Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Path: utzoo!utgpu!water!watmath!clyde!cbosgd!osu-cis!bgsuvax!denbeste
From: denbeste@bgsuvax.UUCP
Newsgroups: rec.games.misc,comp.sys.apple
Subject: Re: zork decoding
Message-ID: <1363@bgsuvax.UUCP>
Date: Mon, 9-Nov-87 14:09:59 EST
Article-I.D.: bgsuvax.1363
Posted: Mon Nov  9 14:09:59 1987
Date-Received: Wed, 11-Nov-87 23:26:48 EST
References: <400@ur-tut.UUCP>
Organization: Bowling Green State University B.G., Oh.
Lines: 25
Xref: utgpu rec.games.misc:951 comp.sys.apple:2965

in article <400@ur-tut.UUCP>, dmw3@ur-tut.UUCP (David M Walsh Jr.) says:
> Xref: bgsuvax rec.games.misc:816 comp.sys.apple:2495
> 
> 
>    Well I heard a few years back that Infocom used a 3/2 encoding scheme.
> They put 3 characters into 2 bytes by only using 5 bits per character.
> This leaves 1 extra bit for God knows what.

That extra bit is used to tell if the 3/2 encoding scheme is used, or if the
next 7 bits contain a straight ascii character.

> assume that they do use the "left-over" bit because they do use more than
> 32 characters (I think.)  The alphabet is 26, the space key makes 27,
> adding punctuation can take it over 32...

The lowercase alphabet and space are in there.  I don't know about the
other 5 characters.  All of the stories that I know of use both upper and 
lower case text.  For machines, such as the //+, that don't have lower case,
the text is converted on the fly.  Word wrap is also handled on the fly.

I have a friend that spent some time looking into this, but the accuracy
of the above is subject to the reliability of my memory

---
          William C. DenBesten | CSNET denbeste@research1.bgsu.edu