[OT] Re: A vision for Parrot

ruby-talk
1

  1. There is a fundamental difference between the two projects: at least
    both Perl and Ruby people seem to be interested in using the VM as a
    language backend, but I don’t think anyone outside the Guile community was
    interested in writing the aforementioned translators.

  2. Using Parrot as a backend has speed as a carrot, while using Guile as a
    backend has no carrot that I am aware of (ideologigal issues besides).

  3. As far as Schemes go, Guile is rather bloated: if the application is
    small, embedding Guile is a significant overhead.

– Nikodemus

···

On Wed, 6 Nov 2002, Andrew Dalke wrote:

It didn’t happen. Writing those translators are hard because each
of the languages has different object models, which must be implemented
nearly perfectly. Guile is a full-fledge language built on years of
research into Lisp and scheme, so I would be surprised if it was any
easier to implement in Parrot’s byte code. It didn’t happen with
Guile, it’s even less likely to happen with Parrot.

2

In article slrnaskqgb.3pa.akuchlin@ute.mems-exchange.org,

In article Pine.GSO.4.44.0211061146570.12517-100000@kekkonen.cs.hut.fi,
Nikodemus Siivola wrote:

  1. Using Parrot as a backend has speed as a carrot, …

It does? Upon what evidence? Since Parrot doesn’t run any
languages yet, how can you claim it’s faster than existing
implementations of those languages?

At the Ruby Conference last weekend Dan Sugalski gave a talk
on Parrot. He showed a slide comparing the speed for counting down from
100million to 0 (or was it counting up to 100 million?) using various
scripting languages (Ruby, Perl, Python).
There were three Parrot versions impemented in Parrot assembly code -
non-optimized, somewhat optimized and very optimized. As I recall, the
scripting languages all took around 100 to 150 seconds. The worst of the
Parrot versions took something like 18 seconds and the best did it in 4
seconds… That’s a pretty dramatic difference, although it certainly
isn’t covering a lot of operations. At any rate, it seems to show that
for counting the parrot VM’s performance is quite good :wink:

Phil

···

A.M. Kuchling akuchlin@mems-exchange.org wrote: