GNU Automake - Future directions

This page tries to explain some of what we have planned for Automake's future, both near and long term.

Warning

This page was written before the 1.5 release and is completely out of date.

Short term

In the short term, we are working towards Automake 1.5. I want this to be a fairly major release, implementing a number of frequently-requested features. Whether or not this goal can be met in a timely way is uncertain, to say the least, so don't be too disappointed if 1.5 doesn't live up to the expectations I'm about to set.

Rewrite the manual. Progress is slowly being made here.

Medium term

Get rid of aclocal in favor of some better solution. This probably requires autoconf changes as well.
Support texi-to-html generation.
Finish multilib support, and in general better support cross compiles (this probably requires autoconf support).
Add features to automake to support packaging tools a bit better. In particular it should be possible to associate an installable file with some subpackage name.

Long term

In the long term we are going to rewrite automake in Guile. That should improve the internal structure a bit, and provide opportunities to regularize its feature set.

We'd like to change automake so that implementation variables and targets in the generated Makefile.in follow a naming scheme so they can easily be avoided by user code.

We're also interested in changing automake so that it can more easily be used to generate a single large <tt>Makefile</tt>; see Peter Miller's paper Recursive Make Considered Harmful for more information.