[its-hackers] Muddle Oblist observation

Chris Reeve chrisxreeve at gmail.com
Mon May 3 02:27:13 CEST 2021


Hi Rick,

You reminded me of something. OBLISTS were initially as described in the
language manual. We discovered OBLIST searches were very slow. We changed
things such that there was one  hash table of ATOMS (probably the notorious
ATOM-TABLE?). An atom pointed to the name of its OBLIST. .OBELIST was a
list of OBLIST names. It's scary that I remember that from 35 or more years
ago. I don't think this change was visible except to code mucking around
with OBLISTs. The other vague memory I have is writing something like
release notes for new MDL releases. These release notes probably explained
the change. When I have some time I'll investigate further.

-clr

On Sun, May 2, 2021 at 7:28 PM rrs0 at earthlink.net <rrs0 at earthlink.net>
wrote:

> Hi Everyone,
>
> I Just notice a difference between how ITS Muddle handles oblists and
> what's stated in GFP and SWG's "The MDL
> Programming Language". The manual says that an OBLIST is a of PRIMTYPE
> UVECTOR and UTYPE LIST where
> The LISTs are hash buckets for ATOMs. The default bucket count is 13
> (Note: a prime number).
>
> What I have observed all OBLISTs have one hash bucket in them. See blow:
> I admit this makes getting all the ATOMs in an OBLIST easier when there is
> only one hash bucket per OBLIST,
> but this isn't what the Muddle Manual said. Is that any different in MIM???
>
> <LENGTH .OBLIST>$
> 2
> (<GET <1 .OBLIST> OBLIST> <GET <2 .OBLIST> OBLIST>)$
> (INITIAL ROOT)
>
> <SET FOOBL <MOBLIST FOO 3>>$
> #OBLIST ![()!]
> <SET BNARBL <MOBLIST BAR>>$
> #OBLIST ![()!]
>
> <FLOAD "ITSPKG >">$
> "DONE"
> <LENGTH .OBLIST>$
> 2
> <GET PACKAGE OBLIST>$
> #OBLIST ![(PKG!-PACKAGE)!]
> <GET PKG!-PACKAGE OBLIST>$
> #OBLIST ![(SEARCH!-PKG!-PACKAGE IPKG!-PKG!-PACKAGE)!]
> <SET BIGOBL <MOBLIST BIG 237>>$
> #OBLIST ![()!]
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://its.victor.se/pipermail/its-hackers_its.victor.se/attachments/20210502/f363afae/attachment.html>


More information about the its-hackers mailing list