[Foundation-l] [Internal-l] Relocation Announcement

[Anthony]

On 9/25/07, Stephen Bain <stephen.bain at gmail.com> wrote:
> On 9/25/07, Anthony <wikimail at inbox.org> wrote:
> >
> > Hitting Amsterdam to read while
> > logged in and checking in with Tampa when you actually hit "submit" on
> > an article, on the other hand, isn't.
>
> Don't forget that reading works differently for logged-in users too.
> Even if you try caching only parts of a page (which involves all sorts
> of challenges on its own), you're still going to confront the problem
> that the site looks different for different users depending on user
> preferences and so forth.
>
> To give one example, I have the preference for numbered headings
> turned on. Sure, you could cache one version with and one without
> numbering, and then check that setting and serve one or the other.

Or you could cache the wiki source and generate the page dynamically
based on the cached settings.

> Plus is it really going to be more efficient trying to paste together
> dozens of different cached components than it is to just render the
> page anew?
>
I don't know about efficient, but it'd certainly be faster, probably
by at least a couple orders of magnitude.  It doesn't take hundreds of
milliseconds to parse wiki source.

I'm not volunteering to implement it, and I'm not complaining that it
hasn't been implemented.  I'm just objecting to the fact that some
people on this list seem to be claiming it's impossible or infeasible.