RnRS-tools

[Lassi Kortela]

There now exists srfi-tools
(https://github.com/scheme-requests-for-implementation/srfi-common/tree/master/srfi-tools).
We wrote it with Arthur and it is proving to be helpful with a growing
number of SRFI-related tasks.

I'm thinking it'd be useful to have a similar rnrs-tools for working
with the Scheme reports. Would somebody like to help me create it?

We could have separate r5rs-tools, r6rs-tools, and r7rs-tools. But
there's a lot of common ground between the reports (same TeX template,
etc.) so it would seem profitable to combine them.

I also expect it would be useful to separate the "apolitical" tools from
the "political" reports themselves. Different people are interested in
each kind of work (or the same person wearing two different hats). The
tools would just take the contents of the reports as handed, and do
whatever technical things people need done to them.

What tools should there be? The first order of business would be to give
a proper framework to the scripts in
https://github.com/schemedoc/standards (reproducible builds of
"official" and "errata-corrected" RnRS PDFs), figure out Dorai Sitaram's
tex2page for making HTML copies of R6RS, and adapt tex2page to R5RS and
R7RS as well.

Ideally the R7-large Publishing committee would incorporate their tools
into rnrs-tools as well.

[Lassi Kortela]

One more point: Once finalized, the Scheme reports are set in stone
(modulo unofficial errata corrections, which are sparse). But tools are
software, which is intrinsically prone to bit rot and benefits from
continual maintenance work such as porting.

The RnRS repos have thus far combined the TeX as well as makefiles and
scripts into one repo. Inevitably, the TeX markup is fine whereas the
scripts have bit-rotted. Some of them require obsolete dependencies
(e.g. PLT redex for checking the R6 operational semantics). For this
reason, I'm partial to importing known-working versions of dependencies
like tex2page and redex into the rnrs-tools repo.