🚀 go-pugleaf

Paul wrote:
> Centre Parting wrote:
>> Paul wrote:
>>> Franc Zabkar wrote:
>>>> On Mon, 6 Apr 2009 11:24:53 +0100, "Centre Parting"
>>>> <nokia.account@ntlworld.com> put finger to keyboard and composed:
>>>>
>>>>> The Google study presumably polled a lot of PC users about HDD
>>>>> failures.
>>>> The Google study concentrated on Google's own huge installed base
>>>> of hard drives spanning several manufacturers, models, and
>>>> generations. Nobody was polled.
>>>>
>>>> Failure Trends in a Large Disk Drive Population:
>>>> http://research.google.com/archive/disk_failures.pdf
>>>>
>>>> - Franc Zabkar
>>> What struck me about their study, is the word "humidity"
>>> doesn't appear anywhere in the report. And humidity is a
>>> factor in disk reliability.
>>
>> Not heard that touted before.
>>
>> You sure you're not just confused about the air conditioning issue ?
>> Reasoning that since big installations are kept in a/c'd rooms (dry
>> atmosphere), non-a/c'd (relatively humid atmosphere) installations
>> must be less reliable ?
>>
>> Big installations aren't air-conditioned because of humidity.
>> They're a/c'd because that's the only way TO cool a room.
>> Humidity is purely incidental.
>>
>> That's some pretty weird-ass reasoning you've got.
>>
>>>    Paul
>>
>
> I judge studies by their attention to detail.
>
> Without knowing what combination of temperature and humidity
> is present in each facility, there is no way to judge whether
> the missing factor is important or not.
>
> If you download an OEM spec for a disk drive, there is
> a temperature/humidity graph in it.

Yes, but that's for un-a/c'd conditions (under which, the majority will be 
used).
In a/c'd conditions, it become irrelevant because the temp is kept low 
enough that humidity is of no consequence to the cooling capacity of air.


It documents the allowed
> values of temperature and humidity for a disk drive. Which
> means in any study, humidity should be documented (because the
> disk manufacturer acknowledges it is a factor). Or at least the
> word should appear in the study at least once, as in
> "we chose to ignore humidity".
>
>    Paul