Hacker News Reader

No. It's not the Java language. It is the developers. They are the ones who have to have 35 levels of abstraction. Why? Because for some reason they think this will make modifying the software they write easier. In reality, no, it doesn't because everyone gets frustrated wading through the abstractions. If I can hold ten items in my brain's working memory and all ten are taken up by keeping track of how many levels deep I'm in, then it is really hard to make any real changes to the code.

It's not a language causing shitty abstractions, it's not a framework causing shitty abstractions. It's shitty developers.

One can write a Java application and never touch XML, use AbstractFactoryWhateverTheFuck, and write sane, clean code. I've done so for years, worked on green and brown-field projects. I've also written, maintained, and modified monstrosities in Java... as well as C#, Javascript (front and back-end), PHP, Python, and plenty more.

Java is not the Hot New Thing™ and never will be. It's also not always the answer, but for a run of the mill application running on a web-server you're going to have a hard time finding a better language and ecosystem that will keep on trucking and be reasonable to work with.

The presence of many of these frameworks, spring, AOP etc, is proof that many people do not find Java reasonable to work with. If Java was sufficiently expressive, you would need only libraries.

What, specifically, about having a framework means the language is not expressive enough?

Frameworks do not compose. Libraries do. If you need a framework, your language has failed.

No one needs a framework, but if one makes things easier then what is the problem?

What, in your mind, is a good language that needs no framework?

Everything has a cost. As I said, frameworks do not compose and often fundamentally take away control. If you want to change the behaviour of your framework, you need to hope they have created the necessary callback interfaces.

Not parent, but for example, this is how easy serialization is in Scala Play in the majority of cases:

    case class Person(name: String, age: Int, likes: List[CustomObject]) //simple data object
    implicit val personFormat: Format[Person] = format[Person] //this creates a "Format[Person]" with serialize/deserialize methods

The code to generate a serializer/deserializer is completely independent from the definition of the data, no annotations or XML needed. It is also efficient in that it doesn't use reflection. This is only possible because of how expressive Scala is - macros and strong type system means a lot of powerful code introspection can be done at compile time.

Because of implicits, most of the time I write that one line above, and an HTTP handler as a simple function from (Person => HttpResponse) and dont even think about the serialization/deserialization. And that one line is not an opaque abstraction, it is very customizable if you need it to be, and all via code not config.

Using Jackson in Java can get much more complex because annotations and XML files creep across the whole code if you want compile time generation. I think this is directly because Java's inspect/generate code at compile-time is not as powerful as macros.

> If I can hold ten items in my brain's working memory and all ten are taken up by keeping track of how many levels deep I'm in...

If you find yourself doing this, it is a clear indication that the software design is failing to abstract properly or at all.

(I'd argue that the industry in general has a terrible ratio of "coders" to people who can design programs, thus you're much more likey to come across poor designs than good ones. Beware of this when judging programming languages.)

> a clear indication that the software design is failing to abstract properly or at all

I believe, it is an indication that people tried to fix all the leaks in the abstractions, which turns the abstraction into an ugly twin of the stuff it tried to simplify.

ORMs are a good example. They start easy and simple, but lack various features of full SQL. If you start a new project, the ORM is fine. At some point, you are forced to use plain SQL, which clashes horribly with the ORM concepts. Thus the desire to extend the ORM. Over time more and more features are bolted on and the complexity grows.

IMO that's still bad design on the end user of the ORM library. There's a time and place to use tools like Hibernate, but it became fashionable so it's used all over the place where it doesn't fit and has created huge messes.

There's a really bad trend in the industry of choosing libraries or tools based on trying to improve your resume, rather than trying to find the best solution to a problem.

Absolutely, just look at all the Hadoop users who don't really have big data.

I've worked with a number of ORMs, mostly home-grown, admittedly.

I've yet to see a good one.

Often ORMs fail to implement JOINs efficiently.

Every ORM fails to give you a powerful query language: if they had a powerful query language, someone would write an ORM for the ORM, because the reason ORMs exist is that "SQL is hard", as the doll says, and any powerful query language will get the same treatment. But SQL is not really hard, and not having SQL eventually means you pay a huge price.

And then there's things like LDAP, which are absolutely awful. LDAP's object naming scheme has no canonical text representation. Let that sink in. And that's not even the main problem -- forcing your database to be hierarchical and a collection of objects, with no standard relation/pointer type, is. (Yes, Active Directory has an Object DN type, I'm aware.)

Just say NO to ORMs.

This is a well known issue [1], and isn't due to ORMs being poorly written.

It's not true that people who know SQL always prefer using it to using an ORM. In one of my prior jobs we had strict controls on the queries that ran in production for performance reasons. We screened all of the query plans the planner generated, but we also used an ORM. Sometimes we needed to work around or even reimplement the queries that used to be in SQL due to niche features like index hinting or the HAVING clause. And yet usually we used the ORM because it was more concise and reduced manual mapping of columns to fields.

[1]: https://en.wikipedia.org/wiki/Object-relational_impedance_mi...

> They are the ones who have to have 35 levels of abstraction.

Have you seen FizzBuzz: Enterprise Edition ?

https://github.com/EnterpriseQualityCoding/FizzBuzzEnterpris...

haha thanks for sharing that, one of the most absurd things I've ever seen. Every once in a while I think about picking up Java or .NET and going for the enterprise-y jobs...no thanks.

I'm going to be laughing about this for days.

its beautiful, not only is it absurd and hilarious, but it says so much. Not as a direct diss to java, but how badly a simple task can get over-designed to the point of being indecipherable. What would be great is if they added some kind of CRM or e-commerce to it, and donated the proceeds to charity. So like, could "Pay as a service" or sign up for "fizzBuzz as a service" and pay to run fizz buzz up to certain limits.

I'm not sure you can really separate the two. Sun pushed out the language and also helped develop much of the culture (enterprise beans, the pet shop example). Languages shape peoples thinking.

I disagree, it is some developers. Just as some developers can create a big ball of magic code dynamic code or as some developers can make insanely cryptic code in Scala/Haskell that looks like the user is typing in windings.

Java 8 and the current ecosystem can be incredibly lean and devoid of high levels of abstraction. Like today how I took the lovely Sinatra inspired http://sparkjava.com/ out for a spin. I think if Java can shed the memory of it's horrible past (and the enterprise space was truly bad) then it has a lot to offer developers these days.

I agree that there is a Java culture, but I also think that it developed into the way it is because it was so popular during a time when there werent very good engineering practices by non-technical people. The movement towards wanting managers with a technical background I think is fairly recent. The majority mass of java developers now probably "grew up" in the jaded horror story environments we seek to avoid.

Also, i think new grads tend to only know how to make things dry via abstraction (not considering the trade offs) and b/c of javas popularity and oop as a buzzword, always did it via inheritance.

Java libraries are complex because Java is inexpressive? Not because the libraries attempt to do everything for every situation? Are there some concrete examples you can point to?

It has been my experience and observation that something like a JMS library in Java are complex because of a different interpretation of "do one thing and do it well." In the Java world, possibly because of enterprise adoption, that means pluggable transport for the queue, pluggable aaa, pluggable security and encryption, pluggable durability storage, configurable delivery protocols and guarantees, aspect based hooks to support different ha patterns, pluggable dispatch on delivery and probably some other things I can't think of this early in the morning. Each JMS implementation is "the best damn message queue ever made" with tons and tons of features. That's not an expressiveness issue, unless it's too expressive. Java makes it easy to add a layer of abstraction at each of those decision points. The typical user here wants to get a message from one piece of software to another, maybe with ordering; JMS will do that more ways than you can imagine or care about and can use pieces of machinery you didn't even want to know about. Where does expressiveness play in to that?

> Java libraries are complex because Java is inexpressive? Not because the libraries attempt to do everything for every situation?

Why do these reasons have to be mutually exclusive?

Furthermore, your second point relies on every single cumbersome/complex java library also being a perpetrator of trying to do too much. Surely a lot of people here can probably think of smaller libraries that explicitly don't try to do everything, and are still "complex" and cumbersome to use. The best examples I can think of were of internal libraries at previous jobs, and that happened across many languages not just java. These sorts of annoyances pop up repeatedly in a wide variety of languages, despite not necessarily having the same culture as java.

Not to play evangelist here, but I just haven't had as many of these kinds of issues with libraries since switching to more modular typed functional languages (i.e. ML/Haskell), and it has really played a role in helping me see the forest instead of the trees. Noticing how language complexity and inexpressiveness leads to those kinds of problems isn't supposed to be obvious or intuitive, because it's effectively a Butterfly Effect that causes most of its severe repercussions much further down the line. We can play whack-a-mole with smaller sources of complexity (such as java having kitchen-sink libraries) as long as we like, but shouldn't the goal be to hunt down where all those "complexity moles" are even coming from in the first place? Because if that is the goal, and even smaller more focused libraries still produce annoying amounts of incidental complexity, then there's not much room to blame anything other than the constraints of the language at that point.

> Java libraries and frameworks are complicated because Java the language is not expressive enough.

I have written Java code and rewritten code large parts done by others. The issue I have noticed is architecture and framework mindset of people who just do not think solving problem at hand is important and good enough task. It would be ok if they wrote elegant library/framework but it was still junk and solved only very narrow problem. At each instance it was the problem of culture and people using Java because I was perfectly able to write productive and simple code with just standard plus very few extra libraries.

> Another language in this category is Go. It's still young enough that people haven't used it much outside the niche it was created

Huh, may be you should be able to give example of highly expressive and popular languages which are a hit with developers in very different domains. Because it is easy to pile on popular Java and increasingly popular Go without talking about alternatives.

No Java is simply not expressive enough, that's why so many frameworks exist. Adding Lambdas certainly helps, but you are missing out if you restrict yourself to Java.

Let me give an example of how Java limits ones thinking.

Look at these Stack overflow answers: http://stackoverflow.com/questions/28459498/why-are-java-str...

They are great examples of Java/C# users being unable to explain succinctly what the APIs are doing, because they lack the necessary common language. They have to answer with great essays describing various scenarios and use cases.

A Haskell user just needs to compare the type signatures, showing where the IO is happening. She will then understand the differences and the implications.

I guess I am not obsessed with expressiveness. I think of programing as something to help business/users to get on with what they want to do, as opposed to exercise in elegant self-expression. Towards that goal Java, Go seems very fine languages to me.

But as my example pointed out, expressiveness is often the difference between being able to understand something and not. Expressiveness permits better abstractions, better composition and better libraries, allowing cheaper and faster software. Something business users will appreciate.

I do not disagree that Java, Go could be better. It is just that I remain big sceptic because such better languages leaves very important things which matter in real world and covered by Java, Go etc but not by better languages e.g tooling, GC, powerful runtimes, build tools and so on.

I would wait till Scala does something great natively without standing on shoulders of JVM/JRE etc.

Python does not lack power. It's a different problem.

Monkeypatching is everything you need in terms of power. The problem is that it is too powerful, and Python lack intermediately powerful constructs. So if you have a problem that could be perfectly solved by passing around anonymous functions, creating a DSL, or writing some macro, you will either write everything by hand, Java style, or use monkeypatching and get a mess of mutable code.

Still, monkeypatching will solve the problem. It is more powerful than any of those other features. It's just that killing your fly can be best done with some small tool, and cannons leave a mess wherever they are used.

The last time I looked at Python (which was something like a decade ago) the impression I was left with was that there was a sort of cognitive dissonance or collective self-deception going on. There was a simple, clean language subset which was the one presented in the official tutorial, by and large; and lying underneath there was another language full of introspection and underscores, which seemed to be necessary to get (many) things done without a boilerplate explosion. The impression I got was that everyone writing Python was convinced that those other guys didn't need to think about these dangerous, advanced features, but they were in the 10× hacker elite for which the power was necessary and appropriate. Now probably that's an exaggeration and somewhat unfair, but I think it's probably not completely wide of the mark. ;)

My experience with Python is that 90+% of your code base can be in the "nice" subset, but it requires some discipline. The key is to always, always, always encapsulate the complex parts in helper functions or libraries, and accepting that maybe it isn't worth making your code twice as complicated to save a single line.

The official answer from any Python developer to anybody asking about "magic" is "do not use it". It's simple, gets the message across, and when the ones asking are experienced enough to not need asking about it, they'll just be one disaster away from being experienced enough to know when to use it. (There are plenty of resources on how to do it. Just no assurance that you should.)

Overall, it's a more useful response than the Ruby's community one. Ruby seems to have inherited the Lisp propensity of meta-programing everything so many of its basic libraries are incredibly hard to follow. Even with Ruby's metaprograming features being easier.

Still, providing safe metaprograming features sidesteps everything. I do love Haskell.

Go is as old now as Java was when Spring and Hibernate started to appear on the scene. If it time leads to this kind of pattern, why was Java 'over the hill' by this age, yet Go is still considered young?

i think java came up when oop was being hyped (the nominative subtyping variety) and design pattern mania ruled the land.

such things are less en vogue these days.

Java had widespread adoption.

But, if Go hasn't already achieved the same kind of adoption after this many long years, doesn't that suggest that Go will always be relegated to the niche it was designed for? In which case, according to the parent, it won't need to adopt those patterns in the first place.

Java was adopted very rapidly (as soon as it ran acceptably on commodity hardware) because the industry desperately needed a safe subset of C++. The pain point addressed by Go is mostly deployment and packaging, which is why mostly devops (who have rejected OS packaging for some reason) are excited about it.

> The pain point addressed by Go is mostly deployment and packaging, which is why mostly devops (who have rejected OS packaging for some reason) are excited about it.

What happened to OS packaging was mostly a result of the arcane tools around .debs and .rpms, which led to sysadmins reinventing the wheel (poorly) using symlinks, tarballs, and shell scripts.

FPM helped to a great degree, but even then, what if you want two different versions of the same application on the same machine? And if you hack your way around that, how will you manage which instance of the application gets which port?

When Docker came out, that kind of ended the idea of using OS packages for deploying in-house applications. Why bother with a packaging infrastructure that will only tie you to a certain OS?

I still think OS packages have a place, especially for system level software. Nobody ever complains about how SSH failed to install or upgrade...

But even when using containers for internal software, Dockerfiles can still be a pain to deal with for the simple reason that most languages have horrible build-chains that produce unusable outputs, which as you said Golang addresses beautifully.

However I think you're off the mark when you say "mostly devops" are excited about it. It's entire companies that are excited about it, and for one simple reason: Languages with these bad build chains and bad outputs have led to the creation of an entire teams of people to deal with it.

The realization of this is spreading, and it's going to be an interesting time.

It was always fashionable to hate java. Java was bad and old pretty much right after being born.

That is an insightful comment. What languages do you consider (most) expressive?

Are you saying Scala got it right? What's your take on Kotlin?

Kotlin is a cool language, but doesn't do enough to differentiate itself from Java 8.

There are real benefits to picking the mainstream / popular option, which includes tons of documentation, forum discussions, books and developers available immediately for it. So when you go against the mainstream, those alternative tools have to be worth it.

I think Scala got over the worthiness threshold, by exposing one of the best OOP implementations exposed in a static language, sane generics, higher kinded types and implicits which make it possible to encode type-classes, which led to a community emphasizing on functional programming.

I think Kotlin isn't over that threshold. Some people view its thin layer approach to be an advantage. But it either has to evolve in being more than that, or to divorce itself from the JVM, because it is in direct competition with Java 8 otherwise. Evolving to be more than a Java++ goes against its own philosophy. Divorcing itself from the JVM is a work in progress, they also have a Javascript compiler, but the results aren't convincing enough for now.

It's a nice language, but personally I like those with a strong community of functional programmers and libraries. If I'll see http://typelevel.org for Kotlin, I might reconsider.

> but doesn't do enough to differentiate itself from Java 8

Short of implicits and higher kinded types what has Kotlin not yet copied from Scala? FP in the pure sense is still very much niche, nobody cares about Scalaz and Cats in Java land.

Providing a less powerful Scala as an alternative to Java might very well prove to be a great strategic move. Scala's certainly not going mainstream anytime soon, probably in part due to its being too powerful.

Going on 6 years in Scala, staying put for now, but will jump ship if/when something better comes along (for me Kotlin is not it).

I find Scala enormously impressive because it manages to bridge the gap between subtyping. Not sure whether that is worth the complexity this brings, though.

Subtyping makes a lot of things that are easy in purely functional languages hard. I think that is also the main reason why you don't get global type inference.