Le portail francophone de la géomatique

Hello all,
Belated response to Jarmo's input to the discussion, but I would bring to
the attention of the list (and addressees) a very interesting article on
Mapping the Future concerning the efforts in the USA to produce (new)
high-quality topographic data at federal and state/regional level.
This article did not appear in one of the mainstream GI/GIS magazines, but
rather in one of the more respected scientific journals - Science - Volume
298, p..p. 1874 - 1875, 6 December 2002.
The article points out that:
By the late 1980s, when USGS finished mapping the entire country (having
started in 1879!) at an estimated cost of .6 billion and 33 million work
hours, the survey's 55,000 'topo maps' had become standard tools for
hikers, urban planners, and relief workers such as the Red Cross, among
other consumers. As USGS's priorities shifted toward scientific research,
however, its mapping program languished. As a result, while towns went boom
and bust and landmarks such as airports, buildings, and parks spread and
dwindled, the topo maps lagged further and further behind the landscape
they represented. Today, the maps are only sporadically updated, and some
are 57 years old.
So I think there is some merit in questioning the current quality of
certain of the free USGS spatial data. Question is, how much is of poor
quality and how much is more current and of higher quality?
We just don't have the money to maintain a robust revision program, says
William Flynn, chief of USGS's Mapping Partnership Office in Austin, Texas.
So USGS is trying to upgrade its maps and their quality by drawing on the
spatial information resources of hundreds and thousands of communities
across the USA via a National Map team which began brainstorming ways to
combine existing digital data on elevations and hydrography at USGS with
the high-resolution maps emerging at local levels across the country. 'Most
of the local communities maintain current Geographic Information Systems
(GIS) data sets at very high resolution for their own purposes, anyway,'
says Flynn, head of Texas's National Map pilot. 'Someone simply needs to
develop the partnerships with those communities and take on the technical
challenges of integrating the variety of data and formats into a standard
set of products for distributing over the Web.'
Sound familiar?
USGS envisions the National Map as a seamless, continuously updated map
with seven layers of geospatial data for every U.S. county. The data
include topographic features from aerial photos and satellite images;
surface elevations; locations of water bodies, transportation, major
buildings, and public land boundaries; road names; and land-cover types,
such as open water or high-density residential.
Eight pilot projects are underway to kick-start the process. One, based in
the Lake Tahoe area (desperate need to manage water resources here - Water
Framework Directive participants, pay attention!), straddling
Califorina-Nevada state borders, found that, you guessed it -
We thought that in a well-known and highly studied environment like Lake
Tahoe, there would be a wealth of core data. ... We were surprised to find
a lack of any kind of regional data sets.
Instead, they found a mixed bag of geography: excellent GIS measures in
some agencies, poor ones in others, and, in one county, unusual computer
software that required painstaking translation. ... There's a lot of data
there, but the integration is a huge effort. A representative of the
Delaware Geological Survey and member of the Delaware pilot team, agrees:
Even for our group, with considerable technical expertise, getting the
framework data to an accurate scale is a lot of work. Storing,
communicating, and automatically updating nationwide digital geographic
data pose daunting problems of their own.
USGS estimates that delivering the full-scale National Map in 10 years will
require million a year - i.e. 1,500 million overall ! - about
twice the current budget.
Welcome to the world of high-quality data - at a price.
Who pays that final cost - government or end-user - is another matter,
which is still not wholly resolved in the USA, since states and local
governments do *not* have to make their data available for free. Certain
states even have state law requiring that they charge for spatial
information and/or the GIS and services used to collect or use such data.
No mention in the USGS article as to how these legal issues will be resolved.
Hope this small intervention helps muddy the waters a bit more! There is
probably no clean solution to this debate, but the issues need debating
in any case. (Shame that we aren't hearing more from the GINIE project team
on this! I thought policy issues were their reason for existing - and
being funded by the Commission! ;>)
Kind regard
Roger Longhorn

Of equal interest is the US Census Modernisation Programme, since the Census MAF/TIGER
national data coverage is for some application areas (Business geographics, transportation,
...) the source for de facto US Framework data.
G.B. Le census bureau definit clairement le MAF (master adrss file, geocode) comme une
ressource pour l’industrie geomatique, pas seulement un outil de recensement.

The topographic content of TIGER data (which includes streets and their names, lakes,
streams and their names, railroads, geographic entity boundaries, housing unit locations and
key geographic locations (schools, airports, etc)) originated from USGS data (in 1970's
and1980's subsequently augmented in many ways. The result is a complex set of accuracy
and quality problems. Addressing these is the first (and perhaps the largest) aspect of the
current MAF/TIGER Modernisation programme. A good overview of this is to be found in a
presentation by Bob Marx, Chief, Geography Department,
at:http://www.census.gov/geo/mod/BobM.pdf
The accuracy/qusilty issues are being addressed, driven by the intent to use GPS technology in
the 2010 Census. The first stage is the award (announced last July) of an eight year contract 'in
excess of £200 million' to Harris Corporation: http://www.census.gov/geo/mod/maftiger.html
G.B. L’utilisation du GPS est associee a celle d’orthophoto, qui permettent de suivre la
dynamique (nouveau bati, nouvelles routes). Elle exige de la precision (de quel cote de la voie
se trouve le bati). Elle permet de l’analyse spatiale independamment du decoupage en tracks
et blocs (IRIS, ilots), chaque bati etant repere par ses coordonneesGPS. Cette possibilite n’est
pas utilisable aux Etats-Unis(meme pour les colectivites), les adresses etant protegees
comme donnees personnelles (vois-ci-apres obligations de confidentialite liees au Title 13). Il
reste que Tger comprend une bonne part de la topographie de base (voies, toponymie,
hydrographies, decoupages, bati, equipements publics).

As with the USGS National Map project, the thrust is Partnerships with state, local and tribal
governments. Census have posted this week a robust statement of their postion on 'free
access', including the following:
In the foregoing context, it is hard to imagine a business model (oftenreferred to as,
'privatization') that could succeed in an environment that requires open access to NSDI types of
information, as current federal policy has for many years. The current model with
government/NSDI data and commercial/value-added data each existing in separate
environments has been the most successful in the world. No country with restrictive policies
toward spatial data, including those with highly developed governmental and commercial
sectors, has evolved a robust GIS business environment as exists in the United States. Some
have tried; all have failed. Such approaches always create roadblocks to enterprise systems
and data sharing. (See below)

Several points emerge from this chronicle:
a. as Roger and others have pointed out, the costs of data and the care of data have to be met
from somewhere.
b. rich countries like the USA may be able to afford the kind of cost duplication that arises from
not having a single authoritative framework data coverage, but it should be avoided if possible.
c. in a number of countries, there are very substantial investments to be made/costs to be met
to deal with legacy accuracy issues to make data fit for use in the GPS era. The way in which
these are handled is likely to be an important discriminant between the 'free access' and 'user
pays' regimes.
In the former the prime driver will be Government priorities - hopefully in the latter there will be
more of a 'duty of care' to users, as they are in fact customers.
G.B. Le refus du partage a faible cout des donnees de base conduit a la duplication de ces
donnees, aux frais du contribuable. Mettre les donnees de reference a la disposition de tous
au cout d’extraction est un moyen de reduire la depense publique… Cela suppose que l’IGN
se tranforme de vendeur de produits en integrateur.

Background Information on the MAF/TIGER Accuracy Improvement Project (MTAIP) and
Governmental Partnerships
The Census Bureau's MAF/TIGER Enhancements Program seeks to achieve a high level of
map coordinate accuracy in TIGER by seeking, as a first priority among data sources, digital
files prepared and provided by state, local and tribal governments. The Census Bureau will
have its contractors (the folks who will actually perform the much needed TIGER realignment
work) use all high accuracy GIS files that state, local, and tribal government are willing to
provide without royalty or copyright restriction as part of the MAF/TIGER Accuracy Improvement
Project (which is the largest component of Objective One in the overall MAF/TIGER
Enhancements Program).
Objective Three of the MAF/TIGER Enhancements Program is to Enhance Geographic
Partnerships with federal, state, local and tribal governments. The goal of Objective Three is to
gather reliable information (metadata) about the GIS files that are available (location accuracy,
vintage, etc.), store that information in a repository called the TIGER Enhancement Database
(or TED, for short), obtain copies of those GIS files that meet or are better than the location
accuracy requirements of the MAF/TIGER Enhancements Program when those files are
available without royalty or copyright restriction, to make use of this GIS data from geographic
partners, whenever possible, and to develop a process for the ongoing receipt of acceptable
information from those partners.
The Census Bureau cannot accept (nor does it favor) any restrictions on basic high quality
geographic data of the types included in the National Spatial Data Infrastructure (NSDI), such
as street centerlines, street names, hydrography locations, elevations, boundaries, etc. It is not
an acceptable option for the Census Bureau to report a numeric data total for some
geographic area (such as a 'block') and then not let the receiver of the statistical data see the
definition of that block (in the form of a map or a digital file) unless they insert their credit card.
Information of the type included in the NSDI is inherently governmental. (In adopting the
foregoing position, the Census Bureau is not suggesting that all these geographic data must
be exempt from copyright or restricted status. Value added items, such as speed limits,
numbers of lanes, traffic direction, turn restrictions, the locations of businesses that make use
of the file more convenient, the generation of 'routes of travel,' etc., certainly can be considered
for such protections.)
In the foregoing context, it is hard to imagine a business model (often referred to as,
'privatization') that could succeed in an environment that requires open access to NSDI types of
information, as current federal government policy has for many years. The current model with
government/NSDI data and commercial/value-added data each existing in separate
environments has been the most successful in the world. No country with restrictive policies
toward spatial data, including those with highly developed governmental and commercial
sectors, has evolved a robust GIS business environment as exists in the United States. Some
have tried; all have failed. Such approaches always create roadblocks to enterprise systems
and data sharing.
Some have portrayed the Census Bureau's approach to geographic partnerships as, cannot
[apparently meaning, 'will not'] partner with states [apparently also meaning, 'local and tribal']
for data. Clearly this is neither the Census Bureau's current policy nor its intent. Some of the
foregoing perception legitimately evolved from [apparently very frustrating] experiences in
conjunction with the geographic preparations for Census 2000 when operational schedules
precluded the Census Bureau from accepting offers of good files from then willing partners.
As illustrated in the opening paragraph, the Census Bureau is now able (and anxious) to
actively seek geographic partnerships with the operational pressures of Census 2000 behind
the agency. The Census Bureau's hope is that those disappointed with the earlier rejection of
their offers will recognize -- and accept -- this change in approach and participate now while
the opportunity exists, and before the door may need to close again late in the decade, nearer
the operational pressures for the 2010 Census.
Others have suggested the Census Bureau cannot [will not] partner with states for data
because the Census Bureau is not a fund-granting agency. This perspective ignores the fact
that the Census Bureau has a well-documented history of negotiating agreements that involve
in kind transfers of services and/or products with partner agencies. For example, as part of
the MAF/TIGER Accuracy Improvement Project, if a government has highly accurate imagery,
but cannot afford the feature extraction cost to transform that imagery into GIS files, the Census
Bureau is willing to have its contractor use that imagery to reposition and update TIGER and
return a corrected file for loading into the GIS software of the imagery donor. Thus, the Census
Bureau achieves its objective, the donor achieves their objective, neither agency has spent
more than it originally had intended, and no funds actually needed to change hands.
Another perception that may lead to a view that the Census Bureau cannot (will not) partner
with states for data is a consequence of law. The Census Bureau is bound by the
confidentiality provisions of Title 13, United States Code, which is the collective set of laws
passed by the Congress over several decades. Title 13 requires, among other things, that
information about individuals and establishments (including their addresses and/or specific
locations) cannot be disclosed (except in the form of statistical totals) to anyone who has not
agreed to abide by these restrictions. Thus, much to the annoyance of some who participated
in the Local Update of Census Addresses (LUCA) program conducted as part of Census
2000, only those local and tribal officials who agreed to sign a Confidentiality Agreement could
review the Census Bureau's address information, and they could not retain or use that
information for any local purpose.
The fact that the Census Bureau cannot (is prohibited, by federal law, from) let geographic
partners use the address information it compiles does not mean the Census Bureau cannot
[is unwilling to] be a good geographic partner. The widespread and frequent availability of
TIGER/Line Files, including the location-corrected versions that will be issued as the
MAF/TIGER Accuracy Improvement Project reaches conclusion for each county, are not
restricted by Title 13, because they do not contain confidential (Title 13 protected) information.