Save stacks periodically to avoid crashes? [OT]

[Jim Ault]

Thanks for clearing up part of the mystery for me.. I thought it was some
sort of cubic spline and the volume under that surface.  Glad to know it is
only a random number.

One of my friends still believes it is an inverse proportion to the
importance of the data and the minutes until the deadline...
freq = importance/time = maximizing relative loss

Jim Ault
Las Vegas


On 3/25/06 12:32 PM, "Dennis Brown" <see3d at writeme.com> wrote:

> After a lot of clever hacking, I have found that I can almost always
> find the counter in any application that counts the number of
> keystrokes since the last save.  Once the counter crosses a certain
> threshold, a random number is invoked in each new keystroke.  If the
> random number matches the keystroke count, then a crash is provoked.
> 
> It is all part of a secret programmers guild directive designed to
> help users learn to save and back up their work regularly.  It is
> expected that after 11 generations, the urge to save will become a
> reflex built into the human genome through natural selection.  Those
> who do not learn to save will become failures in life --unable to
> attract a mate. Their "bad" genes will then vanish from the species.
> 
> Dennis ;-)
> 
> On Mar 24, 2006, at 11:49 AM, Jon Seymour wrote:
> 
>> Hi, I've been using Rev for about a year. I'm sure it won't shock
>> most of you to hear that periodically Rev just seems tired and
>> crashes. Now I am sure that coding glitches are sometimes at fault,
>> but generally speaking I think Rev (esp. 2.7) has stability issues.
>> Here's the thing, though: it seems that if I am saving the stack
>> periodically, which I would tend to do to avoid losing data in a
>> crash, the program actually crashes less. It's as if saving has
>> some benefit to memory management or who-knows-what-else in the
>> engine. It's like a "refresh" function. Has anyone else observed
>> this? Is there a rationale? Would it be smart to have a commercial
>> application save its stacks regularly, not only to store user
>> changes, but simply to confer stability?