-- Jeffrey Carter:
-- I don't see a need for an AVM. However, a JVM implementation in Ada,
-- open source and free in both senses, might be a worthwhile project if
-- anyone cares to undertake it.
I could not agree more. With HW support for kJava (see fx. ARM Jazelle =>
http://www.arm.com/sitearchitek/armtech.ns4/iwpList76/5B0F7D3A50CA4252802569
E0003620BE?OpenDocument&style=CPUs) there is no need for another MeToo
technology. What is hot these days in R&D labs for mobile computing:
* product-lines, generative programming
(see fx. Aspects => http://aspectj.org/servlets/AJSite,
Jakata/GenVoca => http://www.cs.utexas.edu/users/schwartz/ etc.)
* XML
* architecture frameworks & quality issues
(see fx. Holt => http://powerlips.ece.utexas.edu/~jolt/,
Bosch => http://www.cs.rug.nl/~bosch/ for links)
* Java support
Programming wise it also back to the "Basics" as C++ has failed to deliver
the promise. Basics is unfortunately in most cases to C! With 0.5-1.5M SLOC
in most mobile (handheld) units these days complexity is a major concern.
Ada is ideally suited for that job.
In my opinion Ada 0x revision should try to support some of the emerging
design/language topics like "aspects" - i.e. less emphasis on pure OO as an
salvage for everything - to get Ada back as a topspot language where it
really belongs.
regards /søren
Søren Henssel-Rasmussen
Nokia Mobile Phones, R&D Copenhagen