alexis@dasys1.UUCP (Alexis Rosen) (07/11/88)
Shortly after my first report on HyperCard, I found a copy of the new version,
1.2.1, and did some experiments. I am glad to say that my expectations for
V1.2.1, as written at the end of my first article, were for the most part
correct. There is a big speed improvement in many areas. One obvious win is
importing data. I was able to import my entire dataset of 33K records (HC
V1.0.1 couldn't import more than 19,017 records, for some reason). For my
dataset (see previous article), import speed is 50% to 100% faster.
Unfortunately, that's where my tests stopped. This freshly-generated stack
CONSISTENTLY CRASHES HyperCard when I try to compact it. Compaction is
necessary for reasonable performance with large stacks. Under UniFinder on
a 4MB Mac ][ with no inits except Apple's PramFix V.2, System 5.0, Hyper-
Card would consistently crash with ID=01 (bus err, I think) within about an
hour. Disk space was NOT a problem (9MB stack, 19MB free). Under MultiFinder,
2560 Kb partition, it would always "unexpectedly quit(1)".
THIS IS RIDICULOUS!!! Here I am, telling everyone how HyperCard is actually
capable of handling large files, and now this!
I do have one possible clue as to why it's crashing. My data is somewhat
atypical in that a certain three or four words appear on almost every card.
Furthermore, each card has a unique word which appears on no other card. If
HyperCard is building some some of inverted list in memory, this could
result an unusual demand on memory... but probably not. I'm certainly not
going to trace through their code to figure it out!
Whether my data is unusual or not, this behavior is *TOTALLY UNACCEPTABLE!*
At the very least it shows a lack of attention to detail in error handling.
At worst, well, it makes HyperCard useless for large stacks. 'Nuff said...
I will attempt to rebuild this stack, again, from scratch. I suppose there is
a one in a million chance that my hard disk has undetected damage which has
trashed a card, causing the compaction routine to fail, but even if this is
so, HyperCard SHOULD NOT CRASH! There is NO excuse for a bomb, EVER!
I'm really pissed off, in case you haven't noticed.
Well, Apple, say something...
--
Alexis Rosen {allegra,philabs,cmcl2}!phri\
Writing from {bellcore,harpo,cmcl2}!cucard!dasys1!alexis
The Big Electric Cat {portal,well,sun}!hoptoad/
Public UNIX if mail fails: ...cmcl2!cucard!cunixc!abr1