Re: planning the next steps

[Seth Teller <teller@lcs.mit.edu>]

michael, peter, patrick, vitaly, [& olivier],

since you're all back for the start of the term, i
wanted to re-send this to remind everyone of the
development directions for BMG.  i'd like the four
of you to meet regularly, at least weekly, while
i'm away, to make progress on these fronts.  if
possible, i'd like to incorporate the stata center
plans (2D floorplans only?) into BMG so we can do
route-finding inside stata.

please think about this and write back to the 
group with the pieces you wish to concentrate on
over the coming term.  i'll write more over the
coming week with suggestions for coordinating all
of our efforts.

seth.


Seth Teller wrote:
> 
> bmg'ers,
> 
> here are several directions in which we can take BMG and
> walkthrough next.  these ideas span the spectrum from well-
> understood implementation issues to full research (i.e.,
> things we don't yet know how to do).
> 
> summarizing, the directions [with suggested student teams
> in brackets] are:
> 
>    o merge space-type and adjacency extraction into existing
>      BMG pipeline to form single source for space/route data
>      [michael, patrick]
> 
>    o merge spaces and adjacencies from basemap with interior
>      info, to enable inter-building route generation.  note
>      that this requires solving a naming problem for outdoor
>      spaces.  [vitaly, patrick]
> 
>    o incorporate topo (height) information into whole-campus
>      map to adjust terrain height and building floor placement.
>      [vitaly, michael]
> 
>    o resuscitate "stairmaster" utility and generate interior
>      vertical connectors (ramps, stairwells, chair lifts,
>      elevators).  [michael, vitaly]
> 
>    o generate vertical connecting elements (ramps & stairwells)
>      at building perimeters, i.e., to connect surroundings with
>      entryways.  [vitaly, michael]
> 
>    o generate purely exterior vertical connectors (e.g., ramp
>      and stairs at entrance to ames courtyard; stairs into
>      medical center from eastern side).  [vitaly]
> 
>    o extend visibility compilation and query capability to
>      whole campus case.  phase one:  existing algorithms,
>      axial models, building rotations suppressed to produce
>      axial geometry.  phase two:  new algorithms, with
>      locally axial, globally non-axial geometry.  [peter]
> 
>    o incorporate metric texture samples (photographs of, say,
>      one-meter square texture patches from campus surfaces)
>      into generated model in a principled way.  note, this
>      raises another hard naming/geo-location problem.
>      [olivier, vitaly, michael]
> 
>    o resuscitate furnpop utility, and merge extracted space-type
>      and furnpop invocation to furnish (most) interior spaces
>      [students:  ?]
> 
>    o extend furnpop to populate exterior spaces, with trees,
>      hydrants, benches, newspaper boxes, mailboxes, telephone
>      poles and boxes, light posts, etc.  [students:  ?]
> 
> i'd like each of you to think about these and your own ideas
> for next steps, your estimate of their scope/difficulty and ideas
> for how we can pursue them, and your own involvement over the summer
> and fall.  also, give some thought to a compelling app (perhaps
> several) that we can deploy in the same period, to give the
> rest of the community some understanding of what we're up to.
> 
> let's discuss at tomorrow's BMG meeting, T 415pm in the G lab
> as usual.
> 
> prof. t.