web apps with large volumes of bidirectional http traffic

[Nomen Nescio]

Ryan Lackey writes:

> I need to find some relatively widely deployed applications which have
> frequent user interactions (rapid clicking on links, from as large a
> population of links as possible, and also form filling and such).
>
> (it should be pretty obvious what this is for)

It's not, really.  Maybe some kind of cover traffic for an underlying
hidden data stream?

Or maybe you will create a client which mimics a human person, and
sends information by choosing what to click on?  Sounds like a pretty
low-bandwidth channel.


> I'd like:
>
> 0) *rapid*/frequent user interactions; fast clicking on things (like every
> second, no more than 5 seconds)
>
> 1) "sticky"...long interactions with a given site (on the order of hours)
> (also all links need to be under the same url/same server)

These two don't make much sense together: someone is going to sit
there, interacting with a server for hours, clicking every second?
That sounds more like a job than something people would do voluntarily.
You're either going to need to accept a much lower click rate, or else
a much shorter connection time.

Okay, maybe one of the online fantasy games services like EverQuest?
I hear people get addicted to those and spend hours on them, but I
think they do a lot of typing rather than mouse clicks.  Or one of the
shoot-em-ups?  Those probably use the mouse a lot more.  They aren't
really "web applications" though.  Does it really have to be a web app?

> So far, the best ideas:
> 1) Porn
> 2) Mailing lists with lots of internal links (next, reply, etc.)
> 3) Sites with search engines with lots of linked data (encyclopedia,
> etc.)
> 4) html games (or flash, maybe) -- either imagemaps, or just tables,
> things like chess, or puzzles, or whatever

Most of those won't satisfy your click rate and connect time goals.

> I'd definitely appreciate any suggestions on possible web apps which
> meet these criteria; reply to lists or ryan at venona.com.
>
> I'll post when it's ready.

Maybe you could give some more clues...

---------------------------------------------------------------------
The Cryptography Mailing List
Unsubscribe by sending "unsubscribe cryptography" to majordomo at metzdowd.com