Re: How to slow down an AS

["Raul A. Jager W." <raul@xxxxxxxxxx>]

I did not blame the OS, I recognized it was my mistake.  But, I think 
that if possible, the OS should protect me from my mistakes.

I fully realize that protecting me from my mistakes requires disabling 
some "nice features", the main question is: Is the feature worth the risk? 
Not an easy one, implies: how usefull is the feature?  How often people 
commit the mistake? how bad is the consecuence?

The circular reference is not usefull, is it hard to test?,  Can IBM set 
a limit to the references? like 10 levels deep?

I don't think cascading DDMs is the best way to reach a file in a 
computer that is not in the same net, having the DDM point to the 
Physical file and seting the route in the IP will be faster and more clear.
_______________________________________________________________________________________
Paul Morgan wrote:

> Raul,
>
> (flame on)
>
> You could also have had a DDM file on PRODUCTION pointing back to your DDM
> file on DEVELOPMENT.  This would also be a circular reference but would be
> more difficult to detect.  Why slow down a DDM connection even further that
> it already is to check for circular references?
>
> Don't blame the OS for your own mistake.  This behavior of DDM files is by
> design.  You might not have visibility to another Iseries on the network but
> you can point the DDM file to another Iseries with a DDM file which does
> point to the Iseries you can't see.  .  Nice feature which I wouldn't want
> to see removed just so you won't have stupid recursion problems.
>
> (flame off)
>
> Paul
>
>