Near Future of Programming Languages [pdf]
dev.stephendiehl.com
dev.stephendiehl.com
- The borrow checker in Rust is a great innovation. Previously the options were reference counts, garbage collection, or bugs. Now there's a new option. Expect to see a borrow checker in future languages other than Rust.
- Formal methods are still a pain. The technology tends to come from people in love with the theory, resulting in systems that are too hard to use. Yet most of the stuff you really need to prove is very dumb. X can't affect Y. A can't get information B. Invariant C holds everywhere outside the zone (class, module, whatever) where it is transiently modified. No memory safety violations anywhere. What you really need to prove, and can't establish by testing, is basically "bad thing never happens, ever". Focus on that.
- The author talks about "Java forever". It's more like Javascript Everywhere.
- Somebody needs to invent WYSIWYG web design. Again.
- Functional is OK. Imperative is OK. Both in the same program are a mess.
- Multithread is OK. Event-driven is OK. Coroutine-type "async" is OK. They don't play well together in the same program. Especially if added as an afterthought.
- Interprocess communication could use language support.
- We still can't code well for numbers of CPUs in triple digits or higher.
- How do we talk to GPU-type engines better?
* Javascript everywhere is a function of low barrier-to-entry for it, but almost everybody agrees it is flawed as a language. If that's the future, we are screwed as an industry. One thing I've noticed (and I say this as a guy who wrote Ruby for 10+ years), is that type safety is becoming a hugely desired feature for developers again.
* WYSIWYG web design (a la Dreamweaver) died off a little because the tools saw a web page as standing in isolation. We know however that isn't interesting on its own - it needs to be hooked up to back-end functionality. Who is producing static HTML alone these days? In the case of SPAs it needs API integration. In the case of traditional web app dev, it needs some inlining and hooks to form submission points and programatically generated links to other "pages". Making that easier is the hard part - seeing a web document as an artefact of an output from a running web application container.
* Multi-threaded, event-driven, coroutine-type patterns are fine in Go, to my eye. What's making you think we can't mix this up with the right type of language and tooling support?
* Is it that we can't code well for CPU counts > 100 or that the types of problems we're looking at right now that need that level of parallelism tend to be targeted towards GPUs or even ASICs? I think I'd need to see the kind of problems you're trying to solve, because I'm not sure high CPU counts are the right answer.
* Talking to GPU-type engines is actually pretty simple, we will deal with it the same way we deal with CPU-type engines: abstraction through a compiler. Compilers over time will learn how to talk GPU optimally. GPU portability over the next 20 years will be a problem to solve as CPU/architecture portability was over the last 40.
Everybody is everybody who has used other languages intensively or is into programming languages or, the most negative parties against JS, people who are into formal methods. But 'everybody'; I get often downvoted to hell for being negative on JS on Reddit. And i'm not using a baseball bat; i'm subtle about it as I don't care for language wars. Use what you want, but please don't say it's the best thing to happen to humanity. But no; 'everybody' (as in headcount) thinks it is the best thing that happened to humanity and that other languages should die because you can write everything in JS anyway.
Even the most ardent JS fans I work with call it objectively bad.
A modern JavaScript engineer would point out that:
- The community is very accepting of new engineers.
- The ecosystem is huge and there are great solutions
available to many problems.
- It's easy to write consistent code while avoiding
many problems with the language if you use `eslint`,
`prettier` and `flowtype`. Tooling on the web is
excellent and rapidly improving.
- You can use one language on the back-end and front-end.
You can even use the same language to write desktop,
embedded, command-line and curses applications.
- `babel` drives the adoption of new features into the
language with greater vigor than many other languages.
- The VM is very fast due to lots of work by Google,
Mozilla, etc. There's also methods of interacting with
modules written in WebAssembly or if you're on the
back-end native modules written in Rust/C++.
- JavaScript is driving a lot of momentum for
new programming languages (Reason, Elm, PureScript, etc)
since as a transpilation target it gives you access
to many environments and libraries at the get-go.
It's dominating because of the sheer drive within the community, and if there were award ceremonies it'd be winning 'Most Improved' year-upon-year.
A community accepting of new engineers isn't. But it isn't something to brag about either, because it never means its literal sense. Instead, people say a community accepts new engineers when it has low enough standards that newbies feel empowered without learning anything new. The one language to rule them all mentality is also about low standards.
The ecosystem is just compensating the lack of native features and bad overall design; so is the tooling. Babel is a great tool, but it's just compensating for the lack of a good VM at the browsers; ditto for all the transpiled languages.
Besides, no JS VMs aren't very fast. They are faster than the slowest popular languages (Python, Ruby), but can't get anywhere near the moderately fast ones like Java and .Net.
An accepting community can be a large funnel: it doesn't need to mean that nobody improves or that there are no selection effects. More candidates means more chaff, but also more wheat.
Babel isn't merely contributing to a 'lack of a good VM'. It's more valuable to see it as a playground and streamlined process for large-scale testing of language-level features. (It solved this problem later on. It was originally meant only to solve the problem of running modern code on older VMs.)
I guess you could argue that the VM should have been complete in 1995 but what programming language has managed this?
Also, I don't think that comparing JavaScript's VM to the VMs of statically typed languages is fair. You mentioned yourself that in comparison to similar dynamically typed languages it's faster. Compare apples to apples.
I don't think I'm bending the truth here. There's been a lot of investment into JavaScript, and not all of the problems that have been solved are obvious or easy.
At this point it is ascendant because it has effectively solved lots of problems that relate to speed of change and it has done this seamlessly on a large number of platforms. I think people look at this and pattern-match to 'first-mover advantage' or 'JavaScript has a monopoly on browsers' but neither of these things were what got us to this place: it was innovation in maneuverability.
(It won't necessarily stay ascendant now that it is so easy to circumvent writing JavaScript but yet still plug into its ecosystem.)
That needs substantiation. It's not immediately obvious in any way.
About Babel, notice how we never had a problem playing with language-level features that target the PC? That's the difference a good VM makes (although on the PC case, it's not virtual). Besides, you are conflating VMs and languages somehow - they have only a tenuous relation.
About the speed, Javascript misses most of the expressiveness of Python and Ruby. It's more on line with Java but if you want dynamically typed languages, with PHP and Perl too. Yet, it's not any faster than PHP and Perl. It has reasonably fast runtimes - those are not an issue for the language, but are not a big selling point either.
Overall, the reason Javascript still exists and is viewed as a serious language is all because it has a monopoly on the browsers. It was worked on enough that if that monopoly goes away it will still be a good choice for some stuff, but it is not an stellar language, and can't be coerced into becoming one unless it can pass through a Perl5 - Perl6 level transition.
Basically, this. ClojureScript and others are simply the efforts of sane, experienced developers that don't want to cope anymore with Javascript and its warts.
In fact, the normal stereotype of JavaScript developers as people constantly chasing new technologies and libraries is actually true, but what you are claiming is the exact opposite: "people feeling empowered without learning anything new". People are empowered and learning new things because there is a low barrier to doing so and this is exciting.
> misses most of the expressiveness of Python and Ruby.
> It's more on line with Java
> Overall, the reason Javascript still exists and is
> viewed as a serious language is all because it has a
> monopoly on the browsers.
In my opinion, the existence of NPM and a broad range of environments that you can run your code on will likely mean that JavaScript would still be a productive environment even if the web was to die.
Could you please give examples of expressiveness of Python or Ruby that Javascript lacks?
Counterargument: Read
Douglas Crockford - JavaScript: The Good Parts
Even though this book is somewhat "aged" (it is from 2008), I know no better book where it is presented so well what is so interesting about JavaScript and how this language is so often misunderstood.
If after reading this book you still consider JavaScript as "objectively bad", so be it. But first read the arguments that are presented in this book.
Good tech mostly speaks for itself. When you need entire books to convince you that a language has good parts, then you know the language has a problem.
What language is not flawed? And why are we "screwed"? I don't get this FUD...there are more important things than language-choice such as dependency management system + community + ecosystem. JS lets you get on with the job and get things done quickly. You need performance - use C/C++ bindings. Its been clear for a long time that JS is the safest long-term choice and is slowly creeping into every other language's castle.
You meant descendants?
Anyway, they won't. They will throw away the garbage and rewrite anything good they find. Just like we do.
Just look at the prototype hell. Look at the quirkiness of the comparison operator. Type inference. The mess piles up extremely quickly.
"Disciplined devs don't make those mistakes" is NOT an argument and never was. Decades later, people believing themselves to be godlike C/C++ devs still make stupid mistakes.
But I guess learning from the past won't happen for people who fanboy and have a survivorship bias. And have money on the line.
To some extent, yes, that happens in all languages. To some extent 10^100 and 1 are both numbers. That does not mean you can even compare them.
I don’t think anyone ever hated type safety, I think they hated verbose syntax and unfortunately conflated the two.
Now we’re finding a happy medium with compiler type inference and people are like, wait what?
For example, functions that return a different type depending on the values passed in are a useful pattern, but not allowed in staticly typed languages, except those with dependent types (Idris) or runtime downcasting (Go). There's always a way to acheive the same thing, but usually at the cost of safety or much more verbosity.
This is actually, IMHO, one way Go strikes a happy balance between static and dynamic typing. It provides a (runtime) safe and low-verbosity way to write dynamically-typed code, i.e. `interface{}`. Rust is slated to get something similar with `impl Trait` values.
Look at this verbose OCaml:
type my_return_type = AFloat of float | AnInt of int | AString of string;;
let my_fun n =
if n < 0 then AString "negative"
else if n == 0 then AFloat 0.0
else AnInt n
# my_fun (-4);;
- : my_return_type = AString "negative"
# my_fun 0;;
- : my_return_type = AFloat 0.
# my_fun 42;;
- : my_return_type = AnInt 42
In a dependently typed language, you could reduce verbosity at the use site as well as avoid the extra runtime branch.
Since Go 1.9 was released with its new Type Aliases feature a few months ago, the verbosity is even lower. By putting `type any = interface{}` somewhere in your package, you can just write `any` instead of the verbose `interface{}` everywhere you want some dynamic typing.
It seems to me that at one point JavaScript had a lot of confusing/bad design decisions but that more recent changes have largely eliminated them. For example, I almost never have to worry about "this" anymore.
I recently worked on a project using TypeScript and I really appreciated how it changed a lot of the bugs from being runtime to compile time. I could definitely see how that is a big flaw, but it seems like the community is developing solutions.
But the thing is, just because arrow functions exist doesn't mean people are going to use them. Similarly, just because modern browsers allow "let" and "const" doesn't mean people will stop using "var."
I agree that JavaScript is generally fine if you stick to the new parts of the language, but that in itself is a pretty big problem. Maybe not for me and you, but for software development in JavaScript generally it's really not ideal.
Either that, or an ASP.Net view where the backend is interacting with the user through a browser. But that doesn't work well. It's much better to standardize the backend API than the entire frontend.
I think SOAP successfully demonstrated that rigorous "API standards" are a pipe dream. Swagger/REST is probably as good as it's going to get.
Which is why what we actually need is a la Delphi
https://www.outsystems.com/platform/
JetBrains is probably a good example of a "Borland" like company.
Outside HN/Reddit bubble that are plenty of companies that are willing to pay for software, the supermarket cashier doesn't take pull requests.
Also, the back to native focus on mobile platforms, including Google having to integrate Android apps on ChromeOS, might make it less relevant, given that native IDEs do offer some Delphi like experience.
Btw Visual Basic continues to exist under Visual Basic.NET and if you stick to the basics you could learn to write C# GUI programs pretty quickly.
I agree writing GUIs by hand is very counter-productive.
https://www.jetbrains.com/help/idea/components-of-the-gui-de...
And integrate with Scene Builder for JavaFX.
JavaBeans are the Java version of VCL components.
Qt doesn't have a component eco-system like Java or even .NET enjoy.
The company behind Qt is still trying to sell QML to C++ developers, looking from the set of talks at Qt World Summit.
But yes, Outsystems (I worked with them and their product quite a lot in the past) could be considered a Delphi. But still not modern; it's rather painful building apps/sites with it that people seem to want.
Jetbrains can be considered a Borland; I didn't think of that because I consider them more in the space of 'low level' programming tools (which, like you say, includes Delphi functionality, but a modern Delphi wouldn't be like the old Delphi; it would need a lot more innovation).
The crucial difference between Outsystems(/Bubble/etc) and Anvil is that Outsystems tries to be a "no-code" environment, and we think that's a mistake (or at least, a different market).
Delphi and Visual Basic proved that writing code isn't the problem - code is the best way to tell a computer what to do. But writing code in five different languages to produce "Hello World" on the web...now that will slow you down.
(Count 'em: JS, HTML, CSS, backend eg Ruby/Python, SQL. We do everything in Python, which gets you going much faster. We just got back from PyCon UK, where among other things we got an 8-year-old building database-backed web apps in an afternoon. That's the sort of thing that used to happen with VB/Delphi.)
Full visual development, runs on a scala runtime, can be deployed (visually) on cloudfoundry, docker/kubernetes, etc.
There are already some attempts at it, for example https://anvil.works/
Agree 100%. However, i find Javascript pretty good for quick and dirty "MVPs".
I guess you're referring to languages that are not-quite-without-side-effects, but I'd say the biggest influence the functional paradigm has had on other (imperative) programming languages is actually the addition of higher-level data manipulation operations. The functional utility libraries you see for languages such as Python and JavaScript exist solely because sometimes functional idiom like "map this onto this" or "take that only if this holds" or "let's make a new function by pre-filling these function arguments" is more intuitive than having for loops all over the place. And it mixes just fine with other imperative code.
Interestingly, on Haskell-land the debate is about merging "do notation" and normal notation. On practice the difference is limiting, and arguably the types alone are enough to represent the difference.
https://wiki.haskell.org/Do_notation_considered_harmful
More constructively, it could be argued that IO just does not fall in the remit of functional programming. SPJ doesn't refer to Peter Landin much, but Landin's perspective is valuable. In 1964 he wrote "Is there some way of extending the notion of [arithmetic expressions] so as to serve some of the needs of computer users without all the elaborations of using computers?" (The Mechanical Evaluation of Expressions).
IMO this is exactly what the Haskell tradition is pursuing. Landin, quite modestly, doesn't anticipate that entire computer systems will be expressible in functional form. Implicit data structures are just one example of a mechanism that is incompatible with the idea that everything is an expression whose internal representation is managed by the language runtime (itself presumably involving mutable data).
I don't think Landin was experiencing a failure of vision--I think he saw clearly that some of the "elaborations" of computing are remote from a functional model.
Follow up questions bc I'm still a Haskell noob:
Do you know if async/wait from Control.Concurrent.Async ensure linear sequencing?
Could the sequencing problem be solved if linear types make their way into Haskell? (They seem to mainly be about memory management, but I'm not sure about other potential applications)
I tend to agree. The two big wins from a more “functional” style, from my perspective, are the clear emphasis on the data and the way effects are more explicit and controlled.
I want things like higher order functions and algebraic data types and powerful interface/class systems. With those I gain many useful ways to represent and manipulate data that I don’t have in most languages today.
In a world where most mainstream languages are just discovering filter, map and reduce on their built-in list types, a language like Haskell gives me, out of the box, tools like mapAccumWithKey that work with any data structure as long as it provides the specific, clearly defined interfaces required for the algorithm to make sense.
In a world where most mainstream languages are worrying about accidentally derefencing nulls or whether there’s a proper type for enumerating a set of values, functional languages routinely use algebraic data types and pattern matching, and some go much further.
Arguably, these aren’t really functional concepts at all, in that you could have them just as well in an imperative language. However, in practice it is the functional-style languages that are far ahead in these areas, because they are a natural way to work in languages that emphasize composition of functions and careful, explicit handling of data.
I also want to know that I’m not applying effects on resources unintentionally, or sharing resources without proper synchronisation, or trying to apply effects on resources in an invalid order, or failing to acquire or release resources properly, or leaving resources in a mess if something aborts partway through an intended sequence of effects. This aspect goes a lot further than just making data constant-by-default, but it certainly doesn’t require trying to remove state and effects altogether. These things aren’t so much about making my code more expressive but about stopping me from making mistakes.
I want a language that will stop me from accidentally modifying a matrix in-place in one algorithm while some other algorithm has a reference to that matrix that it assumes won’t change. I don’t want a language that will stop me from ever modifying a matrix in-place. Sometimes modifying things in-place is useful.
I want a language that will be explicit about the initialisation and lifetime and clean-up of a locally defined cache or temporary buffer. I don’t want a language that tells me I can’t cache a common, expensively computed result 15 levels deep in my call hierarchy without changing the signature of every function on every possible path to that point in the code, or a language that will let me do whatever I want but only if I use some magic “unsafe” keyword that forfeits most or all useful guarantees about everything else in the universe as well.
In this respect, my personal ideal programming style for most tasks very much would be a hybrid of imperative/stateful and functional/pure styles, with the key point being that the connections between them should be explicit, obvious and deliberate.
I don’t know much about Rust so I can’t comment much on that. The intended point of my example was that in a purely functional environment, you can’t have a local, low-level cache, because updating a cache is inherently stateful. So either you need something like a mechanism that lets you break the rules locally, like say unsafePerformIO in Haskell, or you need to infect your entire call chain with whatever mechanism you use to manage effects top-down. While that is perfectly reasonable, given the constraints you choose by adopting a purely functional language, I don’t think it’s particularly helpful.
My conclusion was that I’d rather have a hybrid of imperative and functional styles, contrary to the suggestion in the original presentation and in general agreement with stdbrouw’s comment.
They are purely functional in the sense that NO VARIABLE inside your code can be ever mutable.
However, they do have ETS (which is an in-process cache inside the VM) which is fully mutable and people have long made wrapping libraries around it for transparently working with mutable arrays, double-linked lists, queues, matrices, graphs and what have you.
The philosophy basically is "always work with immutable data except when mutable is more performant or is otherwise more practical". They don't shut the door on you, they just force you to make your intention to work with mutable storage very explicit and clear. That helps a lot when you go hunting inside your code for side effects, too.
As a maturing Elixir dev I can say this philosophy works incredibly well in practice. Code is smaller, much more readable, you don't worry about side effects like ever -- except in very, and I mean VERY RARE occasions (for 1 year of working with it I only had to do that twice) -- and finding a bug is times faster compared to Ruby, Javascript, PHP, Java.
(a) strong visibility of and control over effects,
(b) an emphasis on powerful tools to represent and manipulate data, and
(c) straightforward imperative/stateful aspects when required.
Of course that isn’t to say that no such language exists, but if it does then sadly I have yet to become aware of it. New ideas are always welcome…
For example you can do functional, imperative and OO programming in Common Lisp, and it brings extensive features for working in all these three paradigms.
More recent examples, Racket and Julia.
Now, I am by no means a Lisp expert, so it’s entirely possible that I’m completely unaware of something here. However, I’ve yet to encounter much of an effect system in any flavour of Lisp. Indeed, it’s hard to see how the sort of explicit visibility and control of effects that I’d find useful could be achieved in a language with primarily dynamic typing using any of the approaches I’ve encountered so far.
> - Functional is OK. Imperative is OK. Both in the same program are a mess.
What do you mean by that?
My experience is that functional alone is impossible, since the only useful thing a program can do is through state changes; imperative is a-OK, and functional+imperative in the same program is the best way to do things (i.e. well-defined stateful areas surrounded by lots of functional code).
Some programs are simply pure functions. If your program entry point accepts a string and returns a string, then you can write useful things, such as compilers, image processing, grep, etc, entirely as pure functions.
All purity does is force you to change the type of a function that wants to log something, to reflect that it’s returning a value that, when evaluated at run-time, will have the side-effect of printing to the console. It’s still a pure function since it returns the same description given the same argument(s), the only difference is that when this description is evaluated by the run-time system, something will be printed to the console (in accordance with the description).
In other words, you’re not restricting the effects your program can have, only how you can choose to describe them (as first-class values whose evaluation has a side-effect that is not observable by your code).
It’s all explained perfectly in this talk: https://www.infoq.com/presentations/io-functional-side-effec...
You can't counter that with an example of a program that you can't write as a pure function.
The idea of pure functions is a false friend in CS because it tries to solve the problem of state by wrapping it up and wishing it away.
It's true that state causes a lot of problems, but so many useful systems rely on mutable state - at practical application levels - that it might be interesting to design robust systems that manage state, context, and relationship instead of trying to create contrived examples of state-free systems.
There seems to be a cognitive bias against this in CS. Most developers appear to love puzzle systems made of hard little components with solid edges, and thinking in terms of context and relationships seems to be disturbingly alien. So there aren't many programming paradigms that explicitly work with contextual inference instead of trying to rigidly delimit interfaces.
But there are real prizes to be won from associative context-smart computing, and IMO the domain is wide open for innovation - because it may be possible to give up the pursuit of complete safety and predictability (which doesn't exist anyway) in return for new kinds of powerful, productive, and smart features.
I don't understand why any of this means you can't have a compile as a pure function. A pure function can cope with things changing internally - it just creates new data structures to represent things that have changed in the old data structures and then passes the new data structures onto the next phase.
A pure function just means you can't do something like a package manager that needs to read files from disk or download things.
> but so many useful systems rely on mutable state
You don't have to argue this to me - I wrote the first half of my PhD on the importance of mutable state.
I'm just arguing against the nonsense that it is impossible to write a useful real-world application as a pure function.
I work in the VM research group at Oracle, and guess what? Our JIT compiler is basically a pure function. It takes in some bytecode and produces some machine code. It's structured a little differently in reality, but it is logically, and almost in practice, a pure function from one to the other.
I'm writing a presentation about this right now.
I also agree - context and state are important and useful, because many real-life problems or processes rely on such a model and thus translate naturally when state is a "first-class citizen."
Fair enough.
However, they have a mutable in-process cache (living in the BEAM VM) that many people have written libraries around for stuff like mutable arrays, matrices, graphs, and many others.
It goes like this: do 99.5% functional programming and you have the imperative / stateful tools for when they are absolutely necessary.
This works. Extremely well.
+100
There have been a few attempts in this direction, but they have mostly been couched in the form of a whole new language that also embodies at least a half dozen other novel (i.e. unfamiliar) ideas as well. Contra the OP, I think this is an area where incrementalism does work. Extending a language people already know with a few constructs for IPC, much like fork/join or async/await have done for concurrency, is much more appealing. I've been thinking about this for a few years now. Maybe I should write some of that down and let people pick at it.
I'm really interested in hearing more of your thoughts on this, since it touches on one of my personal research interests. What kind of language support for IPC are you looking for? Something in the vein of session types [1], which checks that two parties communicate in a "correct" sequence of messages?
I'd suggest offering two forms of marshalling - strongly typed and non-typed. Strongly typed marshaling means sending a struct to something that expects exactly that struct. That will usually be another program which is part of the same system. Structs should be able to include variable-length items for this purpose, so you can send strings. Checking involves something like function signature checking at connection start. This should have full compiler support.
Non-typed marshalling includes JSON and protocol buffers. The data carries along extensive description information, and the sender and recipient don't have to be using exactly the same definition.
Both are needed. Non-typed marshalling is too slow for systems which are using multiple processes for performance. Typed marshalling is too restrictive for talking to foreign systems.
Human readability however has always been the selling point of terrible formats like XML and JSON.
Not sure how can a binary-compact format ever account for that. Maybe excellent cross-platform tools that allow you to inspect and modify the binary-compact format wherever it is? (I mean not only standalone CLI and GUI software; I also mean native browser support in the Dev Tools space and going down the line in the future -- transparent support for the format[s] natively in the programming languages / VMs themselves.)
I think you mean “lack of memory safety” rather than bugs. Garbage collection doesn't magically free you from finalization bugs, it just makes their consequences less disastrous.
> Formal methods are still a pain. The technology tends to come from people in love with the theory, resulting in systems that are too hard to use.
Clearly, the solution is getting rid of people, but it isn't entirely clear on which end to get rid of people.
Async + Event driven works transparently with higher order frp with the usual tradeoffs of first order vs higher order frp.
Multi threaded + Event driven seems like a mess as soon as state gets involved. Some transaction based system might be interesting.
It's almost like the "problem" isn't the languages themselves, it's the tooling around the languages. I want to be able to do a lot more than "compile", "run", with my code... I can imagine machine learning driven tooling being able to automate a lot of the mechanical aspects of code writing beyond the simple generate a closing bracket that an IDE can do...
https://dlang.org/spec/unittest.html
It was pioneered in the context of logic programming and Prolog by Ehud Shapiro in his 1982 PhD thesis "Algorithmic Program Debugging". The thesis was published as an ACM Distinguished Dissertation by MIT press and is available online from:
http://cpsc.yale.edu/sites/default/files/files/tr237.pdf
Together with Leon Sterling, Ehud Shapiro later wrote a very important introductory Prolog book called "The Art of Prolog".
reverse "abcde" = "edcba"
Also, this is really similar to how Idris and Agda work, except they use expressive types to generate the code (using the Emacs modes) rather than test cases.
[0]: http://nautilus.cs.miyazaki-u.ac.jp/cgi-bin/MagicHaskeller.c...
> tests without cluttering the production code and excessive mocking
IMHO, tests should be an integral part of production code. In MPWTest[1], tests are typically expressed on the class side of the class under test, in a category called Testing, though it's easy to override that. This solves 2 of my major annoyances with xUnit-style testing: dual class/test hierarchies and (with static typing) the need to add public interface for the tests, which don't have privileged access.
Mocking in particular and stubbing should be eliminated as much as possible, but this is not a programming language issue[2], more a "not making architectural assumptions too early" issue.
I don't think Spec alone removes the need for mocking, using Clean Architecture is probably more important for that https://8thlight.com/blog/uncle-bob/2012/08/13/the-clean-arc...
If you structure your app so that any IO lives at the edges, it becomes possible to test all your business logic without mocking. Spec can help at that point by providing generative testing.
Since Spec
Technically, these could very well be simple commented annotations that point to a separate .test file or something, but are shown in-line by your favorite IDE or code-editor.
> I'd also like to specify various injection points for tests right as I'm writing the code
You mean, like methods/functions?
> extract some arbitrary subset of the code and surround it with tests
You mean, like modules/classes?
- runtime state you need to initialize if your code is written in a stateful manner
- other methods/functions in different modules/classes the code you want to test calls out to
- the fact that method/function and module/class separation is (and IMO always should be) primarily driven by needs of production architecture, not testing, means that it may not be perfect for testing needs
Add to that the failure of testing tools and methodologies (like "don't test private methods" meeting "separate out duplicated or complex code into private methods"), and I feel the problems are real.
----
Here's an idea that just popped into my head right now: how about "hash-based testing"? You take a code you want to test, like:
private Integer foo(String bar, Frobnicator quux) {
String frob = quux.invokeMagic(bar);
return memberApi.transform(frob);
}
Method cut = <<
private Integer foo(String bar, Frobnicator quux) {
String frob = MOCK(quux.invokeMagic(bar)) AS("frob");
return MOCK(memberApi.transform(frob)) AS (frob.length() > 3);
}>>;
//continue testing cut()
This way, you end up being able to mock everything precisely, inline, in whatever way you see fit, with your tooling ensuring that the actual code stays in sync with the test, since whenever the original method changes in any meaningful way, you'll fail the code "equality" test.
Might be a stupid idea, I welcome comments.
(INB4 testing to implementation instead of the interface - if you have to mock anything, you're already testing to implementation, and this way you can inject testing alterations precisely, instead of having to turn your architecture inside-out to support IoC/DI/whatever the current testing-enabling OOP fad is.)
Text substitution is impossible to reason about; magic test frameworks sound like a good idea because every kind of magic sounds like a good idea in isolation, but they'll always end up getting you in trouble. Plain old code, plain old functions, parameters, and modules, always win in the end, because every bit of brainpower you can save from understanding magic is brainpower you can spend understanding the details of your actual test instead.
That said, I didn't mean text substitution, but AST transformations (or rather, AST comparisons, since my example is defined in terms of AST equality with the MOCK magic thingie making the particular part of AST be ignored for the sake of comparison).
So really the idea that a good design is also testable is somewhat of an approximation of reality, It has some elements of truth but at some point making your code more testable actually makes it worse, often you need to make other compromises. Then the interactions of pieces of code typically happens in much more brittle system/integration tests. YMMV, this is just my experience.
A lot of interesting pointers in the replies, thanks!
When I was at google, one of the hardest thing I did was look at the marginal intern review scores and tried to pick out the ones from the big pile whom we should look at again and who should not be looked at. There were all kinds of crazy ass stupid interview questions that in my opinion were not very useful to classify capability.
Or they couldn't recognize failure
I wish we were better at communicating about programming languages. I wish we were moving faster. But despair is unwarranted. We really are in a much better place than 10-20 years ago, and the next 10-20 years look set to bring more improvements.
To me it seems we are catching up with the past, and as someone already programming on those environments, it looks we have spent 10-20 years loosing our tools, educating the masses, only to get a taste of things used to be.
The language gap is relative to where he thinks we should be aiming for, what the author thinks is possible. If you think that we're already "there", then there is no gap for you, and there's nothing wrong with feeling that way.
Personally, I agree with the author: we can do vastly better. I think it's a failure of imagination and effort, not potential.
I think JavaScript does some amazing things, but it reminds me of Visual Basic in the 90s or Flash in the 2000s. Anyone could write code (some good, more horrible) with it and do cool stuff.
But the maintainability is horrible. I hope TypeScript comes to the rescue!
We continue to struggle abstracting complex problems using functional decomposition, structured analysis, information (data) modeling, and object-based decomposition.
Many newcomers I meet only know modeling problem domain concepts as data in a database, with behavior and constraints acting on that data in a separate layer, organized using functional decomposition. Of course that layer increasingly approaches a 'big ball of mud' as size and complexity increases. Sounds a lot like we are back data-flow modeling as so popular in the 1980's, in a new guise.
A focus on programming languages in my opinion, masks the real issues we face.
Indeed, major problems of programming languages can hardly be solved in the area of programming languages itself as it is being done now.
I would say that one needs a new programming or computing model so it is not about languages. At least it is my conclusion after 10+ years of research and attempts to develop such a new programming paradigm. And although I have quite significant progress (concept-oriented programming), the more I do and the deeper I go, the more fundamental problems I meet. And these problems are not about programming languages at all. It is more about "how a system works", "what is a system", "what is computing" etc.
Otherwise it's endless and one loses motivation.
language constructs <--> design patterns
In fact, these design patterns and language constructs come in (anti-phase) waves.
Erlang led to Microservices and the Reactive Manifesto. the Open Telecom Platform is still a very modern Microservice library (in my opinion) after more than three! decades
ML led to programming against generic interfaces (http://www.cs.cornell.edu/courses/cs312/2006fa/recitations/r...)
We then have Chinese literature and Indian literature, and these have very different paradigms. Perhaps it is even hard to effectively translate from Chinese to English. But the actual variance within a given language is still greater than the variance between the languages.
And like human languages, the vocabulary is often arbitrary. In human languages a dog is called a dog not because it makes the sound "dog" and not because it looks like the letters 'd', 'o', and 'g' but for entirely arbitrary reasons of phonetic shift and arbitrary initial designation.
And in lisp you take the 'car' of a tuple and in Haskell you take the 'head' of a list. But the two concepts are very much the same. However, the distinction between continuation passing style (CPS) concurrency and the actor model can be expressed in either langauge just as well. And the distinction between CPS and the actors is FAR grater than the distinction between calling the basic function "head" or "car".
Software and business systems are diagrammed with totally ad-hoc “flow charts”, bubble and arrow diagrams, and less ad-hoc sequence diagrams and UML diagrams. We need advances in formal ways to model concurrent processes, from the level of threads to concurrent business processes.
Modeling the above in a business sense, with support in existing paradigms and languages is already a solved problem, just a little known one.
---
[1] The other type of system in industry and commerce being the continuous system that isn't based on recorded events, but instead on logging errors and abnormal circumstances, e.g. an elevator control system or an automated warehouse delivery system, or the engine monitoring software in your car.
[2] Recorded for business or legal reasons.
[3] It's those constraints that prevent an elevator from moving with its doors open, or billing for that product not shipped, or allowing someone to vote twice.
And while them recognition of planned response to events as the unifying abstraction for system analysis is not new (it's at the heart of the structured analysis/design approaches of the 1960s and 1970s which, while displaced somewhat by object-oriented analysis and design, are echoed in more recent tools and technologies like BPMN and Flow-Based Programming, though the analytical methods haven't come back as much as the tools have.)
But, no, modelling for both validation by domain experts who are not computing experts and implementation by developers who are not domain experts is not a solved problem in the general case.
I do agree that a lot that has been learned that is useful in finding adequate solutions to that in specific cases is not as widely known as it should be.
They can do this because they know a set of higher order patterns, beyond just a knowledge of "objects" and "classes". Beyond just a design approach of identifying classes as the "nouns" in a problem domain.
These higher order patterns represent those business processes, as you call them.
Sadly such a skill is very rarely seen, or even in demand, given that data modeling is the basis for most design today.
You mention a "domain object model". Are you talking about diagrams that capture much of the dynamics/asynchronous behavior that I described (in which case, where can I learn about such a diagramming language?) or is it more restricted to the nouns?
Those technical concerns are dependent on the non-functional requirements of the particular required solution. This includes aspects of concurrency.
I suspect your interested in a "framework" that supports both problem domain modeling and specific technical circumstances?
Compiler would infer the types, and the programmer would read it and say, oh, I agree with this type, but I disagree with this type, that's perhaps wrong, this should be rather that type. Then the compiler would infer types again, based on programmer's incremental input.
Data structure selection. Programmer would say I want a sequence here. The compiler would say, I chose a linked list representation. The programmer would look over it, and disagree, saying, you should put this into an array. And compiler could say, look, based on measurements, array will save this much space but list will be this much faster.
Code understanding. Programmer should be able to say just, I don't know what happens here, and the compiler would include some debug code to show more information at that point.
Or take refactoring. Programmer would write some code, computer would refactor it to simplify it. Then programmer would look over it, and say, no, I rather want this, and don't touch it, and he would perhaps write some other code. The compiler would refactor again...
But all this requires that there is syntactically distinct way (so that perhaps editor could selectively hide it) to specify these remarks in the code, both for computer and programmer. So each of them should have a special kind of markup that would be updated at each turn of the discussion. Because you don't want to just overwrite what the other side has just said; both are good opinions (which complement each other - human understands the purpose of the code but the computer can understand the inner details much better). So, to conclude - I wish future programming language would include some framework like this.
[2]: https://isomorf.io/
To some extent, this is the promise of object-oriented programming, that in this particular instance has failed a bit in mainstream languages. It's true that in, say, Java, you have to massively refactor your code to switch between arrays and linked lists, because you use different syntax ([] vs. method calls) to access elements. It could be a bit better in C++ due to operator overloading; you could hide your actual container type behind a typedef, and as long as both container type A and container type B support the same operations with [] that you actually use, you can freely switch between them.
In Smalltalk, every container class derives from a single Collection class, and they have very very similar APIs. There you can, to a large extent, just program against the Collection API and not care much about the actual type of collection you have in your hand. You still have to choose one! The compiler won't do it for you, not in the way you envision. But the idea is that if you program against the generic API, then profile/benchmark your code, it should suffice to change a single line of the program to try a different representation to compare against it. (Of course some things won't work. You can't index into a set.)
Other dynamic languages should be similar, Python for example. But I think you have to work harder to achieve full genericity.
On code understanding, what makes it better than the programmer inserting the debug statements themselves? It saves some misunderstanding from the computer's part.
On refactoring, some IDEs do that. I'm on the fence about its usefulness.
I know a bit of Haskell and want to look at Idris, someday.
My point was, there should be a clean (best if even syntactic) separation between code itself (i.e. what should actually be done) and its properties (like types). Also, because there are two points of view about the properties (human and computer), this separation needs to be there twice (so for example, each type could be specified by computer and by human). I haven't seen a system that would do it, on a systematic level in the programming language. I only gave examples to show where it could be used.
Now I get the entire programming-compiler conversation. It is interesting. I can see some potential there. Yet, I shrug when I think about all the sparse metadata that I will have to check once I discover a low level bug (there is a reason I'm not programming in Smalltalk).
Somehow the best place for all that stuff to live is right there at the source code. That means the compiler (IDE) should be editing your files, so it better have a great integration with your version control system.
That's what I am saying, and that's why it has to be syntactically distinct.
Also it needs to be clear what was input by computer and what was input by human, that's the second distinction. Because of how the conversation works. You don't want computer to erase human input, but the result also has to be logically sound. So you need to know both inputs for the comparison and synthesis, which happens at each conversation turn, both human's and computer's.
And that's what "type holes" and similar systems lack - they only record the result of the synthesis, not the two different opinions. Which is IMHO wrong.
I will contest that one too :)
As long as it is interactive enough, and the history is well marked, there is no reason not to rewrite human code.
Email me at glmitchell[at]gmail if you want to chat more about this.
Kotlin does type inference. You can see what the inferred type is if you enable inline type hints. It's not a part of the source code, but it looks as if it is (modulo styling). As you work with the code you can see the inferred type change in real time.
OK, data structure selection, it won't help you pick between a linked list and an array. That said, I'm not sure that feature would bring much benefit to most programs. You almost always want arrays.
Code understanding: if you're unsure what's going on at a particular point, you can just add a breakpoint and run the program. You can then explore the contents of the program, evaluate arbitrary expressions, you can add "evaluation breakpoints" that print out the values of those expressions without stopping the app (on the fly ad hoc logging, in effect), you can investigate how the code you're looking at relates to other code, what the data flows are, what the location in the type hierarchies are, etc. There's a lot of ways to look at the program.
Refactoring; this is the point I went down this train of thought. Because a modern advanced IDE like IntelliJ can do this sort of thing already. It can do things like spot code duplication and fix it for you by extracting a method, in real time, with a single keypress triggered by subtle visual hints like a soft wavy underline. It can convert code between imperative for-loops and functional pipelines of map/filter/fold/etc, in both directions. It can identify and automatically delete unused variables, function parameters, object properties and so on. "Simple" is somewhat in the eye of the beholder, but it's a pretty close realisation of what you seem to be asking for.
The dialogue is not had through markup in the code but rather, through the IDE giving its suggestions using visual hints, and the user starting an interaction through a keypress that brings up a menu of suggestion options ("intentions"), which may in turn lead to more options and so on.
This is possible today, at least in Haskell. (and I'd guess in OCaml too ... surely also Scala?)
Your proposal for adaptive data structure selection based on benchmarking is intriguing!
In short, the near future of PL is great, and exciting stuff keeps happening. Don't believe the naysayers.
At last, someone noticed! ;-)
Though from that expression, what the author doesn't seem to grok is why having a Smalltalk-like environment is desirable; maybe not as the primary way to program computers but certainly as a tool alongside.
It's a shame that a family of programming languages that build on and expanded that model hasn't gained more traction in the industry (not necessarily for people who dedicate their lives to build complex software with a highly general programming language, but for the rest of us).
And it's not useless; we probably wouldn't have had Elm without Bret Victor's Inventing on Principle[1][2]. And there has been some progress in that direction of interface design of, for a lack of a better term, "programmable environments": look at Apparatus, for example[3]. Where would you even fit that on these slides?
[0] http://worrydream.com/ABriefRantOnTheFutureOfInteractionDesi...
[2] https://vimeo.com/36579366
[3] http://aprt.us/
End-User Development has lots of under-explored ideas on how to build software automatisms that don't require the end user to learn a hard formalism (even if such formalism exists as the basis for the system). Though I understand that programming language theorists are not interested in that angle of the evolution of PLs.
The engineering break through will be when we have a scripting/evolving system that can be more easily distilled into sealed systems. So people poking around in a spreadsheet will be able to turn that into a reliable application.
IMO, TDD and BDD could be seen as attempts to do exactly this.
If you've ever had to make another person's research or prototype quality code more robust it can be easy to wonder how on Earth they work the way they do. Having this dichotomy in mind makes it easier to have more sympathy.
It works both ways. Researchers need to understand why platform development moves so slow. This dichotomy helps reason about it.
But you're right, between evolution steps the system is not sealed. That's a concern and why projects like Debian are keen to have reproducible builds. It's also a big part of why containerization is becoming so popular.
The containerization movement is interesting - while having these 'sealed' virtual environments and reproducible system builds is the right direction, I'm amazed at the size and weight of each of these.
But with things like LD_PRELOAD, different .so minor versions, etc, a compiled program suddenly straddles the categories. Hence my comment about reproducible builds and containerization.
Yeah, I see that as a likely evolution as well.
In some ways, an IDE is already an exploratory system with a focus on delivering a sealed system - the final compiled software represented by the source code. I believe this combination is what made IDEs so successful and universally desirable, and future programmable environments should share these traits.
Auto-completion, refactoring and code-browsing tools support the exploratory work. The problem is that IDEs are not so good at exploration as a REPL because they're unable to store intermediate results and states. TDD could then be seen as an ad-hoc way to store such intermediate state.
The other thing that would help, and my personal wish, is better simulation and the ability to see indirect effects of changes - instead of only showing what is directly modified (i.e. a line of code) and leaving the simulation up to the human.
Yes, exactly. You get that with a REPL (e.g. Lisp or Python) and copy-pasting to the source code files, but it's far from ideal.
> The other thing that would help, and my personal wish, is better simulation and the ability to see indirect effects of changes
So, sort like explorable explanations[1][2], but aimed at developing a new model and not just exploring one that has been already made? That's what I'd like to see, too.
Copying something I wrote in another thread: I want to see not just the immediate program I'm manipulating (text or otherwise) but also the implications - the affected parts of various other intermediate representations all the way to the running, live system that will be affected.
Moreover, having direct introspection of a model stored in a local notebook is still quite limited with respect to having it in the whole environment like Smalltalk or HyperCard did. There are a few systems trying to explore that "structure of the project" approach, like Leo Editor or the Smallest Federated Wiki which could be a better basis for a "Inventing on principle" tool.
"The thesis was that programming in 2030 will have very advanced research languages but mainstream languages will effectively stop advancing and we'll be last [sic - left?] with a vast insurmountable gap between the two."
That provides some context for what the slides are about.
But: that language might be English (or other natural language). Which is to say maybe computer languages are leveraging the human "language faculty" and rightly so. Which may mean that some languages are always going to feel more foreign.
Kotlin and Ceylon both came out of corporate/industrial development. C#, JavaScript and Java continue to evolve entirely due to industrial funding.
I agree that academia is not a great place to do PL research as all the other non-idea aspects of a successful PL don't get enough attention. But industry is doing OK at launching new languages. Perhaps he feels that's "incrementalism" but there's nothing wrong with making solid upgrades that aren't attempts at unicorn style magic cure-alls.
It'd also help if we collectively got over our demand for our tools to always be open source. This has solid justifications behind it but it ensures that programming languages will always be spinoffs of other endeavours rather than having truly focused attention in their own right. When you take away the profit motive, you get stagnation, and that seems to be the main point of the presentation.
I couldn't agree more but I also am not as pessimistic as he is at its chances of happening.
Interestingly Scala has come from Academia, Industry, and Hobbyists. And for me it's already the next great programming language. Yeah, it has some warts and is hard to learn but that's true of all great things. :)
That's an interesting point. The slide implies there will be no next great programming language because academia nor industry nor hobbyists can deliver it, but overlooks the possibility of a combination working together. For example, Rust started as Graydon Hoare's hobbyist language and then development was sponsored by Mozilla.
Now, whatever innovation this proliferation of languages may bring, it also brings fragmentation which acts as a counter-weight on the value created for the industry as a whole.
Whether the net effect is positive or negative is very difficult to tell.
When given enough time, if math can stay pure and adequately abstract to eventually describe our entire universe concisely. Shouldn't code similarly have no limitation?
> The UPenn dependently typed Haskell > program shows a great deal of promise and > is likely to manifest a decade before other > DT languages generate practical backends.
I don't agree with this at all. The design problems with tacking on dependent types to an existing system are massive - it's a research problem for a reason. On the other hand, writing a "ghc quality" backend for Coq/Agda/Idris/F* seems difficult, but at least it's an engineering problem instead of a rough idea.
In particular, CertiCoq is a compiler for Coq written in Coq, and the main problem here is verification. Simply writing a compiler is no harder than writing a compiler in any language.
> Interesting ideas out of Microsoft Research on SMT solver > directed programming editors that enforce invariants and can > generate code during development.
Another interesting Microsoft Research project along the same lines as Dafny is F. Both are nice, but F is closer to modern dependently typed languages.
> Lots of non-local reporting problems associated > with using unification during type-checker.
I would argue that the problem is that we use constraint solving for type checking. For example, strict bidirectional type checking leads to more tractable errors, since it's straightforward to follow the compiler's reasoning. On the other hand, bidirectional type checking is less powerful, so it's not like there's a silver bullet here.
> Type-safe OTP.
I wish more people were working on things like this. There's some theoretical work on calculi for distributed systems, but once they encounter the real world they inevitably become horrendously complicated.
In practice, though, a lot of people quote this who don't know what options are available, what their strengths and weaknesses are, and what the problems of the job will be, and so when those people quote it, it's just a cliche.
Gambit Scheme: A fast scheme implementation. http://github.com/gambit/gambit
Gerbil Scheme: Provides full module and syntactic tower on top of Gambit Scheme. https://github.com/vyzo/gerbil
(My only exposure to it is from Andy Wingo's blog, and I don't know much about Scheme.)
1. Yes, if the extent of the imagination is languages like Idris and ideas like effect systems, that follow a gradient descent from Java, and always in the same direction: being able to express more constraints. What you get "for free" from such languages may not be significant enough to justify the cost of adoption, and the valuable stuff you can get is not much easier than the options available today, which are too hard for anyone to take. If you were to consider truly novel languages that think out of the box (e.g. Dedalus/Eve) then maybe one will stick and make an actual impact rather than just a change in fashion.
2. How do you even know that we can do much better? NASA engineers may not like it, but they don't complain that we're "stuck" at sub-light speeds. Maybe Brooks was right and we are close to the theoretical limit (that we know must exist).
> We talk about languages as a bag of feelings and fuzzy weasel words that amount to “It works for my project”.
Can you find another useful way, available to us today, of talking about languages?
> “Use the right tool for the job” Zero information statement.
That's right, but it's not a dumb cliché so much as it is a tool we've developed to shut down religious/Aristotelian arguments that are themselves devoid of any applicable, actionable data. One, then, is often confronted with the reply, "but would you use assembly/Cobol?" to which the answer is, "of course, and it's not even uncommon, and if you don't know that, then you should learn more about the software industry before giving any more Aristotelian arguments."
Idris can often infer entire functions from their types if the domain is amenable to accurate type-based specification. For instance, taking the common "sized vector" example, where `Vec n a` refers to a length-n vector of values of type a, functions like
zip : Vec n a -> Vec n b -> Vec n (a, b)
This is a great example of what I'm talking about. The kind of functions Idris can generate is that of functions that you could manually write with only marginally more effort -- if that -- than the effort required to write the precise type. I don't think those functions ever form a significant portion of a significant program, if at all (those functions must be so generic, or else there would be a search problem that Idris can't solve, that they would already likely be in the standard libraries). When Idris is able to generate an efficient sorting function given some constraints, then I'd be impressed.
> can be automatically generated via the interactive "proof search" mechanism
If you've spent a significant amount of time with such proof assistants, you'd see that the proof search is rather pitiful.
Afaik, Idris' proof search was hacked together in an afternoon just to see if it would work. And it did, surprisingly well considering. Don't know how much it has been worked on since then though. But yes, still leaves a bit to be desired compared to other systems.
I don't think that's a fair reflection of what these languages feel like. When your constraints mean that only the right thing can be done, you can make the thing implicit. The things that are tacit when humans talk about the problem can be tacit in our programming as well. You have ways to factor out concerns that are separate, even when they are entangled. "Better decomposition of the problem" doesn't sound exciting, but ultimately that's the very essence of programming. I don't think we need new grand ideas, at least if we're just talking about the next 50 years of industrial programming; I think we need to follow through on the ideas we already have, and I actually think that gradient ascent from Java will happen, and will move the industry forward, even if it ends up being a lot slower than I'd like.
> How do you even know that we can do much better? NASA engineers may not like it, but they don't complain that we're "stuck" at sub-light speeds. Maybe Brooks was right and we are close to the theoretical limit (that we know must exist).
I think we can go at least one or two substantial steps better than Idris: I want Noether's levels of stratification (beyond just total/not-provably-total), and I want Rust-style linearity, not as a specialised capability but as something integrated into the regular type system. But even if Idris is the limit, it's still a big step ahead of Java, and it's going to take the industry a while to make that step. (Maybe it will have to be two or three steps to get there - it feels like most of the industry is only just adopting the things that OCaml offered, and not yet familiar enough with them to appreciate what Haskell offers above that. And I can't blame them - I only have an appreciation of those things because I've been able to work for a long time in Scala and very gradually work my way up towards using the less familiar facilities). "The future is already here — it's just not very evenly distributed."
> Can you find another useful way, available to us today, of talking about languages?
Yeah, that's the problem :(.
The whole industry is a huge iceberg in terms of what's visible about the practice of programming versus what's actually happening. I'd like to see more programming happening in public - which I think probably means more publicly-funded programming (I was thinking about this a few days ago in the context of research and scientific code - we have a relatively good system in place for providing grants for research, but we're a lot less good at funding the infrastructure that's used for research - we end up with a lot of researchers repeating the same work because they're only ever funded for a specific project, rather than being funded to produce something that's useful for many research projects). What makes sense in an open-source webdev project is all that most of us ever see about most languages, and while it's a lot better than nothing it's still a seriously distorted picture of what most programmers are actually doing.
Except that sometimes, when you think about the practical issue long and hard enough, you can find that some constraints need not be stated at all, and can be completely factored out of the language, at least in most cases. For example, GC removes resource management out of the equation (it's not always the right solution, but it is the right one in a large enough portion of software). Or, while FP people pore over effect systems and how to constrain effect, languages like Eve that adopt synchronous programming (which is based on ideas from temporal logic), get a mathematical framework in which effects and computation are one and the same, and effects don't pose a challenge to begin with.
> I actually think that gradient ascent from Java will happen, and will move the industry forward
Perhaps, but I doubt it's in the direction pointed to by Idris. We have much better ideas around. The author of the slides mentioned only the segment of PL research that he find appealing. Other segments may have some more promising ideas.
> I think we can go at least one or two substantial steps better than Idris: I want Noether's levels of stratification (beyond just total/not-provably-total), and I want Rust-style linearity, not as a specialised capability but as something integrated into the regular type system. But even if Idris is the limit, it's still a big step ahead of Java, and it's going to take the industry a while to make that step.
I think you have fallen in love with a specific formalism (I've fallen in love with others) and can only see progress within it. If you take a step back, ignore formalism altogether and ask yourself -- or better yet, conduct research -- on what are the actual problems in programming, you may find solutions in completely different directions.
> it's still a big step ahead of Java
I don't think so, and I don't know why you'd think that. I'm not even sure that Java isn't the limit (modulo small improvements). Of course, when I say something is a big step ahead I don't mean that I enjoy it more or that it can pull some more technical tricks, but rather that it reduces development costs by 30%. If Idris is a big step ahead of Java, by how much do you think costs at, say, Google or Amazon or Citibank would drop if they decided they would all switch to Idris tomorrow? I'm fairly certain that it wouldn't be anywhere near 30%, and I'm not even sure costs wouldn't actually rise.
> it feels like most of the industry is only just adopting the things that OCaml offered, and not yet familiar enough with them to appreciate what Haskell offers above that.
I see it not as progress but as change in fashion. Tastes change, and so people adopt some ideas. Have productivity levels risen in the past three decades? Significantly less so than even Brooks's prediction, which was considered overly pessimistic.
It's hard to know what I don't know. I try to look at what people are doing/liking. But I feel confident that programming is inherently going to be about expressing the essence of the problem you're solving - a program will necessarily be as complex as the problem it's solving - and I think that limit is within reach of the approach I'm advocating, which is a conservative extrapolation of things we mostly already value. So I don't see the need to do anything more radical.
> If Idris is a big step ahead of Java, by how much do you think costs at, say, Google or Amazon or Citibank would drop if they decided they would all switch to Idris tomorrow? I'm fairly certain that it wouldn't be anywhere near 30%, and I'm not even sure costs wouldn't actually rise.
A lot of the value of better programming would be new opportunities, not just cutting costs. I think the industry switching to Idris-level languages will bring a factor of 10 to 100 improvement in productivity - i.e. I expect we'll see companies of ~100 people putting the Google/Amazon/Citibanks out of business, and companies of less than 10 people competing heavily with the giants. That's the sort of thing technological improvements enable - compare the rate at which new videogames are being produced now, and how small the teams producing them are, or look at WhatsApp being treated as a serious competitive threat to Facebook with a staff of 10. I think there's been a lot of productivity gain recently that isn't fully captured in economic measures, because we've (rightly) become a lot more quality-of-life focused culturally, and so our productivity has gone into quality-of-life improvements that aren't so visible in GDP terms.
But I fully expect the transition to take decades, with good reason - I agree that a large company switching tomorrow wouldn't reap much benefit from Idris (though I do think we're at the stage where a technologically-aggressive startup should consider it).
I do, because the things that you mostly value (I guess I value them a lot less) seem to me incredibly complex for the good that they do, and we already have some alternatives that seem both more powerful and vastly simpler. But I am skeptical of them, too, though :)
> I think the industry switching to Idris-level languages will bring a factor of 10 to 100 improvement in productivity - i.e. I expect we'll see companies of ~100 people putting the Google/Amazon/Citibanks out of business, and companies of less than 10 people competing heavily with the giants.
That is a very powerful statement, and I'd like to know why you'd think that. I would even like to know what makes you think that any language (never mind which) can make such an impact, even though we've seen nothing of this sort. I do not share your belief that Idris could make such an impact or anywhere near it, nor probably any other language.
I am skeptical, because having looked at Idris and played with it a bit, I was completely underwhelmed. Java + JML can do pretty much the same, if not better, and for a lowe cost, and this is not a statement about Java+JML being a leap forward. I have no idea if Eve could make an impact as big as the one you predict, but at least I don't feel I can instinctively reject this possibility as easily as I could Idris, as Eve actually brings a lot of new things to the table, where Idris feels like some ergonomic improvement over things that may have had some potential that never materialized, in the hope that better ergonomics is what should do it.
I would like to point out one more problem with viewing the abundance of research as a proxy for utility or of an "established idea", one that touches on my own personal biases :) When you publish a paper, certainly in good journals/conferences, you must exhibit some significant novelty. This, ironically, rewards the more complicated (and so arguably less useful in the real world) frameworks and punishes the simpler ones. For example, I've seen papers talking about embedding separation logic, or specifying amortized worst-case complexity in Coq. You won't see such papers on TLA, because doing both in TLA is rather trivial. Lamport himself complained that he had trouble publishing a paper on specifying and verifying realtime systems in TLA because it is so straightforward that it was hard to demonstrate mathematical novelty. He had to artificially add an extra novelty (the use of a model checker) to make the paper worthy of publication [1].
[1]: http://lamport.azurewebsites.net/pubs/pubs.html#real-simple
I think we have; I think we've seen dramatically accelerating programmer productivity. I think language improvements compound heavily: if a language is slightly more efficient in the small, that means code has to be broken up less, which give it more and more of an efficiency advantage as the codebase gets larger.
> I am skeptical, because having looked at Idris and played with it a bit, I was completely underwhelmed. Java + JML can do pretty much the same, if not better, and for a lowe cost, and this is not a statement about Java+JML being a leap forward. I have no idea if Eve could make an impact as big as the one you predict, but at least I don't feel I can instinctively reject this possibility as easily as I could Idris, as Eve actually brings a lot of new things to the table, where Idris feels like some ergonomic improvement over things that may have had some potential that never materialized, in the hope that better ergonomics is what should do it.
All of programming language design could be called ergonomic improvements. I always found all the pre/post contract stuff too complicated and too different from normal programming to get any value out of; being able to capture more in plain old functions, values and types is where I see value.
Really? I think that the change that has contributed to the lion's share of that boost has nothing to do with language features, and has everything to do with the availability of open-source libraries. The other major contributions have been the widespread adoption of automated unit tests -- also not a language feature, and the practicality of GCs -- if considered a language feature, it's a transparent one.
When Brooks said, around '85, I think, that we won't see a single improvement contributing to a 10x improvement within a single decade, people said he was pessimistic. After 30 years I don't think we've seen a 10x improvement with all methods combined, and if we have, most of it is due to libraries.
Do you think that writing, say, an ERP, a power-station control and management software or an air-traffic control system from scratch (without off-the-shelf/open-source libraries) today would be 10x less costly than it was using C++ in 1987, or even 3x less costly than using Java in 2002 (I'm taking 30 and 15 years respectively as milestones)? Remeber that 10x within a decade was seen as pessimistic.
Although I was just a novice programmer in 1987, I think that the answer to both is an emphatic no. Even when it comes to small/simple programs (which are a completely separate domain), I'm pretty sure that aside from the availability of libraries, Python doesn't have even a 3x factor over VB/Delphi/other "RAD" languages, as they were called way back when.
I'd want to amend that to no domain-specific libraries - I think part of the language advantage is making very general libraries possible. But yes, I do think writing such a system to the same standard would be better than 3x cheaper than in 2002. (I don't think that's what would be done - our implicit expectations around software usability are very different today to what they were in 2002. But I do think we'd get more than 3x better value for the same money than in 2002, certainly risk-adjusted - I think in 2002 there would have been a substantial chance that the project would fail outright)
For smaller programs I'd agree that we haven't improved that much, although it's very difficult to talk about how hard it would be to write small programs without libraries because small programs are mostly just using libraries.
How much would something like Idris lower costs at Experian? That is, in your "cost" metric, what weight do you assign to the cost of catastrophic bugs?
(Yes, I know, the Experian bug was in a third-party component that wasn't updated when it should have been. I assert that it's still a relevant question...)
Also note: I am not actually claiming that Idris would prevent catastrophic security bugs. But the question adds another axis where Idris could give you a 30% cost reduction.
As a practitioner (and evangelist) of formal methods, you won't hear anything from me against the use of formal methods, although I have plenty of reason to believe that Idris (as an example of a particular design) is not very good at formal verification. The cost of bugs must be factored in, but we haven't seen the Idris approach effective at reducing bugs at a cost commensurate with their impact, except when their impact can be truly catastrophic. All the effective formal methods I know of look nothing like Idris.
As someone who has used formal methods to good effect (I believe) and is now learning Lean (which is can be put in roughly the same camp as Idris), I find it very interesting, but I cannot even imagine how it can form the basis for something that will one day be used at scale, especially given that there are such better alternatives.
For programming interactive systems, a very interesting approach which has been quite successful in realtime software and hardware systems, is synchronous programming[1]. It has a deep mathematical theory based on temporal logic, and has been shown successful both in terms of being amenable to particularly powerful forms of formal verification[2], as well as being useful for socially managing the development process[3]. The approach has mostly remained in the realm of safety-critical realtime, but has started making its way to "ordinary" software in languages such as Céu[4] and Eve[5]. Eve, in particular, incorporates at least as much cutting-edge PL research as Idris, but is very different from it, being based on a new and interesting language semantics called Dedalus[6].
In synchronous programming, side effects are not an issue, as the mathematical framework cleanly treats them the same as computation (not requiring embedding in monads), and so allows easy specification over both computation and effects, and a particularly easy way to express global correctness properties (e.g. [7]). The use of monads/algebraic effects in pure-FP is not substantially different from classical imperative programming, other than restricting the effect type, which doesn't seem to me to yield any substantial benefits in general. Using what is perhaps a very bad analogy, given the task of counting the number of passengers on a moving train, classical imperative programming would correspond to a person standing outside the train with a pair of binoculars, pure-FP would be like someone filming the train and then studying the film, and SP would be like someone standing inside the train. Rather than controlling what effects can be emitted by a piece of code, it controls at which point the passage of time can be observed.
In terms of formal methods, TLA+ [8] (a personal favorite of mine) has proven to be very effective, both in theory as well as industry practice -- perhaps more so than any other formal tool -- especially when it comes to specifying and verifying global correctness properties, such as "the database is always consistent", or "data is never lost, even if messages are dropped and nodes fail". It, too, is based on temporal logic, which is a theory that is much better suited, IMO, to interactive/concurrent/distributed software than the theory pure-FP, which is more harmonious with sequential, batch software like compilers. The difference in the difficulty of use of TLA+ vs a tool like Lean [13] for the purpose of software specification and verification (as opposed to doing formal "high math") is the difference between riding a bicycle and flying a Boeing 747.
As far as code-level formal methods are concerned, I find contract systems (like JML [9] for Java, and ACSL [10] for C/C++) to be more promising, again, in my personal opinion, than dependent types, because they separate specification from verification. This is important because while specification is always helpful, the chosen form of verification can make the whole difference between something that's more effective than testing and something that's completely infeasible. The most rigorous form of verification is formal proof, which is both the level of confidence least required by software and more or less the only form of verification afforded by dependent types, at least today. In contrast, contract systems allow you to verify your specification at a level of rigor of your choosing -- even on a per-contract basis -- be it formal proof, model checking or static analysis, concolic testing [11], random test generation [12], or even just inspection.
[1]: https://en.wikipedia.org/wiki/Synchronous_programming_langua...
[2]: https://en.wikipedia.org/wiki/Esterel
[3]: http://www.wisdom.weizmann.ac.il/~harel/papers/Statecharts.H...
[5]: http://witheve.com/philosophy/
[6]: https://www2.eecs.berkeley.edu/Pubs/TechRpts/2009/EECS-2009-...
[7]: http://witheve.com/philosophy/#correct
[8]: https://lamport.azurewebsites.net/tla/tla.html
[10]: https://frama-c.com/acsl.html
[11]: https://en.wikipedia.org/wiki/Concolic_testing
Most of my programming is in the domain of "take some data, do stuff to it, produce an output" and my experience with Idris so far is very positive -- more so than Haskell (which I admittedly gave up on early on), the language does feel ergonomic, and like the type system is "guiding me" rather than getting in the way. I also find its approach to metaprogramming interesting. Whether this translates to Real World™ productivity / maintainability gains vs a mainstream language, it's too early for me to compare.
It does not prevent the browser from eating all the resources.
What I really want is a type system for reasoning about distributed, stateful, programs. Has to deal with mixing strongly consistent eventually consistent (CRDT) data with synchronisation points (see Bloom/Lasp), hot code reloads, migrations, messaging across nodes, possibly some sort of row polymorphism combined with clojure's namespaced symbols... Like sort-of like Cloud Haskell but with zero-downtime deployments. Or like Pony but with a better distribution story.
https://www.doc.ic.ac.uk/~scd/Disributed_Pony_Sept_13.pdf [pres], https://www.ponylang.org/media/papers/a_string_of_ponies.pdf [paper]
I'm not sold on Cloud Haskell or Clojure.
"One True Pairing"?
"Online Transaction Processing"?
probably not "On The Phone".
Interesting. I thought algebraic effects + handlers were the new hotness in modeling effects with types.
This is a good companion piece for the recent Rich Hickey Clojure/Conj talk :)
Languages that have forgone static typing aren't necessarily trying to solve everything using language types at runtime. See eg Clojure's approach. Or SQL's.
But at the base popularity-contest level the answer is clearly yes: Dynamic languages have been on a roll for the last 30 years - Perl, Python, Ruby, Erlang, Clojure, even JS.
Good dynamic systems have long done live-coding against the running system with instant feedback. Even recent static languages like Rust and Scala have notoriously long compile times and thus slow feedback loops.
Also there's along tradition of querying the system for things like auto-completion / doc display.
The cognitive load of fancy type systems like Haskell's is seen as very high.
Beyond what might be purely PL research, there are some other neat ideas being explored. It does function currying, lazy evaluation, grapheme-level Unicode support, and custom operators from any Unicode grapheme.
It's definitely a language I'd be watching if I was writing about near-future languages.
For programming languages, popularity is crucial. Making them easy to implement, having the compilers as widely available as possible, having the most number of developers pick up the language, etc. That's effectively the thesis of the famous "Worse is Better" essay [1].
As trivial as a name may seem, you're really shooting yourself in the foot if it's something that's awkward to say out loud in a lot of contexts. Workarounds, like pronouncing it differently (as suggested elsehwere in this thread) aren't good solutions, as they hamper communication. Unless the language is already very popular, how do you ensure that everyone else is familiar with the non-standard pronounciation?
There are countless startups whose names in English mean something else entirely in another language. For example "Cocu Social", a cooking startup, whose name in French would mean "gregarious cuckold".
But they definitely knew that this pun was there.
Most French people seem to know English quite well, so I would expect they'd be aware of it. Not saying I know this is true for sure, though.
Let’s also mention San Francisco’s Coit Tower, which always amuses the french tourists (coit = coitus)
Actually, this varies with dialect. Route can be pronounced either way: |ˈrüt| or |ˈrau̇t|
https://www.merriam-webster.com/dictionary/route
https://www4.uwm.edu/FLL/linguistics/dialect/staticmaps/q_26...
Btw, why is it "pronUnciation", but "pronOUced". English is hard.
In Australian English, it is "pronOUNCiation"
BTW, in case there has been some confusion here, since the topic was how things were said, I have been assuming this is what the person I was replying to was talking about, not spelling.
Does he let you use git? (Git is a mild insult in the UK)
3real5me
https://www.haskell.org/onlinereport/haskell2010/haskellli2....
How is it an 'industrial' language? Haskell is the main 'academic' language for a branch of Functional Programming.
'industrial' is just another weasel word.
[0] (Unlike applying it to the frontend, which is a comparatively less mature situation with multiple proposed solutions.)