I'm reading all these possibilities for prefetching in freenet, but I
guess I'm still not really clear what is being gained over the
prefetching built into mozilla. (You are using mozilla, right?)
ok, I'll grant you the first page. Mozilla won't know where to start
until you point it at something. but active-links aside, the gateway
page has always been the one thing thats very quick to load. Pushing
the prefetching work out to the client means that it's in the
customers hands, and is adjusted/customized to their own preferences
and bandwidth. (Possibly before they even download freenet.) It also
means not having to re-invent the wheel, detect modems, or deal with
weird interactions of having multiple prefetching mechinisms in place
at the same time. (And end users WILL use mozilla's prefetching,
whether intentionally, or because someone makes it a default in a
future release.) If you're really worried that your bookmarks will
fall out of the system and you also really don't want to point you
browser at that page, then there are other ways to go about that which
don't require extra code in the freenet node. So I'm wondering if
there is really a big enough advantage to building this in?
In all fairness, I have precaching turned off on my mozilla, so it may
have shortcomings that can be addressed by a freenet specific
implementation, I'm just asking what those would be.

-michael

"Colin" <[EMAIL PROTECTED]> writes:
> Actually, a nice feature might be the ability to specify a number of
> layers in to spider/crawl/preload. Ie, it will load all of your
> (user-settable) bookmarks, then X links deep.
> There would need to be an option to disable pre-fetching the bookmarked
> pages anyway, so this could just be an extension (levels, rather than a
> binary yes/no)
>
> Colin
_______________________________________________
devl mailing list
[EMAIL PROTECTED]
http://hawk.freenetproject.org:8080/cgi-bin/mailman/listinfo/devl

Reply via email to