Geoff Canyon wrote:
>
> On 7/21/99 3:30 AM, dohna <[EMAIL PROTECTED]> wrote:
>
> >> but at least HC seems fairly robust. On the other hand, I haven't yet
> >> used any stacks with thousands or tens of thousands of cards.
> >
> >Well, I have, and that was the main reason for using Oracle as a database:
> >HyperCard is just not robust enough. And you get so much in addition. If
> >you once have seen what you can do with SQL... Hey! I'm not selling
> >Oracle here... get any database.
> >
>
> I'm not really arguing, but I have to throw my two cents in. I wrote a
> HyperCard stack that contained, at its peak, about two thousand cards. It
> was dynamic, in that over the course of time it added and deleted cards
> on its own, to match an external data source (the directory lists of
> several Microsoft Mail servers). It maintained an index of the cards, and
> was active twenty-four hours a day, seven days a week, for over a year.
> It never had a problem.
>
> Let's hear it for the robustness of supposedly trivial tools!
In the book lending system I wrote there where ~ 30'000 books and ~
3'000 users lending ~ 100 books a day. There where index stacks that
maped the user or book numbers to card ids of the stacks where the real
data was. Every card had ten fields for the last digit of the user or
book number (read in from barcodes), 1000 lines for three more digits
and the rest of the digits was the card number, so access was
sufficiently fast. It would have been much easier, if in HyperCard you
could have set the id of a card as you can in MetaCard.
That worked all quite well until under heavy duty when one of the index
or data stacks got corrupted about once every month. I could reduce that
to about once every three months by compacting the stacks at least every
day. That also gave a solid method to find out IF a stack was corrupt. I
then built "redo log files", so I could rebuild the data when a stack
got corrupted.
It was tedious with a lot of hand work. And there was constantly that
certain uneasyness: "When will it crash again... can there be crases
that cause loss of data?". So I was really happy when I was able to
migrate the system to Oracle.
Just an experience report... they say sometimes somebody even may learn
from mistakes somebody else makes ;-)))
Regards
R�diger
--------------------------------------------------------------------
| Ruediger zu Dohna GINIT GmbH 0721-96681-63 [EMAIL PROTECTED] |
| PGP-Fingerprint: F1 BF 9D 95 57 26 48 42 FE F8 E8 02 41 1A EE 3E |
--------------------------------------------------------------------
