John Porter wrote:
>
> Here's my attempt to summarize the issues in detail.
In my haste to get this out to the list, I forgot some things.
An additional Pro for Approach 1a:
> 1a. Write and run native Perl5.
>
> Pros:
. Ready to segue into Approach 1b when possible, i.e.
when the perl5 compiler is finished.
And another set of possible appraoches, involving the use of some
back-end language(s) other than C*. So:
4. Write perl6 in Perl; compile the Perl into HLL (e.g. Scheme).
Pros:
. Generating HLL from Perl may be significantly easier than
generating C*.
. The possibility of generating multiple HLL's is opened up.
. Generated code will generally be more compact and/or platform
independent.
Cons:
. Performance may still turn out to be an issue.
. Integrating the HLL code with the unavoidably necessary C
stuff on the target platforms may be a major headache.
--
John Porter
Aus tiefem Traum bin ich erwacht.