Re: Win11 explorer bug?

"Don Y" <blockedofcourse@foo.invalid> wrote in message news:vjennd$24vi6$1@dont-email.me...

On 12/12/2024 5:47 AM, Carlos E.R. wrote:

On 2024-12-12 12:00, Don Y wrote:

Yeah, I'm REALLY eager to turn on the factory's WiFi interface
for the stove/oven... NOT!

>
There are devices that put the actual interface on the phone, via WiFi. The physical interface has a reduced set of features.

>
Yes.  Via a server located at the manufacturer's facility!
>
So, you have the application layer in the appliance, the network stack in the
appliance, all of the network infrastructure from your AP to the manufacturer's
server, then, back through the phone network, up through the stack in your
phone and, finally, through the app to the display.

I hate this too.
I'm resistant to cameras which bounce off the manufacturer's server, which could be anywhere.

>
Nothing can go wrong, there, right?  <rolls eyes>
>
If I can manage to hang (if not outright CRASH) the appliance using the
FEW controls available to me, how many more wonderful and exciting ways
might it be at risk with all this other fluff involved?
>
Do I *really* need to be able to turn the oven on as I leave work so the
roast has had extra time to cook while I'm busy driving?
>
How might my "blind" actions interact with some activities initiated
by whomever happens to be IN the house (by the appliance) at the time?
>
How many races remain in hiding in the implementation?  (clearly they
didn't test for ALL of these if I can tickle several of them so easily)
>

I'm thinking of a particular heating system with thermostat. You can program the times when the heating turns on automatically
and the temps only via internet. On the thermostat on the wall there is only a manual control that sets the temp for "now", a
knob.

>
This is a false design economy:  "Let's skip the interface on the actual
device in favor of one on some OTHER device."  It invites the two falling
out of sync with each other as there is nothing ensuring updates to one
are also propagated to the other.
>
I'm dicking with UPSs this morning.  In theory, all of them should be
configured identically -- with the exception of specific instance data
(e.g., host name, IP address, SNMP traps, etc.).  I can attempt to verify
this by dumping the configurations (in text format) and doing a line-by-line
compare.
>
    "Gee, how come this UPS has a whole set of settings that the others
    don't?  Same version software..."
>

Oh, and it comes with no manual, no docs.

>
Of course not!  That would be a THIRD thing that would fall out of sync with
the others!
>
The ideal design is the one where you can remove nothing MORE from it.
Yet, we see so many products built on Linux kernels (from which a LOT
can be removed -- including the bugs associated with all that cruft!)
>