Path: utzoo!utgpu!news-server.csri.toronto.edu!cs.utexas.edu!hellgate.utah.edu!caen!sdd.hp.com!hp-pcd!hpcvra.cv.hp.com!rnews!hpcvbbs!akcs.gadiel
From: akcs.gadiel@hpcvbbs.UUCP (Gadiel Seroussi)
Newsgroups: comp.sys.handhelds
Subject: Re: HP48 compact arrays
Message-ID: <27b83c55:1997.3comp.sys.handhelds;1@hpcvbbs.UUCP>
Date: 12 Feb 91 19:40:20 GMT
References: <5966@hplabsb.HP.COM> <27b4e2a1:1997.2comp.sys.handhelds;1@hpcvbbs.
Lines: 18

Derek Nickel asks why not use Array of System Binary for compact arrays.
There are a few problems with that: I wanted to represent binary 
numbers of any length with the minimum possible number of nibbles
(up to 16 nibbles). Even if we ignore 5-nibble boundaries, we would
still need to have a multiple of 5 nibbles. But, worse, as you point
out in your notes, the system thinks that any array that is not of 
Reals is a Complex Array, and will let you edit it on the Matrix
Writer, which, in our case, could lead to catastrophic consequences
(try it on an Array of System Binary with just one element. BACK-UP
before you do this!!!).
What I wanted was a data type that could not be easily manipulated except
for the provided functions. That more or less rules out all other 
types, except Code, but I don't need to tell you why using Code would
be a terrible idea. So, so far I have not seen any convincing argument
against using Linked Arrays. By the way, one assertion (at least) in 
a previous posting of mine was incorrect: I do use a built-in object
(NEWOB) on linked arrays in the IPUT function, and nothing bad happens.
Gadiel