Submitted by alex pang on August 10, 1997 at 14:22:31
first, allow me to applaud the authors of this document.
very well written and understandable even for someone not
in the area. however, i don't think i fully understand
everything. let's use the task of soliciting/integrating
comments for the "methodology for eval. of collab. system"
document as an example.
so, this would be a type 4 task. or, it could also be
a type 5 task depending on participants' background,
ulterior motives, etc. here, i'am also taking the
liberty of assuming that either of these task types
can accommodate introduction of variation to choices.
the point being, these task types can sometimes be
fuzzy, hence, the resulting evaluation metrics/measures
need to be selected properly.
let's say that a splinter discussion group is somehow
formed which later comes back with a set of comments/
suggestions to this document. how would such actions
be captured by the scenario description (example on pg 16)?
i.e. does the scenario description allow for parallel sub-group
activities and/or non-sequential events?
----
for task type 6b (bargaining task), where does the role
of mediation fit in? would that be at the service layer?
---
for task type 9 (non-mcgrath), where does the role of
incentives fit in? e.g. professor's use of incentives
and other mechanism to improve info dissemination;
e.g. incentives and other mechanisms to improve fairness
in floor control, ...
---
pg. 30, support for object types:
the list seem to be limited to image, sound, video/animation.
a. what are computational images?
b. what about other forms of object types?
e.g. text, raw data, etc.
---
pg 34, recording.
assume that also implies ability to playback recording
---
one other general comment: an added advantage of scenarios seem
to be their general nature -- i.e. unlike benchmarks (e.g. linpack)
developers cannot tweak their system to excel in any one particular
strength. the downside is the tradeoffs of general do-it-all
systems against specialized system. for example, in visualization,
the "quality of the visualization" really depends on how it is
going to be used by the person. but, i think there's sufficient
cautionary text to this regard that's already in the present
document.
sorry for the rambling nature. regards,
alex pang