[SensorNetArch] Non-outline

[Scott Shenker]

Upon re-reading the proposal, I have come to the conclusion that:

(a) we don't have 14 pages of content behind "why we need a sensornet 
architecture."  In fact, we merely assert that the single factor 
holding back research is the lack of an architecture.

(b) most of our material is reviewing various aspects of a 
proto-architecture.  While identifying SP as the narrow waist is novel, 
the rest of the material is mostly review or speculation, neither of 
which will cut it at Sigcomm.

(c)  I think we might be better off heading towards HotOS.  I don't 
know HotOS very well, and Hari is the only person on the PC who has 
ever seen a mote, so perhaps we would be doomed there.  But we could 
whip out a five pager that:

1)  Makes the controversial assertion that the lack of an architecture 
is holding the field back

2)  The narrow waist belongs at SP

3)  Sketches out the rest, very briefly.

While (3) is old news, (1) and (2) might be considered more novel.

(d)  The HotOS deadline is Tuesday, midnight.  That makes it a very 
doable task, with limited downside.  Besides, none of the students were 
planning to sleep between now and then.

(e)  We talk on Monday (1:00-2:00).  If HotOS has any attraction, we 
should probably start writing before that.  So, what do people think 
about HotOS?