On Wed, Mar 13, 2019 at 7:46 PM Lassi Kortela <xxxxxx@lassi.io> wrote:
> Just realized: Do you mean that you'd eventually want to publish a
> standard syntax for static type signatures as part of the SRFI
> submission guidelines, so that authors of new SRFIs could directly write
> their own types in that syntax if they want to?

The short answer: no.  I don't think it is feasible;  nor will Arthur
accept it, neither will the other editors want to go with it.

All my previous statements take into account the workflow I've
proposed a few emails ago:
* the author writes the SRFI as he wants (and is accepted by the editor);
* the editor together with the author finalize the document in the
"free-form" HTML format;
* from here on the "volunteers" (or if time permits the editor and/or
author) would take the HTML, reformat it and augment it as needed, and
add the signatures;
* (nothing prevents the author to start with this last form, but it is
not part of the SRFI requirements / guidelines;)

> That would be really
> cool, but how and when would we know what the right syntax is and that
> it won't change anymore..

The volunteers can know that the syntax is right by running an
automated tool that validates and extracts data from the document and
additional files.  And if it changes we do as you did with the ISO
dates:  someone proposes the change, tries it out, changes the tools,
and submits a pull request for the existing dataset.  (But as the
format will be more geared towards automatic processing, this last
part should be easier to achieve than it is today.)

> I know the initial plan is to develop a
> Dialyzer workalike but Dialyzer is still being improved

Just to be clear:  I don't propose that we write any Dialyzer-like
tool.  I was just stating that having a more thorough procedure
signatures we could in the end write something like Dialyzer.

Ciprian.