le 2001/04/03 16:09, Stuart Price à stuart.price@... a écrit :

>> Might it be an idea to enquire about the arrival time of the Object Runtime.
>> Methinks that a Record Object would know it was the same > as another,
>> sister, Record Object, and then it would just be a
>> question of testing contents,
>> again, you could probably code this method into your Record Object.
> 
> Isn't that the same problem only worse.  With a record all of your data is
> stored in one contiguous piece of RAM, so being lazy I was trying to find a
> way of comparing it all with another block the same size but somewhere else.
> With an object you may not choose to store all of its data in a record so a
> specific comparison method is your only option.

Well I believed that old records, so-called pseudos, stored info in
contiguous RAM, so the only trouble in that case was skipping padding bytes
in strings.
So-called 'real' Records, I have no idea where they do/put anything: they
seem some sort of abstraction like Abjects, so it just seemed to be a good
idea to make them bedfellows.
If you try to do 'real' Records comparaisons, you'll need to know the
structure (an FN by record type?) or, as I thought, just abstract further by
going Abject.

Perhaps I'm not ready to be abject orientated after all.

:-j

[gnomes are not upwardly orientated, it's just the dumb spikey hats that
give that impression]