The bit is currently meaningless, we should either admit this fact and
recommend not using it or fix the problem. To keep it as is is
essentially the former.
On Wed, 15 Apr 1998, Tom Hastings wrote:
> It looks like a number of products do set the broken bit as well as the
> unavailable bit, so we shouldn't change the spec for the top 25
> out from under them.
> So I would think that we should add a note to application writers
> that the broken bit doesn't necessarily mean that the device is broken.
> At 08:56 04/15/1998 PDT, Ron Bergman wrote:
> >The Dataproducts implementation conforms to the Top-25 alert definitions
> >reporting "Unavailable because Broken".
> >I would, however, prefer that the HR MIB and the Printer be revised to
> >allow more flexibility in these conditions to allow the true condition of
> >the printer to be reported.
> But if we make such a change, how is an application to know whether a device
> that has set the broken bit is really broken or is one of the existing
> devices that sets the broken bit because the top 25 says to set the bit?
> Unfortunately, I don't think we can use the broken bit to mean broken
> without impacting existing implementations.
> > Ron Bergman
> > Dataproducts Corp.