still remember how ugly and tedious STL iterators were in C++. Now it’s just “auto”. LINQ also wouldn’t work without type inference.
Having worked recently in it, I found Java to be relatively (as a language, not for tooling) backwards compared to C#.
Every generation must learn this for themselves though.
You’re also likely to discover that while C# the language is amazing, the CLR runtime looks frankly minor league when compared to the JVM.
This is comically backwards from my experience. I often see near-native optimized performance on C#/CLR, where I see similar code (SIMD friendly loops) literally hundreds if not thousands of times slower than it should be in Java.
https://docs.microsoft.com/en-us/dotnet/csharp/language-refe...
Lambdas can also have attributes now, too.
The F# style is enabled by a set of features - some of which would be really hard to add to C# (such as currying and global type inference) - but even if they were added, the millions(?) of C# developers would be unlikely adopt the functional style just because it was possible. A language is not just a list of features; each language has its own culture and "idiomatic" way of doing things.
https://devblogs.microsoft.com/dotnet/whats-new-in-fsharp-6/
I hate they delayed the introduction of Algebraic Data Types again. Maybe they will appear in C# 11.
Using ADT in F# makes it a breeze to do domain modeling.
For example file scoped namespace is a very simple but great change, improvement for lambdas, improvement for structs, all of this make it nicer and simpler to write readable code.
Edit: I would also love to see ADT in C#!
This version brings quality of life changes that make it simpler and nicer to write readable code, which is great to see IMHO.
I do wonder what was limiting them before
Heck, I'd say, by now with local functions most lambdas that previously would have been a local could now just be a local function.
Java had two tech advantages over C#: It officially ran on platforms other than Windows and ran well, and their JVM was far ahead of the MS CLR.
Both of these things went away/are going away.
What really remains is Java's reputation over "Embrace, Extend, Extinguish Microsoft" (tho Oracle has been busy trying to destroy Java's reputation ever since they bought Sun) and the larger eco-system/community/existing code bases and sunk costs.
That's all of no use to me (I don't work on projects like that), but let's not just shove that aside.
I also don't really agree with this:
> look and wait and do a half-assed implementation of a lot of tried-and-true C# language features some years later
There weren't other buyers in town, besides IBM, which I probably would have managed it just as Oracle has, given that the trio have been around in Java world since the early days.
Oracle engineers did the first implementation for Java 7, more recently it's mostly contributions from Intel, AMD and ARM [1].
[1] https://cr.openjdk.java.net/~vlivanov/talks/2019_CodeOne_MTE...
OpenJDK, GraalVM, Open J9, Aicas, PTC, microEJ, Azul, Amazon,....
main = do
putStrLn "Who are you?"
name <- readLn
putStrLn ("Hello, " ++ name)
Do you need to know how do syntax and the assignment operator <- works? Sure. But that has nothing to do with category theory. That's just syntax.
a. do notation is converted to haskell b. what the bind and return functions do, for IO
So you can figure out;
c. Why your types are not lining up
To understand Haskell in general you need to realise do, bind and return are generic and can be used for not just IO but say for Maybe, List etc.
Basically you need to know most practical things about monads!
I had a bad time writing Haskell do notation until I understood monads. I used to write imperative code, try =, try <-, always undo typing back to known working states etc. to try and magic the code into compiling.
I’m doing a take 2 now of getting into Haskell again but ignoring advanced language features (unless forced on me by a library) and ignoring category theory. The goal this time is using Haskell to just build stuff.
I find it kind of weird that we've normalized not being able to read the code outside of the proper program for it. Since those IDEs often cost money it starts to feel a bit like steps towards a walled garden to me, and I'm not sure its good for actual computer science.
Besides, the paper itself should optimise for clarity and global type inference languages allow type annotation where it might help. They simply let you skip type annotations which only add noise.
That's also the difference in 'philosophy' between OCaml, F# and Haskell. Haskell has the 'same' global type inference but the community sees them as (often only ;) documentation.
Java has higher peak JIT performance though after Rosalyn was introduced C# has made strides at closing that gap on many metrics. GC algorithms are more advanced, namely ZGC and Shenandoah. I would elaborate but I don't want to turn this post into an actual essay. JVM is more configurable. This is a double edged sword, it heavily favors users that have spent the time to really understand the platform but it has created a reputation of the JVM being hard to operate. JVM class loaders and the new(ish) Jigzaw module system are very powerful, especially for shipping customized/cut-down JVM + stdlib when you need to do distribution. JVM has better portability. .NET Core has obviously changed the game but other platforms are still treated second class to Windows in many ways. GraalVM is reviving the guest language ecosystem of the JVM. Clojure, Scala and Kotlin have all remained reasonably strong but there was a dark period there for Jython/JRuby and friends. GraalVM is changing the tide here and I have already been able to leverage this over Nashorn to embed essentially native speed JS code into one of my Java programs.
CLR has better primitive types, JVM's project Vahalla may resolve this. The CLR/.NET tends to introduce features at a quicker pace than the JVM/Java, this I feel is CLR's double edged sword. Async/await being the key example. It's been in the CLR for what feels like forever now vs JVM which is only now just getting close to landing Loom. However I feel like Loom is an infinitely better solution to the problem that leverages the unique advantages of JIT compiled VM languages that rely on relatively little unmanaged code. The experience on the CLR is much more integrated assuming you are on the well trodden path. i.e VS + Windows 10. MSIL is much much more pleasant than JVM bytecode in my experience and the community developed tooling for working with it is great. Optimizing programs on CLR is easier than JVM, it takes comparatively less effort to write allocation free code and get great memory layout using primitive arrays etc for high performance code. CLR interacts better with native code. JNI is poor in comparison. Project Panama could make this better on the JVM side.
This is by no means a complete list of ways the two differ, this is just my experiences having written a decent chunk of code for both.
FWIW I prefer the JVM but if I am forced not to use the JVM the CLR is the very next thing I propose.
Not really. Maybe a tiny little bit, but not really. Microsoft has discovered that the "other platforms" are very important, because they want to sell Azure, and people there will use linux predominantly. Same as MS invested heavily in WSL to support "linux" workflows.
dotnet on linux might not be completely to parity with dotnet on windows when it comes to platform integration, but it's getting there. dotnet on macos is somewhat similar in this regard and maybe a little bit slower to catch up because may be a little lower in priority than linux.
Anyway, it's not my impression that platforms other than Windows are treated second class. It's just that Microsoft has already invested decades in the Windows implementation, while the linux and macos implementations are rather fresh (although dotnet using mono as a base certainly helped) and therefore less optimized yet in certain regards.
>Java has higher peak JIT performance though after Rosalyn was introduced C# has made strides at closing that gap on many metrics.
That isn't really true. JIT (peak) performance is pretty much comparable between Java and dotnet, maybe even a little better in dotnet/RyuJIT. There are certain areas where one is better than the other.
Roslyn is the C# to MSIL compiler, btw, while the JIT compiler for MSIL is RyuJIT. While Roslyn helps produce MSIL that is (hopefully) great to JIT compile, the actual JIT compilation work happens in RyuJIT.
> GC algorithms are more advanced, namely ZGC and Shenandoah.
Yes, GC is where the JVM really shines - if you go through the trouble of finetuning it to your particular need. At the same time the CLR GC seems to work better by default, but it's a lot less configurable too. In the CLR you basically have only 3 modes of operation: interactive GC (concurrent, trying to avoid GC pauses, with the trade off of less performance, for things like GUI applications), non-concurrent GC (which works well for a lot of CLI type tools and some services, where small GC pauses are less noticeable), and "server" (which is e.g. a lot less eager to "return" memory to the OS, aimed at long running processes that are then usually the only major thing running), and that's basically it, where on java you have tons of available schemes/algorithms and each one is usually heavily configurable as well.
As you already hinted at, dotnet made great strides in the last years to avoid a lot of heap allocations and the GC with value types, Span<> and friends, avoiding boxing a lot of times and bringing stackalloc to managed code, and their use in the standard library. So peak GC performance is a little less important in the CLR than in the JVM (but still very important, of course).
You say "may" a lot, referring to proposed java changes e.g. to add value types, make native code interaction less painful, add some form of async-await. Great. And I hope they will finish up and ship these things eventually, and it will be great implementations. The thing is dotnet already has all these things shipped and in production already, often for quite a while. And while java is still figuring out the design, dotnet can already improve their shipped implementations based on real world experience and data in every new version, and usually does so. Java/JVM is playing catch up here.
>Async/await being the key example. It's been in the CLR for what feels like forever now vs JVM which is only now just getting close to landing Loom. However I feel like Loom is an infinitely better solution to the problem that leverages the unique advantages of JIT compiled VM languages that rely on relatively little unmanaged code.
From what I have seen of Loom so far (which is admittedly not that much) my impression is the direct opposite. To me it so far looks worse conceptually and API-wise than what MS did, while still borrowing a lot of general async-await concepts and even C# and CLR specific concepts, but worse. Don't get me wrong, C# async-await and the CLR support for it has a lot of problematic areas too and is far from perfect and easy, aside from general async-await conceptual problems, but it doesn't seem to me like Loom is avoiding or solving those things, instead adding quite a few of their own warts on top. But as I said, my exposure to Loom is very limited, and it's not a final product yet anyways, so I may be very wrong here.
Loom reminds me of when Java tried to copy a lot of collections-based LINQ stuff but did a terrible job doing so in my humble opinion (they since fixed some of the worst mistakes). Or maybe I am just biased, which is certainly a possibility.
Very important to be able to run applications, not to develop them. I spent a few years doing dotnet development on Linux and have had enough of it (so I'm moving away to JVM which looks more and more interesting as long-term research projects and some language improvements land). It is very much a second class citizen. They still want to force developers use Windows and their tooling.
* `operator=` and `operator ,`
* `operator new` and `operator delete`
* `operator +=` overloaded differently from `operator +`
* `operator &&` and `operator ||` overloading works differently in C#, so that it preserves the short-circuiting behavior
* `operator ++` only can be overloaded once in C#, with the compiler automatically handling the difference between pre- and post-increment
* everyone needs to learn about how to overload `operator=` (for memory management)
* the standard library encourage abuses of operator overloading such as shift operators for iostreams
The language kind of put itself into a corner when what C# calls `MoveNext` is called `++` (with two variants of course) and what C# calls `Current` is called `*`. Taking what mathematically worked for iterating an array and taking it piecemeal to define your syntax is the kind of operator overloading that is a bad idea.
These were the major pain points for the Java lawsuit that Sun did against Microsoft.
The irony with all these attempts to make COM easier to use is that they always fail flat when the Windows dev team has their turn, and then they undo everything and we are back into C++ and IDL land (latest version of it, C++/WinRT).
We who? From where I stand people are fine with operator overloading where it makes sense and C++ is far from being the only language with this feature.
Problem comes when shitty programmers get orgasmic about some language feature and try to use it everywhere to everyone's detriment.
Gross exaggeration. Just a bit of plain common sense. Do override operators to perform matrix ops. Do not override + operator to do multiplication instead. And do not subtract apples from cars.
As for your further example as I said give shitty programmer code in any language and they will manage to fuck it up.
Serialization libraries need to be decoupled from the records being serialized. These two things are compiled into different assemblies, and often written by different people.
Still possible to replace with compile time codegen, but the implementation gonna be complicated and fragile.
I would also disagree with the characterization as shitty code. The operators are actually really nice if you manage to learn them all well. It can make the code both very concise and readable. But it optimized for people who know the libraries extremely well over people who don't, which usually is not a great tradeoff for widely used libraries.
This sounds suspiciously regex like. Concise it is, readable it is not. But regex is unique case and and harassing people with similar patterns in every line of code should be punishable by at least 10 years of maintaining complex software written in Brainfuck.
If a library doesn't have an API for using them with generics, you can paper over that with:
T fn<T>(T throwAway) => Library.Deserialize<T>();Sometimes people lump these in with dynamic objects but those are a very different class of thing indeed.
Now, what would be -really- nice is if we had a way to specify that an anonymous type implements an interface. That could potentially simplify a -lot- of DTO modeling.