- From: C H <craighubleyca@yahoo.com>
- Date: Thu, 21 Jun 2007 16:38:43 -0700 (PDT)
- To: paola.dimaio@gmail.com, Gavin Treadgold <gt@kestrel.co.nz>
- Cc: W3C Ontology List Disaster Management <public-disaster-management-ont@w3.org>
(I request that we actually update our subject lines,
using "[was]" to clarify which thread they came from)
----
SUMMARY (of this post only)
The constraints on an ontology to be used for machine
reasoning are very different than those on definitions
used by human beings with specific training. What the
UN and legal authorities have done is useful to start
but is not disjunctive/operational enough to define a
list of processes, activities and transactions that a
W3 ontology enables. More refinement will be needed,
both from stakeholders (as Paola suggests) but also to
determine how clusters emerge and how many
perspectives on one problem are efficiently combined
to ensure all are consulted before major decisions.
In any case we need a plan for how to review terms.
I suggest the following refined clusters to organize
the (excellent, thorough) list of terms Gavin
published
at
http://www.plan.net.nz/wiki/index.php/Category:Terms
* camp coordination and management (arrivals and
departures, internal security and inventory)
* recovery (focusing on the early phases but with
clarity on when resilience processes take over)
* shelter (emergency housing and evacuations)
* grid (emergency telecommunications and power and
the specific protocols and skills to keep 'em
up)
* health (including aid personnel, remote experts)
* logistics (including GPS, movement of aid/help)
* nutrition (including feeding of aid personnel)
* protection (including credentials, evacuations)
* water and sanitation (including aid personnel)
* resilience (all auditing and risk measurements
that describe the long-term situation
developing)
I elaborate them somewhat below, as a guess at their
scope, and suggest that each of the terms Gavin has
identified could fit under one (or at most three) of
them. Even if this is only an ad hoc means of review.
Each cluster seems to be a particular profession or a
perspective that enables applying particular skills or
techniques.
In the long run, clusters could be identified simply
by noting who gets involved in what type of decision.
Copyright issues should be minimal or nil if we all
agree that our contributions are to be released under
Creative Commons attribution-sharealike-noncommercial
license at some future point, and withheld from those
who do not accept these release terms. A commercial
user may use the material simply for remitting any and
all patent, trademark, copyright and domain name right
they may have in terms essential to the ontology's
use.
Until this is worked out, we own our own
contributions.
------
--- paola.dimaio@gmail.com wrote:
> to analyse the terms one by one, or ten by ten, say,
> and see if we want to use them as they are given
> or amend them for adoption in this group
I don't see what logical coherence "tens" have that
"clusters" do not. The UN Office for the Coordination
of Humanitarian Affairs has given us a breakdown of
how to classify processes, and I think we should use
it (but please, you MUST use lowercase in mediawiki,
any other convention inhibits linking from pages)
* camp coordination and management
* early recovery
* emergency shelter
* emergency telecommunications
* health
* logistics
* nutrition
* protection
* water and sanitation
Under these clusters we should/must apply a
process/activity/transaction pattern structure such as
OASIS uses. That is, each cluster provides a
perspective on a set of processes. Each activity is
part of at least one process. Each transaction is
part of at least one activity and its success or
failure is described only in terms of the activities
that it enables.
> The purpose of this evaluation is to allow a wider
> community, with different perspectives to
> make a contribution to the terminology, and helps us
> evolve, and refine the vocabulary if necessary
For that we must continue to rely on public wikis, not
a mailing list. The vocabulary does need refining but
not for social acceptance, but rather, for rigour.
The terminology Gavin has compiled is suitable only
for directing HUMANS who are performing tasks they are
trained in, NOT suitable for machine inference. There
is really no chance at all that the definitions will
be usable in a W3 ontology in the form in which they
are received from legal and organizational sources
EXCEPT those that describe mere data items that are
never going to be reasoned about.
A P-code, for instance, may be described on a map to a
human, but must be precisely delineated as GPC co-ords
for a machine so that a simple function can return the
P-code a given coordinate is in, and the distance from
the centre of the P-code or from a border with
another.
Getting to these more precise definitions will not be
easy. I agree with Paola it should be done slowly but
there's really no reason not to use Gavin's wiki (if
its accomodating of say a CC-by-sa or CC-by-nc-sa type
license) to work them out (I'd suggest CC-by-nc-sa and
any commercial organization that agrees to surrender
its rights in the ontology and certain derived works
receives a commercial license to use it, others
don't).
That solves the copyright problem: Refined precise
versions of the terms are available for use to any
commercial organization that remits its own
improvements under a CC-by-nc-sa license and accepts a
commercial license to use them freely as long as any
improvements are remitted and other rights (such as
patent rights or trademark in any term incorporated in
the ontology) are remitted to W3 in perpetuity. (If
not for these additional rights and abuses we could
use CC-by-sa or GFDL).
> Secondly, it would help us make decisions regard the
> copyright, as you say
> there might be problems there - only if we decide to
> adopt that given term
> precisely as it is given, we need to worry about
> permission - if we come up
> with our own concoction, I dont think we need to
> worry although we may want
> to make our references public
I'd say, don't worry about it, the non-commercial
organizations should be sharing their work under
CC-by-nc-sa anyway, and the commercial ones can only
adopt W3 works under reasonable and stable licensing
terms.
If there's any reason for concern or anyone knows of a
non-commercial organization that actually sues over a
re-use of a definition or term, then everyone involved
could sign an agreement that they INTEND to remit the
work under CC-by-nc-sa at a future date but meanwhile
have agreed to WITHHOLD their work from parties that
claim commercial rights in definitions essential to
disaster management, until such time as they agree to
participate in a common licensing regime / consortium.
That way, holdouts will see this rising tide of public
work that they are explicitly forbidden from re-using
because of their legal stance, and will therefore come
under pressure to change that stance.
Commercial organizations are harder to pressure, but,
if the material is CC-by-nc-sa, then the only way they
can use it is to agree to the W3 consortium's terms.
> if we agree roughly with this process, we could
> selecting the first batch
> would that be an alphabetical exercise, or a
> logical/semantic exercise?
I suggest the latter, and further suggest that we need
a slight expansion of scope, and a "tenth cluster".
The US military, in my opinion correctly, has lumped
"telecommunications" and the power that they require
into a general term "grid". I'd like to replace the
cluster "emergency telecommunications" with "grid" so
that it implicitly includes the power required to run
the telecom system and all the other powered devices.
I'd also like to ensure that management and auditing
concerns don't get dumped into "logistics" which must
be reserved for actual movements of actual goods and
persons via transport networks. It's hard to know
where a term like "risk management" fits.
I think the safest term is "resilience" because this
implies attempts to prevent disasters and to learn
from mistakes, and an overall risk reduction process,
and would include training in the terminology or the
protocols required to minimize damage in the future.
It implies handing off control to locals and leaving
as much as possible in their hands.
Finally, I'd advocate a few changes that accomodate
resilience thinking, like removing "emergency" from
"shelter" as some of these things end up in use for a
very long time, and resilience advises finding ways to
anticipate long-term adjustments and prepare for them.
That leaves these ten clusters, with some guesses as
to what kinds of processes might be "clustered" under
each one; These could be compared to both UN
Inter-Agency Standing Committee and say NZ and US
definitions
* camp coordination and management (arrivals and
departures, internal security and inventory)
- once something is accepted into the camp's
manifest or inventory it's no longer part
of the other clusters? are camps
effectively
groups of cooperating households under some
temporary authority, or some semblance of a
village authority? with some legal status?
- lifecycle of camp from initial selection of
location (from many) to final shutdown and
passing-off of records to local resilience
- communicating negotiated agreements and any
complaints or challenges about other
aspects
of relief/recovery/reconstruction affecting
camp residents, acting as residents
advocate
(e.g. if rebuilding compromises their lands
or in case a rival group enroaches on them)
- or is this part of protection or
recovery?
* recovery (focusing on the early phases but with
clarity on when resilience processes take over)
- identifying damaged infrastructure which is
possible to save with early intervention,
and prioritizing this to minimize later
work
- identifying lost human capital assets and a
gap in authority or expertise that results,
and possibly mitigating efforts like school
building or recruiting foreign volunteers?
- initiating peacemaking processes between
any
groups in conflict during the disaster -
not
very easy to differentiate from protection?
* shelter (emergency housing and evacuations)
- ensuring everyone has cover over their head
to sleep, protection from insects/moisture/
cold/sun, and storage for key possessions
- identifying long term housing that can be
occupied and possibly later purchased, to
serve quarantine, isolation or services to
family during extended hospitalizations -
or was this called "emergency shelter" to
specifically exclude any such arrangements?
- arranging orderly movements among shelters
and maximum compatibility of those sharing
- or is this considered a camp concern or
even protection (abuses inside the tent)?
* grid (emergency telecommunications and power and
the specific protocols and skills to keep 'em
up)
- hard to differentiate this from backup
power
- Internet connectivity between aid personnel
and mission-critical systems local or
global
- credentials and identities of personnel who
perform authentication and credentials
tasks
- electric power backups and alternate means
to pass life-critical signals if these fail
- any non-telecom means to move signals to
the
telecom-capable (carrier pigeons, notes
from
local authorities, paper forms) and
protocol
to reliably pass them on via radio,
Internet
* health (including aid personnel, remote experts)
- identifying any systemic health risks that
existed prior to the disaster and persist
- identify any aggravated/propagated/created
by the disaster;
- identifying external expertise required and
local data to be gathered by field
personnel
to aid identification of a disease or
spread
- identifying surge capacity limits and when
they have been exceeded, relief priorities
for medical personnel, anticipating supply
outages and resupplying before they're out?
* logistics (including GPS, movement of aid/help)
- if more than one of them has to move, or
if one of them has to move more than once,
it's logistics; (excluding the victims
except for those who participate in aid)
- all requests, orders, labelling, tracking,
authenticating, delivering, and
transporting
(very heavy interaction with the grid which
should be capable of tracking all of this,
there must be many intersections between?)
- standing orders that can be diverted to a
more urgent purpose, verified orders that
cannot be?
- physical control and tracking of vehicles?
* water and sanitation (including aid personnel)
- tracking where potable water comes from,
ensuring optimal use of this, spotting a
likely shortage and resupplying in advance?
- identifying water sources actually used
and instituting testing and safety programs
even if these are inadvisable water sources
- identifying waste water and sewage streams
and tracking e.coli (for instance) to
source
(more the duty of water than health
people?)
- ensuring good animal watering and pasturing
practices to avoid doing damage to sources
- building permanent cistern/sewage systems
where this is simpler than relying on more
temporary solutions and will free effort or
temporary rigs (water trucks, port-a-johns)
- an obvious interaction with early
recovery
* nutrition (including feeding of aid personnel)
- identifying numbers and cultural
constraints
for appropriate supply of compatible meals
- ensuring food actually reaches those hungry
- discovering malnutrition, hoarding, abuses
such as profiteering, and then compensating
- obviously involves also health/protection
- identifying long-term nutritional
challenges
(lost arable land, farmers forced out, lost
livestock, no funds to buy seeds, and etc.)
- obviously involves recovery & resilience
* protection (including credentials, evacuations)
- physical security of the grid and water and
any health-critical facilities (hospitals)
- security of camp borders, and aid vehicles
and supplies, to avoid looting or stealing
- communicating decisions not to protect or
to facilitate otherwise unauthorized
actions
(allowing hungry people to "loot" a grocery
store on grounds the food is going bad,
etc)
- identifying any inter-factional conflicts
or
criminal activities exploiting the
disaster?
or is this ruled out for political reasons?
- identifying local groups suppressing danger
or discouraging exploitation of the
disaster
and the help that they require to continue
- an obvious intersection with resilience?
- broadcasting assurances and how to get help
so people feel more confident and secure
(and if this includes actual peacemaking
then)
- systematically searching out abuses of
power
by people in positions of military or
police
authority, or those "in charge" otherwise,
so as to apply pressure via sanctions like
a
withholding of certain aid or to run a
guns-
for-food or guns-for-shelter program or
etc.
- creative conflict resolution methods, such
as organizing games and recreational trips
* resilience (all auditing and risk measurements
that describe the long-term situation
developing)
- risk assessments involving more than one of
the above cluster or interactions between
- conflicts with long-term authorities or a
powerful local group with informal
authority
likely to persist past the disaster relief
- ensuring documentation of early recovery
efforts sufficient for long-term efforts to
maintain and improve what has been started
- maintaining vulnerable persons lists and a
manifest of what they owned or brought with
them to camps, especially identifable items
that could help identify looters or abuses
- identifying all single points of failure
and
actual failures of key grid, water, sewage
and shelter systems that may have cost
lives
- overall morbidity and value of life
analysis
to determine whether aid was optimally used
- identifying informal roles and volunteers
to
be recognized and trained or credentialled
- identifying capability and surge capacity
gaps that would occur in a similiar
disaster
- understanding the lifecycle of the disaster
If you're looking for just one to start with, try
nutrition, as it's the easiest to separate from all
the others, and the least affected by resilience (you
can't reasonably expect people to grow food for their
own use in an emergency, it'll primarily be imported).
I already outlined (previous note) what the grid does,
though I didn't deal at all with credentials and the
logistics of getting telecom gear into place, on which
questions I defer to the latest US work on this. Nor
with protecting that grid against sabotage or
overload.
Every one of these clusters probably has a dozen or
more distinct identifiable processes "under" it. I
think the best way to think of the cluster is that it
is a perspective of a particular profession or set of
skills, and accordingly many processes (those that
need multiple professions and diverse skills involved)
are going to be classified under multiple clusters.
It should even be possible to identify, based on
actual analysis of transcripts of communications and
field anecdotes, exactly which types of people get
involved in which types of situations, and so identify
the clusters empirically. Which I'm sure is where the
UN list comes from. But the clusters won't stay the
same as professions and roles and resilience
approaches change, so stating how they are derived is important.
____________________________________________________________________________________
Park yourself in front of a world of choices in alternative vehicles. Visit the Yahoo! Auto Green Center.
http://autos.yahoo.com/green_center/
Received on Thursday, 21 June 2007 23:38:53 UTC