Hacker News Reader

Comment by avinium | Hacker News Reader

Salient line:

While Storm's Clojure implementation served it well for many years, it was often cited as a barrier for entry to new contributors. Storm's codebase is now more accessible to developers who don't want to learn Clojure in order to contribute.

Very common story that I hear from a lot of shops that tried to go Clojure first. Great language, but too little penetration amongst developers. Makes it very difficult to effectively recruit employees/contributors.

At some point, you just accept that it's more important to get a new warm body than to continue pursuing some idealized programming perfection.

View on HN

At another point you just accept you'll have to hire people who don't fit the classic job requirement of "X years experience in Lang" and hire people who don't know Lang at all but want to learn. A growing Clojure shop in Seattle seems to be having success with this, onboarding the language is a small fraction of overall onboarding.

On the more meta level it's weird what people expect devs to learn on the job or know up front. Companies expect employees to keep up with the latest nonsense in web dev, but can't pick up a new straightforward language? Unless maybe another Algol like Go or TypeScript? This Steve Yegge quote captures some of it:

"For some reason, programmers love to learn new stuff, as long as it's not syntax."

> A growing Clojure shop in Seattle seems to be having success with this, onboarding the language is a small fraction of overall onboarding.

I can attest to this fact, as part of a unicorn upstart with dev offices in Sao Paulo and Berlin. Clojure onboarding has never been a major roadblock for new engineers. And we never hire asking for "X years of lang experience". Almost everyone in the (~300 strong) engineering team started with zero-to-little Clojure background.

What do you look for? I mean the "X years of experience" is a shorthand proxy for "can get stuff done in the language we use". So I imagine you must have some other way of getting an idea of how a candidate can get things done. Would love to hear more about it.

What we have found is that it is enough to verify 'can get stuff done', and leave out 'in the language we use'. So, much of the interview turns out to be a process of validating two things: 1. can do what is claimed in the language of choice. 2. gauge enthusiasm and interest to pick up our language.

FWIW, that's been our experience as well (both at a largeish publicly traded company and now at a small private consultancy). Go get work sample tests :)

(It's fine if your work sample test really just focuses on one language! But then make it part of the rubric if that's the answer you want, and tell the applicant about that rubric.)

Thanks!

I can see 1. being pretty easy to suss out (paid take home coding assignment is how I'd do it).

How do you assess 2.?

Mostly in our conversations, in an informal way. I would say the assumption we make is that developers with a penchant for any form of functional prog and/or Lisps, and with an acute focus on testing have an easy time grasping Clojure. And we bias our rubric towards this.

Also, we encourage our candidates to choose Clojure for their take-home assignment, even if they do not have any prior taste of it. This, for those of them who choose that option, gives an idea of if they would enjoy working in it. Our rubric however does not have a bias against developers who do not choose Clojure for their assignments, and we make this explicit to the candidates as well.

I guess the difference here is if they have to learn Clojure just to contribute to a single project, then they may be put behind some other things they want to learn and contribute to first.

Yes it seems strange to me that a Java developer is willing to learn Spring, a massive ad-hoc framework which exists in part because of Java's weaknesses, rather than functional programming and Clojure (a lifelong skill that will benefit their Java programming).

I've worked with plenty of devs in my time who don't really understand what they're doing and just copy and paste code from Stack Overflow until it seems to work. A lot of the enterprise takes the monkeys with typewriters approach to software development.

I think there are cultural reasons for that because enterprise treats devs like fungible cogs that you replace when they burn out. It tends to be highly dehumanizing environment where devs have no say, no autonomy, and no ownership. This naturally pushes out people who have options, and you end up with a pool of devs who just want any job that pays the bills.

>>I think there are cultural reasons for that because enterprise treats devs like fungible cogs that you replace when they burn out.

Its mostly approach to any work in general. If it were upto them, these are really the kind of the people who would dig earth with spoons and shovels, instead of heavy machinery.

The idea is simple, should you use Spring + Java, you are likely to need say 60+ people to run a product well. It could take 15 to make it work in Clojure. But having 60 people has its own advantages from their perspective. Say 10 people decide to leave, they can hire 10 replacements for the lowest prices from the market, train them on the application and get it going. Eventually the whole team could leave and get replaced this way, reducing the problem to a bit like the Ship of Theseus paradox. But it works for them.

If they use Clojure, they have to treat people well, pay them well to retain them. Because there teams tend to be small and a exodus puts everything at risk.

They have to use spoons and shovels, personnel trained to work on heavy machinery are expensive, training is expensive and replacing them is expensive.

Also, to get there, you're already accepting the assertion that the move isn't a management decision :-)

Spring is so large that a lot of job offerings require X years of Spring experience.

Benefit their programming maybe but I don't see who Clojure will benefit anyone's Java as the two languages are poles apart.

They both share the underlying JVM library ecosystem, and as Java increasingly applies functional programming to its mainstream frameworks, clojure is a good illustration of how a near-pure functional style can work within the JVM environment

I think it's interesting that people (willingly?) pigeonhole themselves to the extent that they're a "Java developer" or a ".NET Engineer" rather than more simply a "Software Engineer".

I'm a Software Engineer, and while I have my go-to tools, I'd like to think I make an informed choice to use what fits the task at hand .

An Engineer who's a comfortable polygot is perhaps also a better Engineer overall.

I'm willingly pigeonholing myself as a "software developer mostly working on the JVM" just to be honest about my (current) main strengths/experince. Throwing me into a Go/Python/whatever project would first need some time for me to get up to speed with the language.

I do the same, but I also happen to acknowledge that I slowly transition into X Developer, depending on which project I am on.

It is the usual "Jack of all trades, master of none." consulting, but it does come in handy when we need to deal with mixed stacks end-to-end.

This is not the answer that HN wants to hear, but Spring has enormous market share and makes your resume more marketable for tens of thousands of jobs. Clojure, or any other niche language you might insert here, does not help with career building on that level.

For every one shop that looks at darling niche stuff and thinks, "Ooh, that probably makes him or her a better programmer", there's ten shops who look at it and think, "They're probably going to be insufferable to work with, undermine the architects at every turn, and spend their whole time complaining about the tools we use".

As an aside, if you see Spring as existing to correct issues with the underlying Java language, then I'm highly skeptical that you've ever had meaningful experience with either one. Spring is a portfolio of libraries, stretching across over a dozen business application development domains, with inversion of control as the only common thread. NO language has all that baked into its standard library, and you'd have to pull down a bajillion NPM packages (or whatever) to even come close to approximating it.

> if you see Spring as existing to correct issues with the underlying Java language, then I'm highly skeptical that you've ever had meaningful experience with either one.

Let's please avoid personal attacks if at all possible. Spring started out with the dependency injection framework and "aspect-oriented programming". It was marketed as an alternative to EJB. Web frameworks and other add-ons came later. If you don't see DI and AspectJ as working around Java's weaknesses, then I encourage you to try a more expressive language.

In an expressive language (e.g. Clojure or Haskell), DI and AspectJ could be, for example, replaced by reifying a program as data and using modular interpreters. One interpreter using mock data, another talks to the database, uses transaction boundaries and logs. This sort of solution is unfortunately just not very practical in Java.

DI is just another name for "passing values to functions". We did it in Java just fine before Spring became so popular and Spring doesn't add that much for this particular way of writing Java code. Of course Spring is nice as a whole, even if huge and much more than DI, but the benefits of using it just to specify lazy proxy singletons using annotations are overblown. Autowiring becomes seductive so you get situations when you still need to test and some devs now have a component class with 10 private autowired dependencies and ask questions on how to test it that lead to in order of awfulness using reflection in the test to bypass 'private' and assign dummy values to the dependencies, using Spring's JUnit test runner and mock beans, rewriting the class a bit to use constructor or setter-method autowires that you can set without bringing in all the spring machinery, or break up the test so it depends on less and individual parts may not even need to be Spring components at all.

In another expressive language like Common Lisp, it has generic functions which 1) let you dispatch on more than just the class type of the first argument (implicitly in Java, explicitly 'self' in Python) which is a feature that by itself removes a lot of Java back-bending and 2) you can specify :before and :after and :around methods to provide the benefits of aspects with the built-in function compute-applicable-methods available to help see everything that might apply when invoking a method with specific arguments. There's also metaclasses which open other doors. There's just a whole lot of built-in expressiveness in the languages beyond the Java and JavaScript ecosystems.

Software developers' skills are mostly fungible– supply meets demand.

I've been happily employed (and regularly solicited) as a Clojure developer for over 8 years.

There's no shortage of Clojure dev positions for developers that want to write Clojure.

I think it's a vicious circle: management used year of experienc with <tool/language> as a hiring filter, so devs focus on building years of experience with in-demand tools/languages to maintain maximum career opportunity, and the laser-like focus developers exhibit reinforces the idea that devs can't be flexible on tools and languages, supporting the use of the classic hiring filter.

Well, this isn't a hiring situation, this is a getting contributors to an open-source project. When you're looking for volunteers, the tolerable amount of friction is often less than if you're paying someone for their time.

Yes, I was responding mostly to the grandparent's move towards discussing this sort of tradeoff in the commercial space, not just in the open source space. To stick with open source, it could become the case that moving away from Clojure leads to a cooling of contributors, existing or new. If you're looking for an Apache project to start contributing to, there's no shortage of Java projects. But if you like Clojure or want to learn it better, hey look there used to be this Clojure one, cool. If you liked the Storm project specifically and wanted to conribute, I don't think the language would get in the way that much. (Speaking personally if I had more time and energy I'd like to try and contribute to the Godot project despite my desire to never have to program in C++ again. It's a sufficiently cool project to overcome that.)

Optimizing for new volunteer contributors goes beyond core language choice anyway. For an interesting description of how the game Battle for Wesnoth tried to do it, see: http://aosabook.org/en/wesnoth.html

My team has been using Clojure for close to a decade, and we found the opposite to be the case. While the pool of applicants is smaller, so is the noise ratio. Clojure being niche means that you get people who are willing to look outside the mainstream, and are typically genuinely interested in programming.

We also rarely hire devs who already know Clojure, and we typically just train on the job. It's really not hard to do when you already have people who know the language on the team. In fact, if somebody can't learn Clojure, I would question if they can learn to work on a large project in general. Dev practices, architecture, code styles, tooling, and so on tends to change quite a bit from company to company. The language is only small part of that.

In case of Storm, Apache commons is run by Java devs who have zero interest in learning Clojure. So, it's not surprising they would rewrite Storm in their preferred language.

Same (sure; anecdotal) experience here. Hiring for Clojure has been so much easier for us than hiring for JavaScript.

Granted, these are two different products, and maybe the Clojure one is just more attractive to the right kind of person. Nevetheless, in terms of quality vs. quantity of applicants, Clojure has been massively winning.

This is salient too though:

The new Java-based implementation has improved performance significantly, and made Storm's internal APIs more maintainable and extensible.

It's not totally clear if this is because it's been rewritten Java, but the intrinsic qualities of the language do matter; there are real tradeoffs between maintainability and dynamism/flexibility. I feel like this sometimes gets shortchanged in these narratives.

The announcement says they have a new architecture. At the very least it's not a direct performance comparison to the same architecture implemented in Clojure.

> Storm 2.0.0 introduces a new core featuring a leaner threading model, a blazing fast messaging subsystem and a lightweight back pressure model. It is designed to push boundaries on throughput, latency and energy consumption while maintaining backward compatibility. The design was motivated by the observation that existing hardware remains capable of much more than what the best streaming engines can deliver. Storm 2.0 is the first streaming engine to break the 1 microsecond latency barrier.

Idiomatic java is like idiomatic C: pretty fast. Most smaller languages prioritize other things. Clojure has immutability and Haskell has purity.

I have noticed this with every "X produces code faster than C": you begin with two programs that use a suboptimal algorithm, then you take your non-c language and try to write C in it. The result is always awful and removes most reasons not to use C in the first place.

This has somewhat changed with rust and in some sense C++,but for other languages my point still stands. They have a nice idiomatic golden path that is fast enough for most cases. Once you need performance badly enough you have to treat you language as an assembler, and then you will always lose to languages that actually are good at that.

I say this as a scheme/Haskell weenie. Writing really performance scheme and Haskell code often means writing ugly code.

Kotlin generally is not slower than Java. One exception is that it inserts a lot of null checks. And there are features that will make it even faster than Java, I'm talking about inline lambdas.

The compiler is significantly slower than Java's, though.

That's the sad truth for a lot of modern compiled languages (I think Go is an exception). While I did not dig into their compiler internals, it just seems an inevitable consequence from a powerful language (Go is pretty simple in that regard). C++, Rust, Scala, Kotlin, Swift, they all have significantly longer compilation times compared to their predecessors. Probably that's the price we have to pay.

Not really.

.NET Native, D, Delphi, Ada and Eiffel also have pretty fast compilers, while offering quite powerful languages.

Lets even pick Turbo Pascal 7.0 for MS-DOS, which was quite feature rich, and was able to compile very large programs in a couple of seconds.

Also Visual C++ with modules preview support, incremental compilation and linking is also reasonably faster than the FOSS alternatives.

It is only a matter how much money one is willing to invest into improving tooling support.

As far as I know Google is supporting JetBrains in making the Kotlin compiler faster.

Chez compiles a 80Kloc project of mine in less than a second, and apparently that is still 2x slower than before they switched to a nanopass compiler.

There is no reason compiling things should not compile at similar speeds to chez unless you are telling your compiler to ootimise everything.

C++/ruae are just exceptionally slow and it became the new bottom line somehow. I know rust does a lot of housekeeping, but after using chez and sbcl, 45s for a 1.5kloc project is killing me inside.

I say this as a scheme/Haskell weenie. Writing really performance scheme and Haskell code often means writing ugly code.

The same with Clojure, suddenly you start writing Java with parenthesis.

In the JVM, if you want absolute control over performance its better to just write Java for those parts (don't know much about Kotlin).

Clojure is marketed as a very fast language (bunch of clever optimisation gets done because data structures are immutable). If the gains were from switching to Java they'd probably have said so less ambiguously.

Surely what they mean is they've rewritten it to be faster, and switched to Java at the same time because they want to open up the contributor pool.

BTW, if anyone's looking for another language that stresses immutability on the JVM, you might want to check out Frege, a Haskell-like language with some "tweaks" that make it work well in that environment. https://github.com/Frege/frege

And as a sibling comment mentions, the announcement does clearly state that a rather extensive rewrite was done, including switching to a new architecture, with likely performance improvements coming from that - it's nowhere close to a pure Java vs. Clojure comparison.

My impression is that Eta is a better choice, if only because it is compatible with a wider variety of Haskell libraries.

https://eta-lang.org/

Eta is a better choice, if you care about Haskell compatibility in the strictest sense. Frege doesn't really pursue this, but the flip side is that it optimizes for working as smoothly as possible with the JVM ecosystem.

I did find that bit interesting. It seems it's missing a leading "due to the commiters' better familiarity with Java."

Since you can argue knowledge of a language does help with maintenance and extensibility. With this, it also makes more sense for the following bit about finding people able to contribute. No contributors is pretty bad for maintenance.

At my work, we use Clojure, but I do often wonder what would happen if the current Clojure devs left. New devs unfamiliar with Lisps, FP, and Clojure would most likely have a hard time maintaining the code base and extending it, and if they try to rush it before really learning the language, would quickly degrade the quality of the code base, compounding the effect.

I think as long as you continue to have one strong Clojure dev on the projects, you'll be fine, as they can direct newcomers and help them transition to Clojure, but if you lost that, I think a rewrite in a different language would make sense, or there's a risk of the code base degrading quite quickly.

It's a sad state of reality, because it drives down the quality of software to the common denominator of fresh junior devs, which is not much.

Another view of this would be: progress in this industry is made by those who succeed while sticking to their guns and not optimizing their infrastructure for the available worker pool.

So, you’re saying the quality of the code dropped with the move to java?

Could there be other reasons? Maybe they rushed?

He is not saying Java caused it, just that the priority was cheaper labor, and a more popular language makes that more likely.

This is in the context of software development in general, not this specific project per se.

> It's a sad state of reality, because it drives down the quality of software to the common denominator of fresh junior devs, which is not much.

There's no reality here, just myth. That switching to Java would drive quality up makes at least as much sense as it driving it down.

If you're referring to the particular languages, then the reality is that no big effect has been found for the choice of programming language on software quality. As lack of evidence in favor of a big effect is evidence of its lack, the most probable explanation is that no such effect exists (it's possible a small effect exists, but we don't know in which language's favor). And, indeed, while there is a theory explaining (and even predicting ahead of time) why there is no big effect to the choice of the programming language, not only is there no empirical evidence suggesting there is such an effect, there is no theory to predict it, either. That language has a big effect on quality is, at this point, no more than wishful thinking among a minority of developers who are big programming language fans. It's a bedtime story. (I am not saying that language didn't ever have a big effect or that it never will, just that both theory and observation show that there isn't one currently among "reasonable" languages in common use; also, I have used and I like both Java and Clojure)

And if you're only referring to the priority of making the project more accessible, then I don't understand your comment at all. There are many, many more experienced Java developers than experienced Clojure developers, and experienced developers are at least as likely not to bother learning a new language in order to contribute to a project (although perhaps for different reasons). Picking a popular language makes the project more accessible to experienced developers.

What has been shown to drive quality up or down is process. If you have a good process, then making the project more accessible can help, and if you don't then you're screwed anyway.

If we're going to talk about empirical effects, we first need to agree on what quality means. Parts of the thread, including your comment, lead me to believe we sometimes mean an abstract, desirable but ineffable property--but then we're not speaking empirically.

Do you mean "defect rate" or something else?

I mean any reasonable metric you pick. The studies that have tried and failed to find a big effect used defect rate as a metric, but studies are not the only way to detect big effects. Because there are strong selective pressures in industry that can detect big effects, we can say that no big effect on the bottom line has been found.

> then the reality is that no big effect has been found for the choice of programming language on software quality

Substantiate that. I say it's garbage.

<https://www.cse.iitk.ac.in/users/karkare/courses/2010/cs653/..., and this BTW is from 1994 so this evidence has been around for ages

  Language LOC Documentation Lines Dev time (hours)
  (1)Haskell 85  465  10
  (2)Ada 767  714  23
  (3)Ada9X 800  200  28
  (4)C++ 1105  130  –
  (5)Awk/Nawk 250  150  –
  (6)Rapide 157  0  54
  (7)Griffin 251  0  34
  (8)Proteus 293  79  26
  (9)RelationalLisp 274  12  3
 (10)Haskell 156  112  8

(Edit: all aligned with tabs but looks like they've been stripped. See paper instead)

comments of note:

(2) The Ada solution was written by a lead programmer at NSWC; in this sense it represents the “control” group. The developer initially reported a line count of only 249; this was the numberof lines of imperativestatements as reported by theSun Adacompiler, anddid notinclude declarations, all of which were essential for proper execution. The line count of 767 is based on the actual code in and does not include lines with only termination characters.

(4) TheC++solutionwaswrittenbyanONRprogrammanagerafterhavingfirstwrittentheAwk solution described in the next paragraph. In addition to these 1105 lines of code, the developer also wrote a 595-line “test harness.” No development times were reported [Note: C++ has evolved a lot since then, but perhaps so has haskell]

(10) Intermetrics, independently and without the knowledge of NSWC or Yale University, con-ductedan experimentof its own: theHaskell Report was given to anewly hiredcollege graduate, who was then given 8 days to learn Haskell. This new hire received no formal training, but was allowed to ask an experienced Haskell programmer questions as issues came up during the self study. After this training period, the new hire was handed the geo-server specification and asked to write aprototype in Haskell. Theresulting metrics shown in row 10 ofthe tableare perhaps the most stunning of the lot, suggesting the ease with which Haskell may be learned and effectively utilized.

Also mentioned was "The use of higher-order functions [in haskell] is noteworthy" so language features did help. Actual evidence.

> As lack of evidence in favor of a big effect is evidence of its lack

yeah right. Let me rephrase that for you - Lack of knowledge by an HN poster is not evidence of lack of evidence.

> What has been shown to drive quality up or down is process

Evidence please? My experience is that process can be rigorous but useless crap. My old boss said "process is not sufficient to produce quality, but it is a necessary"

There's plenty more in that PDF, please read it.

> this evidence has been around for ages

Evidence of what? This is not a study of software development at all, but of prototyping using a program with no more than a few hundred lines. I'm talking about effects on software development. The issue of prototyping/tiny programs is a completely separate one.

> Substantiate that

I don't know how I can substantiate the lack of evidence, but here's a recent failed attempt to find evidence: https://arxiv.org/pdf/1901.10220.pdf

> Lack of knowledge by an HN poster is not evidence of lack of evidence.

True. While I have been following the subject for at least the past 20 years (and I have read your PDF several times already, including shortly after it was published), it is certainly possible I have missed something. If there's any evidence of a large effect you believe you've found, let me know.

> Evidence please?

Sure. Here's code review, for example (I don't have time to discuss each paper -- as they're quite different -- but they all paint a similar picture). BTW, while programming language studies are debating effects in the 0-15% range, these report effects in the 30-80%:

* What We Have Learned About Fighting Defects, 2002 -- https://www.cs.umd.edu/~mvz/pub/eworkshop02.pdf

* The Impact of Design and Code Reviews on Software Quality: An Empirical Study Based on PSP Data, 2009 -- https://www.pitt.edu/~ckemerer/PSP_Data.pdf

* Large-Scale Analysis of Modern Code Review Practices and Software Security in Open Source Software, 2017 -- https://www2.eecs.berkeley.edu/Pubs/TechRpts/2017/EECS-2017-...

* What Types of Defects Are Really Discovered in Code Reviews?, 2003 -- https://ieeexplore.ieee.org/document/4604671

* Modern Code Reviews in Open-Source Projects: Which Problems Do They Fix?, 2014 -- https://www.testroots.org/assets/papers/2014_beller_bacchell...

* The impact of code review coverage and code review participation on software quality: a case study of the qt, VTK, and ITK projects, 2014 -- https://dl.acm.org/citation.cfm?id=2597076

* Best Kept Secrets of Peer Code Review, 2013 -- https://smartbear.com/SmartBear/media/pdfs/Best-Kept-Secrets...

* Modern Code Review: A Case Study at Google, 2018 -- https://storage.googleapis.com/pub-tools-public-publication-...

If we summarize our current state of knowledge (not myth) it is this: process matters a lot; language matters little, if at all (with all the caveats I mentioned in other comments, such as diminishing returns etc.).

You're moving goalposts. You denied language mattered:

> then the reality is that no big effect has been found for the choice of programming language on software quality

Then I show some evidence giving small programs rapidly developed, which you then say doesn't count because it's 'tiny' and 'prototyping'. You also don't define quality which gives you lots of wriggle room.

Nope, 1000 line programs aren't tiny. They aren't industry monsters but you can't dismiss them because it contradicts you. It's not proof, but it is strong evidence.

...arxiv paper... Interesting, thanks. It would have been helpful to have posted that in your original post.

However from <https://soarsmu.github.io/papers/A_Large_Scale_Study_of_Mult..., and note this paper is not mentioned, as in not debunked, in the above arxiv paper.

"Bhattacharya et al. study four open-source projects which use C and C++, i.e., Firefox, Blender, VLC Media Player and MySQL to understand the impact of languages on software quality [20]. They compute several statistical measures while controlling for factors, such as developer competence and software process. They find that applications previously written in C are migrating to C++ and C++ code is often of higher quality, less prone to bugs, and easier to maintain than C code."

and from that same paper:

"As can be seen in Tables 8, 9, and 10, the mean de- fect density values for the C sets can be up to an order of magnitude higher than the mean values for the C++ sets."

> Sure. Here's code review...

I was talking about your claim about languages not mattering, I did not mention code review. All these papers are about code review. This is relevant to my code review comment (edit: I meant process comment), they don't say anything about language vs bugs (unless you wish to point out a paper that does).

> process matters a lot

And I agreed with you, I just said it didn't deliver quality, it just prepared the ground for it. With enough process you can close any hole in a language, but it gets exponentially expensive.

> You're moving goalposts.

Only if you insist on uncharitable reading. I'm talking about software development; if I mistakenly assumed that could be left implicit, I'm sorry. And ~500 line programs are positively minuscule. JQuery is ~50KLOC, and an average business system is ~5MLOC. We're talking four orders of magnitude between those prototypes and a rather average industry system size (systems commonly run to tens and even more than 100 MLOC).

> It's not proof, but it is strong evidence.

I don't think it matters if it's strong or weak evidence, as it's not even about software development.

> the C sets can be up to an order of magnitude higher than the mean values for the C++ sets.

As I mentioned in another comment, C is not a good example. It's a ~50-year-old language, and the theory that correctly predicted that languages won't make a big difference was based on diminishing returns. I.e. not that no two languages have ever been or could ever have a big difference, but that over time the ability to affect quality with language would severely diminish. The original prediction of that theory was that no 10x improvement would be made by a single language improvement over a decade, and was called overly pessimistic by PL fans. It's not been over thirty years, and we have doubtfully made a 3x boost with all language features combined.

If you want a more precise statement: no theory or empirical evidence supports the claim that a reasonable choice among production languages developed in the past three decades or so has a big impact on any measurable bottom-line metric.

> I just said it didn't deliver quality

But those papers show that, unlike language, process does have a big impact on quality.

> With enough process you can close any hole in a language, but it gets exponentially expensive.

You are now making an unsubstantiated claim that contradicts a substantiated claim. Programming languages (with the caveats above) have not been found to have a big effect (nor is there a theory that suggests they do), while process has.

> I'm talking about software development

And the value different languages may add is entirely about software development.

> I don't think it matters if it's strong or weak evidence, as it's not even about software development.

Programming languages are not about software development? Oh do go on.

> As I mentioned in another comment, C is not a good example.

A study of C vs C++ strongly indicates your claim about languages is false and suddenly C is not a good example? I'm trying to believe I'm just misunderstanding you but it's getting more difficult.

> The original prediction of that theory was that no 10x improvement would be made by a single language improvement

This isn't what you said. Let me remind you "If you're referring to the particular languages, then the reality is that no big effect has been found for the choice of programming language on software quality"

The goalposts aren't being moved, you've just bought them plane tickets to barbados.

> If you want a more precise statement: no theory or empirical evidence supports the claim that a reasonable choice among production languages developed in the past three decades or so has a big impact on any measurable bottom-line metric

Another claim. Show me the study that says that.

> But those papers show that, unlike language, process does have a big impact on quality.

It can if done properly; it does not do so automatically. You can have code reviews that are of little use because you have no spec to review the code against. I have been in that very position. I don't dispute the value of process but it's a necessary but not sufficient condition to bring about quality. I also agree code reviews are good, if done properly.

ME >> With enough process you can close any hole in a language, but it gets exponentially expensive.

YOU > You are now making an unsubstantiated claim that contradicts a substantiated claim.

If I use C I have to worry about garbage leaks, double-freeing pointers, out-of-bounds accesses etc that can be detected with code reviews. If I use python, I never have any of these problems so a code review need not check for these. That's a lot cheaper.

(Edit: weird stuff happening to my post, may appear twice)

> This isn't what you said.

I can't repeat all the caveats every time I say something. I assume readers believe my comments are at least reasonable, even if they disagree with them. For convenience, I restated the current state of knowledge more precisely in my previous comment. If you're interested in what I have to say, assume I'm not an idiot, and perhaps we could have an interesting discussion, and if you think I'm an idiot, then there's no point in arguing at all.

> Show me the study that says that.

Show me the study that says otherwise. I don't need to provide evidence for the lack of an effect (although I did). The lack of an effect is, at the very least, the default hypothesis in this case (as in many others) as there is no theory suggesting we should believe otherwise (while a theory that explains why languages increasingly have smaller effects has made correct predictions).

> It can if done properly

That's your hypothesis. We've been trying to find evidence of that, or others like it, for a long time -- both in academia and in industry -- without success. We simply do not observe that the choice of language today (among reasonable ones etc.) has a big effect, either in studies or in industry practice. But if you've found a language that can drastically reduce software development costs and/or increase quality at scale, that discovery can be easily translated to billions of dollars. Go ahead and make them.

Peter Norvig made a comment here that agrees with you about process and architecture over language (he was talking about why he started using python instead of lisp)

"In terms of programming-in-the-large, at Google and elsewhere, I think that language choice is not as important as all the other choices: if you have the right overall architecture, the right team of programmers, the right development process that allows for rapid development with continuous improvement, then many languages will work for you; if you don't have those things you're in trouble regardless of your language choice."

https://news.ycombinator.com/item?id=1803815

You could be right. But I think there are good reasons to believe that there is more to it than that.

I have a strong suspicion that people smart enough to do advanced meta programming on their own are not smart enough to collaborate with equally skilled coworkers on code that uses all those abstractions to the max. Even collaborating with your own former self is more difficult that writing new code.

So I think _some_ dumbing down is inevitable. The question is how that can be enforced. Choosing a dumber language to bludgeon everyone into submission is the nuclear option so to speak.

I often think of this Dijkstra quote about abstraction: "The purpose of abstraction is not to be vague, but to create a new semantic level in which one can be absolutely precise".

But I think there's something missing in this thought. Abstraction, especially when its precise, is also a form of compression. To understand what a particular abstract expression means in a concrete case requires decompression, which may require a lot of mental effort.

These are admittedly extremely half-baked thoughts...

Because no large effect of programming language choice has been observed, and because no theory explains why there would even be one[1], speculating on the causes of a non-existent phenomenon is not very helpful.

[1]: Once a certain point has been reached; theory does predict diminishing returns.

> Because no large effect of programming language choice has been observed

Do you mind sharing a source for this? It seems trivial to generate a programming language (Brainf*ck) that performs very poorly.

A more likely statement: all commonly used languages are more or less on par.

I meant the latter. I just can't repeat all caveats every time I bring this subject up :) https://news.ycombinator.com/item?id=20074684

It's near impossible to objectively measure anything when it comes to programming as projects are neither repeatable nor comparable and test subjects are never blank slates.

I have read quite a few studies, and I always come away thinking that the methods used are inadequate to support any conclusions at all.

But does that mean there is no causal effect of programming language choice? I don't know and I'm not making any claims either way.

I was talking about abstraction in one person's mind vs abstraction shared between a group a people. And what I said could actually be the reason why the choice of programming language has little effect (if that is the case)

> But does that mean there is no causal effect of programming language choice? I don't know and I'm not making any claims either way.

I'm also not making definitive claims, but we can't ignore observations, either. While small effects can appear or disappear depending on methodology, big effects are easy to find and hard to hide, especially in an environment with strong selection pressures. Therefore, the most likely explanation to why no big effect has been detected -- either in studies or in industry -- is that there isn't one, and that should at least be enough to make it the working hypothesis. So before explaining effects that have yet to be detected we should establish that they exist at all, something we have so far failed to do.

There may well be a small effect, and there could still be a large effect, although the chance of that is small unless we have a really good explanation to why we haven't found it.

Here are some reasons to doubt this working hypothesis:

1) It is inconsistent with the personal experience of most developers. Few would claim that C is just as productive and safe as some garbage collected language (provided that both are suitable for the task at hand), or that a SQL group by statement is not more productive to write than the equivalent procedural code.

2) The methods used to study the subject are completely unfit for purpose.

3) It is inconsistent with the few results that seem at least somewhat well supported because they are simple enough to measure, such as the constant bug count per line of code.

> It is inconsistent with the personal experience of most developers.

I disagree. I've been programming for about 30 years now, and most reports I've received -- both from programmers and managers -- are that language doesn't matter; certainly not much. Carefully selected personal "feeling" report are worthless (especially in biased forums), as that would produce even more evidence that homeopathy is effective than that programming languages make a difference.

> Few would claim that C is just as productive and safe as some garbage collected language

I agree, but that's where the many caveats come in. First, my argument is more one of diminishing returns (following Brooks's theory). I.e., languages make an increasingly small difference. C may have been a big improvement over assemby, and Java may have been a smaller but still quite big improvement over C, but now differences are quite small.

Indeed, the effect of GC and memory safety vs C was one that was almost immediately detected by industry (once mature, performant, etc.) and triggered a huge shift. We do not see such shifts now. Companies move among mainstream languages, among non-mainstream languages, and between the two classes in a way that does not suggest a big effect at all.

> or that a SQL group by statement is not more productive to write than the equivalent procedural code.

That's a completely different matter. I don't think anyone believes that writing a 1MLOC program in SQL is easier than writing a similar one in Java or Python, even if the Python/Java one is 1MLOC, and the SQL one is 200KLOC. When the program is small, there can be large differences, but that's a completely different problem (and also explained by theory).

> The methods used to study the subject are completely unfit for purpose.

I disagree, but it doesn't matter. Big effects are easy to find and hard to hide. And in an environment with strong selective pressures, they're found even when no studies at all are conducted.

> It is inconsistent with the few results that seem at least somewhat well supported because they are simple enough to measure, such as the constant bug count per line of code.

It is not. First, the difference in LOC is not as big as PL fans claim (it can be big for small programs, not large ones). Second, you may want to review those results.

>I agree, but that's where the many caveats come in. First, my argument is more one of diminishing returns (following Brooks's theory). I.e., languages make an increasingly small difference. C may have been a big improvement over assemby, and Java may have been a smaller but still quite big improvement over C, but now differences are quite small.

I don't necessarily disagree, but I'm a bit skeptical about the way in which you use the concept of diminishing returns. If I'm not mistaken then Brooks used diminishing returns in relation to headcount. And this is in fact how economists use the term as well - adding quantitatively more of one production factor.

But that doesn't apply to qualitative changes like introducing new production methods and tools, which is what programming languages are. New programming languages are not simply "more programming language".

If we agree that there have been observable effects of new programming languages in the past, then we cannot exclude the possibility of the same happening in the future.

>First, the difference in LOC is not as big as PL fans claim

What PL fans claim is a bit of a fluffy benchmark, but there are undoubtedly significant differences in LOC, certainly significant enough to make an economic difference.

>I've been programming for about 30 years now...

Shockingly, me too :)

> If I'm not mistaken then Brooks used diminishing returns in relation to headcount.

Oh, I'm referring to "No Silver Bullet" and diminishing returns in the sense that languages at best help with "accidental complexity", so the less of it there is, the less languages can help.

> New programming languages are not simply "more programming language".

But there are inflexible theoretical limitations on their utility (even without Brooks). We know the effect of different programming constructs on expressive compression and reasoning costs. The closer we get to the limit, the less benefit we can have.

> then we cannot exclude the possibility of the same happening in the future.

I'm mostly pointing out that it's not happening at present. I'm less sanguine about future advances because of various limitations (and, BTW, Brooks's predictions were called overly pessimistic by PL fans at the time and it turned out they were too optimistic), but I won't rule out another major breakthrough, or possibly two.

> but there are undoubtedly significant differences in LOC, certainly significant enough to make an economic difference.

I don't think I agree with your first assertion (although that depends on what we mean by "significant" here), and I certainly disagree with your second. We simply have not been able to detect or induce such an effect. Here, too, there are theoretical results showing that increased expressiveness cannot lead to cheaper reasoning, which is not an obvious result even in the worst case (because there are far fewer "compressible" programs than non compressible ones).

>We know the effect of different programming constructs on expressive compression and reasoning costs. The closer we get to the limit, the less benefit we can have.

>Here, too, there are theoretical results showing that increased expressiveness cannot lead to cheaper reasoning, which is not an obvious result even in the worst case (because there are far fewer "compressible" programs than non compressible ones).

I find that very interesting. Do you have a source for it?

Yes. Some interesting papers are:

* http://www.lsv.fr/Publis/PAPERS/PDF/Sch-aiml02.pdf

* http://www.lsv.fr/Publis/PAPERS/PDF/DLS-jcss-param.pdf

* https://www.cis.upenn.edu/~alur/Zohar03.pdf

What's important to put those results in context for those who are not familiar with the subject is that, in the context of complexity theory, the "model checking problem" does not refer to the complexity of a particular model checker algorithm, but to the inherent complexity of the problem of deciding whether a program M satisfies some property 𝜑 (i.e. whether M is a model of 𝜑 in the formal logic sense). The "model checking problem" is the closest to a mathematical description of what we usually mean when we talk of "reasoning about a program."

I've summarized these results and more here: https://pron.github.io/posts/correctness-and-complexity

Thanks!

Java is abstraction over JVM which is abstraction over machine code. But I don't need to know anything about JVM bytecode or machine code to write typical Java code. Good abstraction does not require decompression. You're not going down to assembly to understand some Java code. You think on Java code level. That is good abstraction. If you need to decompress what you're reading, then it's a leaky abstraction. It's okay to use leaky abstractions to reduce repetitions, for example. But ideally your abstractions should not require one to learn its implementations in typical cases. For example if your function name is good enough, reader could trust you to do the proper thing and not read its implementation.

You just explained cornerstone of good design

Abstractions are a spectrum. There's always a degree of benefit to be had from peeking under the hood. I'd argue that it's rare that you find an abstraction that can be used well without having to understand, or "decompress", some of it.

Consider JVM and Java: that you can write Java seemingly without any knowledge about JVM is an illusion - you already learned a lot about how things like JVM work under the hood when first learning programming[0]. Learning more about JVM itself lets you write more efficient code and debug better.

Consider functions: it's near-impossible to contain all the information necessary to use a non-trivial function in its header[1]. That's why good code contains comments and other forms of documentation describing the abstraction in more details. Even then, it's not always enough - sometimes it's really much easier to understand what you need by reading the source[2].

Consider any appliance - be it a car, or a radio, or a dishwasher. You can use it to it interface, to some extent at least. But knowing what's going on under the literal hood really does help with use, and especially helps when something goes wrong. If you don't know what's hidden under the abstraction layer, any failure will likely be incomprehensible to you and leave you helpless.

[0] - In theory, it might be possible to learn some programming without learning anything about hardware or hardware-emulating abstractions. In reality, I've never seen it or heard of it, and I suspect that the simplest mental model of code execution is isomorphic to somewhat simplified computer or virtual machine.

[1] - Function name + name of its arguments + types of its arguments and return value, if available.

[2] - Then again, implementation code may not capture the entire abstraction either. That's why good code often features comments inside the implementation, explaining the rationale behind some of the less obvious code parts.

>Good abstraction does not require decompression

I think there's a misunderstanding. What I'm calling decompression has nothing to do with digging into the implementation of a particular abstraction. What I mean is merely tracking down all the indirections in the public interface and breaking them down to their concrete meaning in a specific case.

Some languages allow for a lot more moving parts than others. For instance, in Java obj.otherName means that otherName is an instance variable (or theoretically a class variable) and access time is guaranteed to be constant (tbd: caveats). It's not redefinable as it is in C#, Python or Swift. So there's one less thing to track down but also less flexibility.

Or take an extreme example like Go's range loop. It's defined only for a handful of builtin data structures. When you see a range loop, you know what it does without following any further indirections. It can be extremely annoying and create a lot of friction when you have to work with custom data structures. But it can make other people's code easier to read than in other languages.

Or take something like this:

  a == b

This simple expression has a far greater number of possible meanings in a language that supports operator overloading and generics than in a language that doesn't.

So transforming an expression's possible meanings into its concrete meaning is what I call decompression. There may not be a need for looking anything other than public interfaces.

Also, you're talking about abstractions purely from the perspective of users. But we are often creators and maintainers of abstractions as well when we model a particular problem.

I think the better word you're looking for might be "reify" rather than "decompress". But there is something here that I can relate to. A long time ago I also made a comparison between various languages of "what do you have to understand to fully understand a Hello World program?" An interesting followup to that would be how much you have to understand on top of that to understand a more useful program.

You might be interested in this book, Patterns of Software, which brings up the issue that inheritance in OOP isn't about reuse but about compression and how the implications of that are why it's not had that much success in following through on the "reuse" promise. http://dreamsongs.com/Files/PatternsOfSoftware.pdf

>I have a strong suspicion that people smart enough to do advanced meta programming on their own are not smart enough to collaborate with equally skilled coworkers on code that uses all those abstractions to the max. Even collaborating with your own former self is more difficult that writing new code.

I'm an unabashed Clojure fanboy and will talk the language up to anyone who will listen, but one thing that I've been thinking about lately is that good judgment on abstraction use is necessary for using such powerful tools. I don't think the place to tackle it is by being "smart enough to understand," but rather by being "disciplined enough to not abstract sometimes." The exciting thing is that discipline doesn't require you to be super smart, just to think it's worth caring about.

Also half-baked, but I think it's something.

The easy solution to this is to create a culture of using powerful tools, but requiring REALLY good reasons to go far to their edges in terms of metaprogramming.

It's not that much harder to enforce than it is to enforce conventions in less powerful languages, and the convenience of working in a well-designed language is so nice.

That's not the only point though, is it?

> The new Java-based implementation has improved performance significantly, and made Storm's internal APIs more maintainable and extensible.

It’s unclear if the author meant that performance is better “because java” or “because rewrite”.

As a Clojure enthusiast who has never been able to find a FT job using the language, I find it hard to believe that the demand for Clojure devs outstrips the supply.

And even if you have to hire devs who don't already know the language I really don't understand the difficulty in learning Lisp - few languages are simpler or easier to grok.

The demand and supply for Clojure both exist, but they are not equally distributed geographically.

Things are progressing, but the cultural addiction to shipping meat around is still too high in my opinion. Geography is a weird and often arbitrary constraint for knowledge work.

I didn't have any problems finding work in Clojure when I was last looking (Aug-Sept 2018) with offers in four locales across the US. I was not able to find remote work at comparable salary for a US based org. (Some context) I have 7 years experience using Clojure professionally across 3 orgs (16 years overall) and was looking for individual contributor roles.

Some things I observed about the Clojure companies:

- Their HR seemed slower than the non-Clojure companies I was dealing with. I don't think this has anything to do with Clojure, but I am no longer working in Clojure full-time because the most credible offers took too long and I've found non-Clojure work I love.

- Outside of trading firms, I did the most whiteboard coding questions with Clojure firms compared to shops looking to hire for Python or Elixir. I like whiteboard work, so this was fine for me.

- 6 Clojure companies, 4 different conccurrent processing models and 6 different service frameworks. Neither of these is necessarily bad, but the community was much more fractured than I realized.

- Of the 6 Clojure orgs, 4 were actively hiring (had open roles) and 2 were passively hiring whenever someone materialized on IRC/Slack/Twitter. This leads me to think if you want a Clojure job, you should probably reach out to people you know inside Clojure orgs and, barring that, the HR people at these orgs directly. This isn't a Clojure-specific thing, but worth keeping in mind.

- 4 of 6 didn't seem to have dedicates DevOps, leading to them looking for candidates also familiar with deploying, monitoring, and debugging Linux hosts. I think I was given heavier weight at some orgs because of a background doing this type of work for Clojure stacks. Might be an angle worth looking into if you don't have production Clojure but you do have Linux and Java deployment/monitoring/debugging/tuning experience.

> At some point, you just accept that it's more important to get a new warm body than to continue pursuing some idealized programming perfection.

I think that is why Kotlin and Typescript are doing so well generally, they build on top of already established bases offering good new things without throwing out everything else.

The trade off is they are constrained from how far they can diverge from their roots but on the flip side that is a strength as well.

I like both, they feel productive without feeling totally alien.

Same thing I repeat again and again: There is a false cost assigned to learning a language. Developers are too unwilling to even try stepping beyond the boundaries of the first thing they learned. The cost is always lower than they may think, and the benefits far surpassing what they may think. We've got to work at showing developers those benefits early; it's as important to creating software effectively as any other engineer's basic toolkit.

> Developers are too unwilling to even try stepping beyond the boundaries of the first thing they learned. The cost is always lower than they may think, and the benefits far surpassing what they may think.

and yet... almost every time I've come in to a project where developers got past those boundaries.... they inevitably recreated patterns/idioms from tech 1 in to tech 2, and make tech 2 horrifically bad, and made it even harder for anyone who actually knew tech 2 to come in to the project and help.

adding other languages is more than just 'get over it' - without guidance from experienced people who can understand good/efficient ways to model the problem domains unique to a business idiomatically in the new tech, you're very likely adding even more technical debt that you won't even know about for months/years. And... IME, the "let's use new stack X" are moved on to another project or company by the time the technical debt is apparent to everyone.

I suspect programming has something similar to Gear/Plugin Acquisition Syndrome in photography and music. It's easy to let tinkering with new toys stand in for doing the actual work of creating.

At some point, you have to sit down and do the work with the tools you have if you want to get anything done. If I were to get serious about programming, I would use C# or Go because I found them the most accessible after trying lots of languages. There's nothing wrong with finding a niche.

True, but I expect you to know which two or three /distinct/ tools you can work with effectively, and which situations they are most suitable for.

Not only that, but knowing multiple languages has a multiplicative effect, massively expanding how you think of programming in all the languages you know, even the stable standbys like C#.

Most CS programs have a course on functional programming or programming paradigms including functional programming.

I don't understand why it would be a challenge to pick up Clojure.

I don't know how it is elsewhere, but I saw many (most?) students struggle with those classes. These are people who have learned to successfully program, but they struggled quite significantly with functional languages (Haskell in this case). Even when they pass, many of them probably wouldn't want to pick up a functional language afterwards.

But Scheme and Common Lisp (and its dialects) are drastically easier to learn than Haskell.

Furthermore, I think it also has to do with the order in which these courses are taught. When I did a small introduction to FP at my university, a lot of people who had no prior experience in programming picked up Haskell much more easily than those who already had a background in imperative/OO programming.

(Though that strengthens the point - FP is often quite a leap for programmers who have used imperative languages for a long time.)

What learnt in school is insufficient to meet with all real world challenges. To claim as a developer, you need to understand the internals, that is the ability to poke and trace to the root of the problem deep underneath.

Debugging in FP feels different to most people, where our thoughts is tuned to the procedural model with observing states and form our hypothesis from there.

Another disadvantage with niche programming language is not enough Stack Overflow help there. That alone will drive most people away.

Many cs courses teach ML style functional programming, not Lisp or Clojure

People have limited time in their lives. If you want to recruit volunteers to your open source project it has to be in a language they either already know or are interested in learning. It may not be “hard” to learn Clojure, but the potential contributor may consider the time and mental energy spent fumbling around frustrated with all the off-putting sigils and different brackets to be better used learning a language with brighter prospects.

FWIW, I have a very different read on this, based on both my experience with Clojure and some of the leading Storm-using orgs (and the Apache process).

Firstly: Clojure isn't actually a problem. We hired people who are highly junior (one who even refused to call themselves a developer) and taught them Clojure. It was fine. It is true that you need a strong dev lead to guide them through that, but honestly, if you don't have a strong dev lead your Python/Java project is screwed too.

Secondly: look at Storm's increased success at Java data processing sweatshop^H^H^H^H^H^Hconsultancies. Of course "can write Java" is what you're going to get if most of your business revolves around the lowest bidder.

A more glass-half-full view of it is that it's a magnificent success for Clojure.

Yes, it got redone in Java, but the fact is that Clojure took it from nothing to one of the world's premier frameworks for distributed computing - something that might have been impossible with the same resources if Java was its starting language. And because of the commonality of the underlying JVM, choosing Clojure allowed the functional approach with far less risk because the transition later was (I am assuming) much less risky and smoother.

Looking at it in that light, this would be a huge boon to people considering starting new projects in Clojure (or possibly other JVM languages): it proves if they succeed and get so popular that moving to Java is the right thing, it can be done.

How long does it take a strong developer to learn Clojure? Is it really a big a barrier to new contributors?

Presumably people who consider contributing to the open source have some spare time and love programming. Wouldn't they find it exciting to contribute while also learning a cool new language?

Maybe, but learning a new language by contributing to a large project built in that language is asking for trouble. Programming languages aren’t just syntax and stdlib. The relevant idioms and patterns may be significantly different. Even making small changes can be hard and intimidating.

And then, even if you wanted to learn something new, to a lot of people, it matters what that something is. Unless the person is really set on contributing to Storm specifically, then the barrier of Clojure is still significant. They may prefer to spend their new language learning bandwidth on something with more momentum.

My experience from hiring and training co-op students for many years is that it takes around 2 weeks to learn with a bit of guidance. when a co-op student with little to know programming experience can learn Clojure in that time I find it surreal that there are professional developers who cannot.

another interesting part:

The new Java-based implementation has improved performance significantly, and made Storm's internal APIs more maintainable and extensible.

Sometimes that backfires.

- Clojure dev

> At some point, you just accept that it's more important to get a new warm body than to continue pursuing some idealized programming perfection.

Describes my dating life pretty well.

Reply on news.ycombinator.com