Hi,

Howard M. Lewis Ship wrote:

>However, I'm loath to do this ... primarily because I'm nervous about giving
>the developer the rope needed to easily hang themself.
>  
>
i think caching on the HTML level is actually useful (you might have 
display logic that is quite complex, or has a relatively large overhead 
but at the same time pretty much static - i know, i do :). In the JSP 
world, i've been using the OpenSymphony cache tags, works fine.. and of 
course, you can hang yourself with it :) but at least it's very easy :)

having HTML caching is _not_ a replacement for various application 
specific caches, it's an additional level.

>More caching problems:  do you cache for a single user, for a single locale,
>etc.
>  
>
the cache key has to be generated by the application developer. also, 
making a runtime decision about _not_ caching for this request is also 
fundemental, as it also remedies the following problem

>Finally, theres' the issue of URLs which may have to be encoded with the
>caller's HttpSession id.
>  
>
not caching in this case is probably a decent solution for most cases. 
how many users have cookies turned off anyway? caching is for the other 
99% :)

>efficiently than Tapestry can quickly generate the appropriate HTML.  If I
>*ever* get a chance to work on Sabertooth again ... it'll have a lot of
>smart caching built in.
>
btw, i've seen your plans for sabertooth. i think you should take a look 
at Hibernate (http://hibernate.sourceforge.net/), seems to be pretty 
close in spirit...

best regards,
    viktor



_______________________________________________________________

Don't miss the 2002 Sprint PCS Application Developer's Conference
August 25-28 in Las Vegas - 
http://devcon.sprintpcs.com/adp/index.cfm?source=osdntextlink

_______________________________________________
Tapestry-developer mailing list
[EMAIL PROTECTED]
https://lists.sourceforge.net/lists/listinfo/tapestry-developer

Reply via email to