[Just J]

It seems clear from the widely varied experiences people are having with notionally identically hardware and software that how the box is used is a critical factor in reliability.

The box has a network port, right? So not have the box record every remote signal received, and every front panel button press detected, in an activity log? Any time one of the menu reset options was selected, this could be sent to D* over the network.

Now, there are lots of things to work out about this process, like how to tie a box to a user (s/n and access card numbers?) and how to tie a reset to a problem (perhaps a web form to submit your data, with known bugs as radio buttons, a text area for new bugs, and a place to put your box S/N and the date/time of reset and, the first time, your connections, devices, and other "environmental" information).

As with Windows, you could make it an "opt in" deal, either case-by-case, or as a system setting - this would protect people's privacy and ensure that D* didn't get buried in data that would never get tied to a particular error condition.

Seems to me that would dramatically improve the specificity, accuracy, and quantity of data available to figure out what conditions are triggering these issues.

 

[wmschultz]

Could you imagine the amount of staff required to look through these files?

Crap Java code induces the behavior that we see.

 

[AlbertZeroK]

Could you imagine the amount of staff required to look through these files?

Crap Java code induces the behavior that we see.

Files like this don't get poored over one by one. A system is set up to capture the files and then queries are built. They see which problems are causing what errors. It's also worth noting that bugs effecting a very few are typically never fixed.

But the bandwidth and support requirements pretty much deter DirecTV from doing this. You're talking about a few programmers to write and manage the system and a server farm. $250k at the least just to get started.

At the end of the day, when they move more programming to MPEG4, we won't have a choicce but the HR20 or DISH (or local cable if you are not in a rural area like I am). So you take what they give you.

 

[PoitNarf]

I fully expect that the beta testers for the development team have HR20s with this capability.

 

[Spanky_Partain]

I fully expect that the beta testers for the development team have HR20s with this capability.

I would have to agree! The netdump feature in Linux is a common way to get this information.

Could it be setup for beta testers like me and you in the outside world?

Yes, but not a likely scenario. Too much to manage!

 

[Spanky_Partain]

There are a couple of other factors to figure in on this scenario...

The netdump feature is for a panic, or PSOD, penguin screen of death, :nono2: not for hangs and lock ups. For that, a log would need to be kept and then sent with some type of if then statement once the machine was up.

Spanky

 

[lguvenoz]

There are a couple of other factors to figure in on this scenario...

The netdump feature is for a panic, or PSOD, penguin screen of death, :nono2: not for hangs and lock ups. For that, a log would need to be kept and then sent with some type of if then statement once the machine was up.

Spanky

Windows tries to do this (not very successfully). When the computer is ungracefully restarted, the OS can detect this and then look through the logs right before the restart to see whether there was some particular event that triggered it. Only occassionally does it truly work well, but it does try.