For The 200 Volunteers And Beyond

[Phawx]

What if there was a transparent layer that was always running keeping statistics? This way if someone has a particular problem with something in the Pandora OP could narrow it down much easier. Obviously, this would be great for the first 200, but I think all 3800 of us using our Pandora's in different ways would reveal quite a lot pertaining to the durability of the Pandora. Not only that, you wouldn't have to pay a company to run synthetic durability tests and OP would actually have valid real world data to glean off of.

All the variables that would need to be kept track of (that I can think of):

- How many times the lid was opened/closed.
***Durability of hinges***

- How long the display has been on for it's lifetime along with brightness level.
***Determine lifespan of display***

- How long the unit was running off of a battery (-sleep) before being recharged.
***Gather realistic battery lifespan and determine the average time people get on Panda's***

- How many times that battery was charged
***Realistic amount of cycles to get out of Panda battery***

- How many times each individual button was pressed or moved.
***Determine the durability of D-pad ABYX, analog sticks and QWERTY buttons. Also, (I realize this is a hot topic) you could generate a heat-map showing which buttons are used more often and see if it is necessary to redesign the keyboard layout. (I know this is a touchy subject. But all of us shouldn't go into this thinking that the perfect design will come out in it's first "physical" build)***

- How many times an SD card was inserted and ejected
***Determine durability of SD slots***

- Record throughput periodically from Wifi, BT and SD
***Verify that all users are sharing the same performance and to help examine IO***

- How often CPU/GPU/DSP/RAM are at full load or heavy usage and what frequencies they are running at.
***Examine lifespan of SoC and the correlation to those that OC or not***

- Record when temperatures fall below or above a specific amount and the duration.
***Having a timestamp along with whatever temperature the device was at (when at extremes) could help determine a problem. Obviously it would have to be very cold to fall below a certain point, but it might help explain the case cracking or something.***

- How many times the NAND is flashed.
***Durability of the NAND***

To bad we don't have an accelerometer or we would be able to detect falls as well.

So. What do you guys think? Could this type of transparent layer run and record the data without having a serious performance penalty? Obviously, for privacy reasons this would need to be an Opt-in case, aside from the volunteers where this data could be helpful.

Is there anything else that should be added or changed to this list? Do you think this is even necessary? My thoughts behind this are that if we want to help OP create a REALLY solid product for the third and fourth batchers and have very accurate data to reflect from, we could make the future Pandora's a well tested device. Helping new comers feel comfortable buying into the Pandora because they have real data to look at instead of synthetic tests.

What do you guys think?

 

[Chip]

Most of those conditions could be tracked, but I don't think any of them would be particularly useful in the short testing period we have available. If we had months instead of weeks, and we wanted to set up proper HALT testing conditions, then that sort of data would be quite valuable. However, under normal usage over the course of 1-2 weeks, it's not going to give us enough data to be worth the system overhead.

Farther out, for the 3rd batch and beyond, the sort of problems this tracking would find will be self-evident. If there is a weakness in the design, we'll know because people will be discussing it here, or more likely sending units in for repair or replacement.

For instance, we don't need to know the exact number of times the lid can be opened before it fails (this would take years of normal use, not months). All we do need to know is if it fails before years of use or not. If it does, then I'm sure people will let us know about it right away.

If somebody wanted to create logging software like you describe, and others volunteered to run it, that data could be quite helpful over the years. I'm not going to say "Don't do it." But the best way for us to determine how durable the design is in the near future is to just let people use it. If something is easily breakable, somebody will figure out a way to break it. If a lot of people break something, then we'll look into how that something can be made less breakable for future units.

 

[Sphinxter]

I think the risk it might create what is known as a single point of failure would not be worth the data it could collect.

 

[DavidBowman]

I think peanut butter and jelly sandwiches are delicious. There. I said it.

 

[Username]

While it might not be a very useful feature, that doesn't mean it wouldn't be pretty cool. Maybe a stats page that shows "number of times screen opened/closed a day" and that sort of thing. It could tell people how the majority are being used.

 

[Consequence_9]

While it might not be a very useful feature, that doesn't mean it wouldn't be pretty cool. Maybe a stats page that shows "number of times screen opened/closed a day" and that sort of thing. It could tell people how the majority are being used.

That'd make a nice widget =]

 

[cameroncros]

And some kind of uptime /average uptime feature would be great for bragging rights.

 

[DavidBowman]

That's a good point. I have to reset my ipod all the time because the applications on it freeze it or don't release RAM properly. I'll expect no less than nearly flawless operation out of my pandora.

 

[lulzfish]

And some kind of uptime /average uptime feature would be great for bragging rights.

CODE
xterm -e "uptime; read"


>_>

 

[dadioflex]

If you start to publicly log statistics about failure rates of any aspect of the device, then sure as shooting, someone will misuse those statistics to make it look bad.

 

[cools]

I'd be happy to subject mine early to the kind of abuse my handhelds get - as long as I can get a replacement when they've fixed it up (if it fails).

Otherwise it'll be a suck it and see