[
Lists Home |
Date Index |
Thread Index
]
As I recall, it did in cases where the declaration was
for a specific set of systems of known quantities and
types. 15 years ago, that was a lot more probable than
today. We tend to forget in this era of homogenization
of hardware and a vague notion that all systems are on
the network and are open what the situation was in the
time that SGML was being designed. Systems tended to
be uniquely connected, realistically closed, and had a
lifecycle that was longer given the expense to replace
it. Most of the applications were batch. As all of
that changed, the simplifications that are XML made
sense.
What I don't remember is if anyone except the markup
wonks actually looked at the capacities. I remember
using them for example in cases where nesting really
did matter.
len
From: Rick Jelliffe [mailto:ricko@allette.com.au]
Also, it is probably wrong-headed, if it supposes that schemas act to
determine or checkthe capacity of a system. Having maxOccurs="4294967296"
will not guarantee that your system will support 4294967296 occurrences.
(SGML allowed a matching of document quantitie with system capacity,
but I don't know whether it was actually much use.)
|
