From: Jason McMullan <jmcc@ontv.com>
  To  : evstack@ontv.com
  Date: 29 Nov 1998 02:36:31 GMT

Re: architectural overview

Brad Midgley <brad@pht.com> wrote with confidence:
> i wish i had talked to you at the expo... argh.

  I should be at next year's....
 
> i've been looking at this and the other stuff i have from you and i think
> i have a good picture of how things work. a few questions:
> 
> is there an fbdev scroller? has it been started?

  No. With some family troubles in the way, and a LOT of internal
redesign work, I haven't had much time to work on that yet.

> when an app writes to /dev/graphn, does the eventual interaction with the
> display circumvent the scroller? 
 
  Yes. Use of /dev/graph imlicitly `unmaps' the scroller, ie a
terminal emulator can `draw' to it, but it doesn't display until
the /dev/graph app releases /dev/graph. (or SAK is pressed)

> your diagram shows the tty device (for reading) appearing after the tty
> emulation layer. is that correct? why isn't it positioned just after the
> keymap filter?

	Because I hadn't redsigned it yet. ;^) Seriously, that is in
the new design, and I should have a (small) paper on that out by
the end of December, along with my next major code release (which
should work on PowerPC, Intel, and (god willing and my Multi boots) 
Alpha.)

> can writes to /dev/eventNN cause vt switches or the reboot sequence? (ie,
> at what point do the events get inserted back into the stream?)

	Yep, they sure can - if they're above the keymapper.
Which is why only root can create them, and only root can tell 
/dev/event where in the stream to attach. (they aren't tied to a 
specific stream point - you (root) can attach them to any filter, 
or between any two filters.)

	Ie, If I wanted speech->text, I would write a user app
to do so, ask root to create a /dev/event attached right below
the keymapper, and inject my text. 

-- 
Jason McMullan - Linux - GGI - http://pepsi.visus.com/~jmcc
  I can't wait until the point where free Linux CD's or other paraphenalia is 
included in cereal boxes.  That truly will be the cherry ontop of our madness!
  -- Mike A. Harris <mharris@ican.net> on linux-kernel@vger.rutgers.edu