Hacker News Reader

Comment by hocuspocus | Hacker News Reader

Scala was a bad fit for the Android runtime and the SDK. On a $200 Android phone with 2015 specs the performance aspect was definitely not a trivial argument against Scala; Kotlin has perfect interop with very little overhead, no conversions between collections (often slow and immutable in Scala), a nullability story that doesn't box everything in Option, ...

And most Googlers do not like clever languages with bad tooling, especially the ones working to productize languages and toolchains for Blaze and other internal tools.

View on HN

All irrelevant when politics are involved.

You will never get me to say anything positive about the Kotlin advocates on Android team.

It isn't only Scala, it is using Java 7 samples to sell Kotlin, the original language for Android, still used for the large majority of Android tooling, where your argument fails flat.

They have begrudgingly being updating Java support, up to Java 17, when Java 25 is the latest LTS, because that great Kotlin interop with Java is useless when Android loses access to the Java ecosystem that keeps moving forward regardless of Android.

Some of them even don't have any ideas how out of date their Kotlin "improvements" over Java are out of date.

I'm not defending JetBrains or Google's decisions. But I don't see why anyone should try to rewrite history.

Scala had a decade head start. Many people at Google are aware of Scala, some being former LAMP students or staff. Google doesn't simply add a language to their small list of sanctioned toolchains.

Kotlin was specifically designed to win the politics argument, using technical merits: full interop, gradual adoption, low overhead, ability to target outdated Java bytecode efficiently (Kotlin used to inline lambdas for Java 6 bytecode for instance, while Scala boxed everything before invokedynamic). It doesn't matter whether one approach is objectively right or wrong, Kotlin fitted the bill and Scala did not. More importantly, Kotlin fitted the bill on Google's server side too!

Then of course JetBrains was tasked to replace the IDE with Android Studio, they could put dozens people backing Kotlin as a first-class Android language, until Google adopted it officially.

But even if Typesafe had somehow stumbled upon $100M in funding to do the same thing, it doesn't change the fact that Scala never had a chance.

Kotlin won Android on internal Google politics, zero technical merit.

Feel free to write another long reply on how Kotlin is somehow special, going to replace Java, while the first iteration of Kotlin Native was a failure with a broken design on its reference counting memory approach, and there is still no KVM to replace the JVM in sight, despite such greatness as pointed out in Android circles.

I haven't defended Kotlin's approach in any way, I don't know why you want to put words in my mouth.

However if you don't see why Kotlin was well positioned, while Scala was not, then you clearly don't know much about either language.

Technical merit is subjective, but different design decisions can precisely be made to win politics or not, and JetBrains made the right ones, first to convince people internally (who knew about Scala, you know), then Google, as they were fairly well aligned.

I know enough about Google politics, and JVM ecosystem is part of my toolbox since 1996, starting with JDK 1.0.1, I have lived through all hype cycles of JVM guest languages since Beanshell was introduced.

Reply on news.ycombinator.com