Mathematicians do research, write papers, referee papers, edit papers.
Mostly paid by university or government.
The result of this work is very often sent to commercial publishers.
A consequence is that mathematicians have to pay in order to read
the work of their colleagues. Pay large amounts of money.
And they do.
Mathematicians are idiots.
Of course commercial publishers are happy with the millions they
get for free. And universities, incorrectly, think they have no choice
but to pay up.
Mathematicians have a new obligation: Make sure the mathematics they
produce is freely accessible on the web. Stuff that cannot be accessed
(without paying a commercial publisher) effectively does not exist.
Material can be made available by publishing in an open access journal,
or by self-archiving.
Where? On a private home page, on a university archive server,
on a preprint server like arXiv, under scientificcommons.org, ...
Never sign unmodified copyright assignments to a commercial publisher.
Always retain the right to make your own work freely available.
Removing your paper from your home page when the journal article has appeared
means destroying your work. Someone else will have to redo it.
Nontrivial parts of mathematics are lost today, captured by commercial
publishers. If it is important material, it is a worthwhile activity
to publish survey papers with full proofs covering such topics,
so as to make these results available to a non-paying audience again.
Directory of open access journals