binary protocol time representation

[Randy Wigginton]

Limiting memcache to dealing with "whole" seconds seems like a good  
choice.  I don't see much use for "expire this item in 4.025 seconds"

On Jul 12, 2007, at 5:52 AM, Roberto Spadim wrote:

> don't forget that for high speed systems (more than 1Hz) time must  
> have fraction of seconds, like microtime() in php with 6 digits  
> after seconds
>
> Dan Farina escreveu:
>> On Thu, 2007-07-12 at 14:16 +0200, a. wrote:
>>
>>> Ticks are easy in .NEt, but I'm much more concerned about the   
>>> "meaning" of that specific time.
>>>
>>> If my server runs DST but one of the clients is on CET, and the   
>>> client sends 2007-07-12 12:00 AM, what does it mean? 12:00 in DST  
>>> or  CET?
>>>
>>> We either should declare that absolute expiration should treated  
>>> as  UTC, or get rid of it and use relative expirations.
>>>
>>>
>>
>> Yes, either of these makes me more happy than the current situation.
>>
>> I generally don't like absolute times unless absolutely required,
>> though. Relative time is robust to all sorts of things for free like
>> massively mismatched clocks and, as mentioned, time zones.
>>
>> df
>>
>>
>> Esta mensagem foi verificada pelo E-mail Protegido Terra.
>> Scan engine: McAfee VirusScan / Atualizado em 11/07/2007 / Versão:  
>> 5.1.00/5072
>> Proteja o seu e-mail Terra: http://mail.terra.com.br/
>>
>>
>>
>