Scala macros are different. We deeply respect the Lisp tradition, in which we are rooted, but we also accommodate and empower our language's values - static types and rich syntax [1]. The end result is admittedly less flexible, but it also opens a number of interesting possibilities beyond those available in Lisp [2]. Our design is also not set in stone, and we're experimenting with different macro flavors [3], but it's too early to say something definitive about those.
Hi, that's me who's doing macros for Scala. Martin & Adriaan have already commented on my choice of the project and its status within Scala, but I'd like to additionally elaborate on your other points.
Speaking of target audience and usefulness. Even before being officially released, macros have attracted attention of popular community libraries and tools (e.g. SBT, Akka, Play, Slick), and later on we got even more people on board [1]. Therefore, even though not many people are writing macros, a lot of Scala developers are, sometimes even unknowingly, using them.
That's I think is the best thing about macros in Scala - natural integration with other language features, to the extent that people aren't even aware that they are using macros and not regular functions or implicits or whatnot. Surprisingly many features can be empowered this way, giving rise to new and effective ways of solving development challenges [2].
Speaking of the future status of macros. Half a year ago, we had a joint EPFL and TypeSafe meeting where we decided to stabilize a subset of macros (so called blackbox macros [3]) and include it as an official, non-experimental language feature by 2.12. Indeed, at times we have fierce debates on our mailing lists about this or that esoteric macro flavor, but that concerns only my research activities [4], not blackbox macros, which have reach unequivocal approval at scala-internals a long time ago.
This brings me to the next point - transparency of development. For the entire duration of my doctoral studies at EPFL that begun in Sep 2011, we were always open with our plans, welcoming feedback and actively answering questions on the mailing lists and stack overflow. Some of the most successful reflection and macro features are there only because of our close collaboration with the community, and I'm very grateful to everyone who has participated and is participating.
Finally, the thing about generating papers is just untrue. Sure there are people who do their research using macros, but if you take a look at the list of my publications, you'll see that over the two years that I've been a PhD student I have authored zero conference papers. This is the cost of seeing through an implementation of a feature in an official language release, addressing the feedback of early adopters, and then maintaining it as it's being used in production. But anyhow I don't mind how it turned out, because I'm very glad that more and more people are picking up macros and finding them useful.