Re: [stack] Is this language concatenative?

Michael Nedzelsky <[email protected]> wrote:
> William Tanksley, Jr wrote:
> > Michael Nedzelsky <[email protected]> wrote:
> > > > But a quotation is not a member of T*, is it? Therefore it's not a
> > > > "simple program".

> > > It is a member if T*!

> > Oh. Then no, I don't think your definition is adequate; it admits all
> > ill-formed strings, such as "]]]]a]]]". Your grammar can't be parsed.
> Which grammar??
> My definition is about languages, i.e. subsets of T*.

Oh. Okay, now I see what I've been missing. Yes, your definition is
strictly about languages, not (per se) about grammars. That's a good
thing.

> S1 --> a | b| [S] (S+S)
> S --> e | a | b | S S | [ S ] | (S + S)
> Examples of simple programs: a, b, [], (a+b), [a b],(+), (a+(+)), (+b)
> The last three programs looks strange, but they are syntactically valid.
> The language L is (syntactically) concatenative by my defintion, since all the
> three requirements are satisfied.

Yes, I see your point.

Your definition seems to have the same problem our previous
definitions had: they allow things which are unmistakably not
concatenative (your infix (S+S) sublanguage isn't concatenative).

But your definition is better than most, because you include the
"unique factorization" property, and because it is expressed in terms
of concatenation and valid splitting, both fundamental to
concatenative languages.

Perhaps I'm wrong. Should our definition allow for languages which are
perhaps not _entirely_ concatenative, but whose "top layer" is
concatenative?

>  Michael Nedzelsky

-Billy