Why?!? Play2 is essentially a collection of specialized Akka actors. When you run Play, you get Akka integration for free. Having an architecture where your Play actions send Akka messages to a cluster of computational nodes/actors requires about 3 lines of code and 5 lines of config. Ridiculously scalable.
>Having an architecture where your Play actions send Akka messages to a cluster of computational nodes/actors requires about 3 lines of code and 5 lines of config. Ridiculously scalable.
Not really. Scala has a lot of good points but it's very ill-suited to web development because of compilation times.
Play1 could compile templates in a few tenths of a second while it takes 7-10 seconds for each template in Play 2, which is unacceptable when you do this all day long.
I predict Play 2 will not be around in a year from now for this reason, there are just better and faster ways to do this.
The Play team or someone else could just re-implement the template layer as Groovy once again.
Scala has a lot of nice things going for it, but using it for the Play2 template language was stupid - the benefits do not outweigh the drawbacks at all.
Weren't there quite some different templating choices (including Groovy) available from the beginning?
Independently from that, the compilation times will automatically improve when they update a) their SBT version and b) their Scala version. Additionally, Play developers can also refine their own code.
When you make changes to a template it takes a few seconds to recompile the template (once ) then after that it load much faster since it doesn't need to compile a second time. In addition the resultant code is rendered much faster the then old Groovy templates ( in my experience ). Is this more of a development time issue for you or are you noticing issues in a deployed system ?
8 comments
[ 4.3 ms ] story [ 29.1 ms ] threadScale what?
Not really. Scala has a lot of good points but it's very ill-suited to web development because of compilation times.
Play1 could compile templates in a few tenths of a second while it takes 7-10 seconds for each template in Play 2, which is unacceptable when you do this all day long.
I predict Play 2 will not be around in a year from now for this reason, there are just better and faster ways to do this.
Scala has a lot of nice things going for it, but using it for the Play2 template language was stupid - the benefits do not outweigh the drawbacks at all.
Independently from that, the compilation times will automatically improve when they update a) their SBT version and b) their Scala version. Additionally, Play developers can also refine their own code.