Re: best paging strategies for large datasets?

>> First u count(*) the rows and select a requested page
>> returning to a client the count result bundled "with a page of rows"
>>
>> (1) client renders the acquired rows
>> (2)__memorize__ what part of the data he just got
>> (3) and stores the count result to calculate "the pager div"
>>
>> all the subsequent clicks on "the pager div" should not immediately
>> generate requests and decides if the request is needed.
>
> Yes, rendering the results throught ajax is a good idea, but one has to
> be careful not to expose one's LIMIT and OFFSET to the client, but only
> the "page" number. Or else the client could query the whole data set. A
> lot of "professional" web site have that hole.
>

this is not a hole, it is only a matter of aesthetic