A few comments and suggestions:

Look into "Reproducible Research" methods and find ways to help more
crusty scientists adopt them.  Emacs Org-mode documents and IPython
Notebooks are two popular ways to write a paper that you can rebuild
when data or analysis is updated.

Your "linear slider" concept misses the fact that scientific research
results are better represented by a (humongous) directed graph of papers
joined by their citations.  More tools to understand and query that
graph would provide real benefit to scientists and laypeople.  But, any
given paper, especially one following Reproducible Research methods, can
be placed in version control (and often is) which effectively gives you
the history (still potentially non-linear) that you describe.

My (possibly faulty) understanding is that you retain copyright for your
own research result materials prior to their publishing.  This includes
preprints and possibly even the as-submitted copy.  At least as far as
the publisher is concerned they only assert copyright on the copy they
publish.  You should be free to submit your copies to an open
distributor such as arXiv.org.

-Brett.