Geoff Worboys wrote:

> Lester Caine wrote:
>> It HAS been demonstrated in the past that Widows did not
>> ACTUALLY write anything to the disk until the application
>> accessing the disk closed, or dropped the connection to it.
>> As soon as the connection was shut, the reported file size
>> would increase.
>
> The reported file size is actually quite distinct from whether
> the data is written to disk or not (partly dependent on what
> you mean by "reported").
>
>> Even with 'forced writes' on, pulling the plug DID loose the
>> whole of a days work! I could recreate it reliably on the bench!
>
> NTFS or FAT? SCSI or IDE? (Different bugs have had different
> effects depending on the combinations you were running.)
>
> I am not denying that there have been problems... simply that
> any description that includes "a whole days work", where not
> even the generators were updated, on an XP machine seems more
> likely to be something other than caching. And these days
> there are many products that will do this for you intentionally
> so it's work checking.

The original question was - has anybody seen this - and the answer is yes.
Yes it was some time ago, but I was experiencing exactly the same problem and I
seem to recall a discussion on one of the lists at the time, but although I have
quite a history I'm not currently finding a link. My own problem was cleared by
a setting in the disk driver, and once this was changed one would see the file
size growing during the day .... but it was most annoying that I had DONE all
the things being listed at the time but in the end lost a few days data at that
site.

Branden has already replied that this WAS the problem in his case! And that it
looks like the same 'write caching' is present in these battery powered machines
as was present back then?

But you are right that some of the 'protection' software can cause similar
problems, and SteadyState could well 'cause loss of a days data' if one forgets
to turn something off. "Learn how Windows SteadyState can return your computer
and hard disk to its exact condition before the user touched it, simply by
rebooting." may have it's uses, but it can also be abused ;)

--
Lester Caine - G8HFL
-----------------------------
Contact - http://lsces.co.uk/wiki/?page=contact
L.S.Caine Electronic Services - http://lsces.co.uk
EnquirySolve - http://enquirysolve.com/
Model Engineers Digital Workshop - http://medw.co.uk//
Firebird - http://www.firebirdsql.org/index.php