I'd like to thank Alexander for all of his hard work. It's entirely my fault that his Neko work hasn't been merged in, but it's something I am going to get done soon.
Does anyone with Scala on Android experience know if it is bogged down by the same size and performance issues? I assume they are running against the same problems on parallel tracks. Just curious.
It's not bogged down by the same performance. There may be times where idiomatic Scala code generates more garbage than similar Java code and thus has more of an impact on Dalvik performance but these are straight forward to work around and well understood. For the vast majority of apps this isn't a problem.
Scala does suffer a similar size issue where the Scala standard library is a few megs and you don't really want to ship that around with your app. You can use ProGuard to strip out the parts of the Scala standard library you aren't using. This is baked into build tools (like jberkel's excellent android-plugin for sbt) so it's usually seamless.
FWIW, I've got a couple Android apps on the Play store written in Scala.
I'm a Scala guy learning Android. There's Scaloid which offers more concise way of doing things, but now tools (esp. Android Studio) are getting better for XML (and faster feedback thanks to live preview), so now I'd rather take the XML route (although probably as I'm getting more proficient with Android, I would prefer programmatic).
The performance is not bad. ProGuard is a must. The problem for me is compile time.
I've been using Scala to develop for Android for a while. There is this sample project to get you started with nothing Java, Ant and the latest Android SDK required: https://github.com/danielribeiro/HelloScalaOnAndroid
I think it's interesting how few comments clojure posts attract. Given pg's love of lisp and how amazing clojure is (at least in my opinion) in terms of language design and what features we'll need in the next 10 years, I would have thought the HN community would be jumping all over it.
This isn't meant to be anything more then an interesting observation.
Well, speaking for myself, I think the article is extremely interesting, but I don't think I have much to add to the conversation. Maybe Clojure just attracts lurkers.
(It should be pointed out that I have plenty of strong opinions I'm more than happy to tell to anyone who will listen, but when it comes to Clojure, I want to hear from Nathan Marz, James Reeves, Stuart Sierra &c, and seriously doubt I have much to add.)
One lurker here - I, for one, am EXTREMELY interested in CoA, and Clojure in general. I have only worked on a couple of Android apps thus far, but it is already enough for me to want to look into Clojure for Android development. I don't know if my Android chops are good enough for me to get away from the Eclipse/Java environment just yet, though.
I write Clojure code professionally and am deeply interested in all Clojure subjects that pass through HN (as well as Lisp/Scheme, Ruby, Scala, general CS, etc. -related posts on HN). But, I probably comment on only a small percentage of them, because, as another poster suggested, I don't have a lot to add other than "this is cool!" and I'd rather hear from folks who are more deeply involved in the projects under discussion--like Daniel Solano (https://news.ycombinator.com/item?id=5804036).
I also suspect that, given how often Clojure posts pop up and hang out on the main page, despite their relative number of comments, the interest is strong here.
If you look at hot topics like reducers, datomic and clojureScript, there's plenty of discussion. Haven't found a better way to find actively responded to threads than slogging thru hnsearch.com pages and Control-F for "comments |"
I think that it betrays the issue. If people are struggling to get started with the language, then that means that no one is bothering to use it beyond 4Clojure and basic koans.
The fact that there is up-voting and about one Clojure article on the front page every week shows that there is a strong interest in the language, but until the maintainers figure out how to make the language accessible, there won't be much commentary because very few people are knowledgeable enough to say A or B about it.
Can someone in the know comment on the size of the runtime? Why is it so large? Is this problem likely to be solved, or are we waiting for more powerful hardware?
The core libraries are very large and there is no standard way of stripping them out. Clojurescript (partially) addresses this by passing the generated js through the google closure compiler. Perhaps a similar pass could be made in the jvm compiler.