[z-machine] [Inform mntce] The Z-machine and Informs 6 and 7

[samwyse]

On 10/20/05, David Kinder <d.kinder at btinternet.com> wrote:
> My understanding of the remaining problems:
> 1) Dictionary resolution. Having only 9 bytes of dictionary resolution
> causes lots of problems when you try to use lots of different characters
> in dictionary words. Sam's V9 work has a solution to this, and in fact is
> probably my favourite bit of V9: it just allows longer dictionary entries.
>
> It occurs to me that perhaps this variable dictionary length could be moved
> from V9 into Z-Spec 1.1, maybe?

"Shiver me timbers! I've been spotted! I'd best hie meself off to the
maze to hide me secret agenda!"

> 2) Arbitrary Unicode output. Earlier Z-Spec revisions had a scheme for
> printing out any Unicode character sequence embedded in a string, though
> after debate it was dropped from the last version. Perhaps it should be
> brought back.

I don't recall the earlier revisions' details, but is there any reason
why a special ZSCII sequence couldn't just flag the rest of the string
as UTF-8?

> 3) Compiler support. Most of the work is in getting I6 to cope with Unicode
> input. The idea of translating Unicode input to @{xxx} style input seems
> feasible, though the error messages might be tricky. Still, if we're worrying
> more about I7 than I6, I suspect we can live with it.

I can live with only allowing the input of characters that have been
defined via Zcharacter statements.  The parser is designed about
languages with alphabets, not ideograms, and most players are
bi-lingual at best.