NYCPHP Meetup

[Rob Marscher]

On Dec 14, 2007, at 1:41 PM, David Krings wrote:
> John Campbell wrote:
>>> Is there anything drastically flawed with my approach?
>>>
>> Since you understand the pitfalls of your approach and choose to do  
>> it
>> that way it is not a problem.  You are also using sessions to prevent
>> repeated hits to the database for "performance reasons".  Don't do
>> that. It typically makes performance worse because session data must
>> be serialized and written to disk on every request even if the data
>> doesn't change.
>
> So you are saying that a database access is faster than a file  
> access? I find that hard to believe, but do not have any evidence of  
> the contrary.

I would guess it depends on a number of factors -- like whether or not  
the session file is in the OS filesystem cache, what type of hard  
drive you are using, etc.  I think the point John was making is that  
if the data doesn't change you might be able to recognize that there  
was no change and not write the session data at the end of the  
request.  With php's default session handler, it's going to write to  
the session file at the end of every request.  If you're not working  
on a medium to high volume traffic site, the performance difference  
probably isn't a big deal.