Image Caching, when to update?

[Sean Gillies]

On Aug 4, 2005, at 9:00 AM, Arnulf Christl wrote:

> Kyle Mulka wrote:
> [...]
>> I looked into squid and realized that thing it needs to know is when
>> the content changes so that it can refresh the cache. From what I've
>> heard so far, it doesn't sound like mapserver has any clue as to
>> when the map it serves might change. It seems that it would be
>> pretty easy for mapserver to simply determine if a map would change
>> based on the timestamp on the shapefiles its serving and either
>> the timestamp or the content of the map file.
>
> Kyle Mulka wrote:
>> Would it be good to move this discussion to mapserver-dev?
>
> Hmmm, i'd say leave them out front, they got enough on their hands 
> doing
> a great job serving maps (not that i would repeat myself... :-)
>
>>> By the looks of what i have seen in your demo you do a great job 
>>> already.
>>
>> :-)
>>
>> I'm using the PEAR Cache_Lite package for PHP right now. Its simple 
>> to use.
>> Only takes about 10 lines of code.
>
> [...]
>
>> Exactly. This is why we need at least something in mapserver that 
>> knows
>> about caching. Mapserver doesn't actually have to do the caching, but 
>> should
>> know when caching would be appropriate.
>
> I don't think any WMS service really knows a lot about what kind or
> actuality of data it actually serves (someone enlighten us?).
>
> Wouldn't this be something that metadata could be used for? Not that
> anybody would care to provide for good metadata...
>
> There are not too many thoughts about misusing metadata for this kind 
> of
> job as far as i know but basically it would fit in nicely?! This said 
> we
> might haul this discussion over to the OpenSDI clan instead of bullying
> the mapserver dev guys, but maybe we should just let them decide for
> themselves.
>
> Best,
>

IMO, MapServer should remain focussed on rendering map images, and 
caching should be left to other specialized applications or modules. 
This is the way the web works. See

   http://www.w3.org/DesignIssues/Principles.html

for arguments why, from a smarter and better writer than myself.

cheers,
Sean

--
Sean Gillies
sgillies at frii dot com
http://zcologia.com