Maude-help mailing list

[Steven Eker <steveneker AT gmail.com>]

For a language with user definable syntax, it doesn't seem
practical to parse a module FOO that imports module BAR before
module BAR has been seen since the syntax of say equations
in FOO may depend on unseen operator declarations in BAR.

Of course one might define a higher level entity; say a package,
that consists of modules/theories/views and when the end of
the package is seen, the import declarations of the contained entities
could be examined to see if there is a consistent order to parse them.
But even a module _expression_ can't be fully parsed without seeing
the user defined syntax.

Steven

On Thu, Jul 1, 2021 at 7:32 AM Marko Schuetz-Schmuck <MarkoSchuetz AT web.de> wrote:

Dear All,

I am using org-mode and its literate programming capabilities to write
and document Maude specifications. From org-mode I can "tangle" the
Maude source code blocks into a .maude file. I noticed that the sequence
of module definitions in the resulting .maude file _is_ relevant. For
some reason I had assumed that the modules could be given in any order
and that the Maude system could handle this. For such a literate way of
specification it would seem much more natural to be able to move text
(including source) around in the document without having to be concerned
with the sequence in which modules are introduced. From my cursory
understanding of the meta-level features, I suspect that some form of
module dependency analysis and reordering could be implemented as a
pre-processing stage. Has someone looked into this before? Would this
seem practical?

Best regards,

Marko