First, let me introduce those at the table here. First is Randall Flynn, Executive Secretary for Foreign Names of the U.S. Board on Geographic Names and the geographer for NIMA. Sitting next to Randy is Kathleen O'Brien, who is from the Canadian Permanent Committee on Geographical Names based in Ottawa. Next to her is Tim Norton, who is chairman of the Arizona State Board on Geographic and Historic Names. Next to him is Wayne Furr, who is Executive Secretary of the Oklahoma State Board on Geographic Names and also the Executive Secretary of the Council on Geographic Names Authorities, which has membership from all the various states in the U.S., observer-status from others. Sitting next to him is Tim Gregg, who is the Executive Secretary of the Washington State Board on Geographic Names. And last but most definitely not least, Mr. Scott Jukes who is with the State of Victoria's Names Authority in Australia.
My name is Roger Payne. I'm the Executive Secretary of the U.S. Board on Geographic Names and also Executive Secretary for Domestic Names Activity. And once again we hope to have a good bit of dialogue here, even though it's late in the afternoon, so we're going to be brief. We have a lot that we would like to cover, and I understand that the last outbound metro train is at midnight, so we certainly have to finish before then. Well I must say I'm not sure exactly where to begin - extraordinary presentations all day long, and I did something that I haven't done in a long time - I felt like I was back in school and I took a lot of notes. So, very beneficial for me already, and I hope that what we say to you this afternoon will have some meaning to you. I want to start by saying something by way of introduction.
We're in the standards business, and quite frankly we believe that standards are the core function of government. Now I know there's an organization called the American National Standards Institute, I've been a representative to that organization for 16 years. But at least that has a government stamp on it as well, so standards are the business of government and at least government approval, we think. Otherwise they're not really a standard, are they?
Standardizing geographic names is all about communication. If you don't have standard geographic names, you cannot communicate properly. Much worse than that, you miscommunicate, and that's really bad news. We've heard a lot about emergency preparedness and digital locations, but also names are going to be used and we certainly don't want to send someone to the wrong place because they're using a map with a different name from the person who happens to be injured, and so on. We can talk about those kinds of situations all afternoon.
I think I worry that a lot of people, even such an assembly as this, tend to think about standards as a barrier, an impediment. Not so. We would prefer that you think of standards as giving you more freedom. You don't have to worry about anything, it's all taken care of for you, you have the freedom to use these digital databases with the knowledge that everybody else is using the same name and using an official name. So it's not a barrier, it's not an impediment, it's certainly not a panacea either, but it does provide you with one name for one feature, and that's the most important thing, everybody using the same name.
By the way, what I'm going to be saying applies to the federal government, the federal government's position. The federal government does not care what the name is - most people have a misconception about this. The Board and therefore the federal government have no preference for the name, aside from a couple of policies that are kind of a big deal which I'll show you in just a moment. We don't care what the name is, we care that everybody use the same name for the same feature, and that's why since 1890 the U.S. Board on Geographic Names had been in business.
I don't care if you interrupt me if you have a question - I find it grows stale maybe if you don't ask it when you're thinking about it. I've already mentioned this: clear and unmistakable references, the ability to communicate, using the same name for the same feature, mutual understanding, mutual intended understanding, and also assisting in international relationships and cooperation.
The U.S. Board on Geographic Names was first organized in 1890 by executive order. It went along and did a lot of neat stuff and then in 1947 Congress said - this is so important that we cannot have it under simply an executive order, so the Congress abolished the Board and re-established it by Public Law 80-242. And this is an excerpt from the mission of the Board: "to provide for uniformity and geographic nomenclature and orthography throughout the federal government". How simple is that, yet how ambiguous can it be? Some people have interpreted that to mean only physical features, because of the word geographic, but others have taken the wider application, meaning all names, and in fact only one category of name was specifically excluded by direct mention in the law, and that's the names of federal buildings. Nothing else is excluded. It is now, but it's not according to the law.
And the second most important piece of the law 80-242 is promulgation. Notice it says "the Secretary": "the law that created the Board said that the Secretary of the Interior and the Board will be responsible conjointly for the standardization of geographic names throughout the federal government". For all practical purposes the federal government through the Board on Geographic Names never names anything, it's a reactive body. We don't name anything. A federal agency may propose a name on its own, but the federal government through the Board never names anything, period.
I want to mention a little bit more about the Geographic Names Information System. The Board and its representatives have for a long time been involved in international activity. That international activity includes United Nations work. The UN group of experts which was established in 1967 or actually a little bit before that meets once every two years to discuss lots of things, and Randy is going to tell you a little bit about data exchange work we're doing with the UN, I think. We do a lot of other work. We're also involved with the Pan American Institute of Geography and History. Once a year we offer a course in applied toponomy under the auspices of that group. I head that teaching team. Randy is still on it, I guess, although he hasn't appeared in a couple of years, we've had a substitute for him. We also have Senor Ramon Rivera from the National University in Honduras, who is the third member of the teaching team - he's really the most important member but he was the third chronologically to join the team. So we're involved a lot in international activity. We provide consulting services to any country that would like to benefit from all the mistakes we've made over the years in establishing databases and establishing policies toward …
(Break in text between tapes)
…experts on geographic names in 1967 a number of resolutions were adopted. The fourth one, which is quite involved and covers about five pages, is one of the most important ones and this is an excerpt from that resolution: "the best way to achieve international standardization is through strong programs of national standardization," and that's what we encourage, that's what we try to develop, and that's how we try to assist other countries who can benefit from, as I said, all the mistakes we've made over the years.
Geographic Names Information System: as I said, in the U.S. that is the only database that any federal agency may use for geographic names applications in federal maps and documents. Along the lines of the question that Randy asked before the break, I'd like to echo and I'm sure he will, all you folks out there compiling information on geographic names, we would love to have it. We would love to process it and verify it according to our own methodology and procedures and put it into GNIS, the one official source. It disturbed me a little bit to hear about all these specialized distributed geographic names databases. Because: one, everybody might not know about them; two, they might have things in them that violate federal naming policy, that may be tempting federal employees to use them, and they may have preconceived notions in them. You see we don't do that; we collect every name for every named entity, period. We don't try to second-guess the user's needs, we just want to collect everything and put it there, so that anybody can extract it and build their own application-oriented database for their own internal use. Well, this database has - yes?
[someone from the audience] - You were talking earlier about temporal problems, trying to figure out the time of certain events in certain areas. In Colorado recently, Boulder County was split up into two counties: one - Broomfield County, and one - Boulder County. How does this take care of names versus counties, especially when counties change their shape?
Roger - We receive a notice of that from our good friends at the Census Bureau and as soon as we do, we go into the database and make the appropriate changes. We update all of the entries that are in the new county accordingly. By the way, it's not just a county, we record every county in which a feature is located, so for example every county that the Mississippi River's in - you want to take a look at the record for the Mississippi River in the database some time. There's a reference to every single 1: 24,000 scale topographic map through which that feature passes, every county it’s in, every state it's in, and also more than 110 variant names recorded for the Mississippi River. So yes, we record everything and, should something dynamic happen like a new county, we will immediately fix it in the database as soon as we receive official notice of it. Yes Ma’am?
[someone from the audience] - Are you dealing with the temporal component of that? It used to be in that county; I need to be able to look that up.
Roger - Yes. There is a field called history and in that we would put a simple phrase - it's not a big deal, it just says "formerly in'.
[someone from the audience] - Yes, how do you reference counties, polygonal features? Are they a point in the middle, and what do you reference it to? You have a USGS topo map there. Is it a 1: 25,000 map that you reference that to?
Roger - Yes, we reference it to it. First of all, let me say this database right now is a text-oriented database; it is not in any way specially enabled. We are redesigning it, and I'm fairly sure it's going to be spatially enabled - I'm hoping it will be anyway - I'm looking at Dave Govoni back there. So the answer to that question is, it is not a spatially enabled database at this time, so we have no polygons in there, we have no boundaries defined by coordinates. It is point, area and linear data, all based on a scale of 1: 24,000 or 1: 25,000. The reference to topographic maps is especially on the linear feature, for example. There is a set of geographical coordinates recorded for each topographic map on which it's located. So, for example, if you plotted out a linear feature it would be a zigzag line like that. That's all we're doing right now. Does that answer your question? But there is a reference to each topographic map on which it's located at that scale. Another question?
[someone from the audience] - Do you have actual authority about where these places are, like if somebody wanted to know what county they're in and there's some dispute about that. Would you be the authority on that?
Roger - Yes, for the federal government, but before we would promulgate that decision we would have received all of the necessary recommendations from the state level and the county level, and then a decision would be made about the extent of feature, if I understand your question. Yes, and here's something to know when you're examining this database, another hotly contested issue but nonetheless, good, bad or indifferent, it's been the policy since 1890. If name placement on a map does not commit a name to tributaries or upper reaches of a stream, the source of that stream is automatically extended up the longest straightest drain, period, until we know differently. A lot of people don't like that but that's the policy. Question?
[someone from the audience] - Just to finish something on exactly that point. I use the GNIS database almost every day and I routinely run into things, because there are lots of canyons in Southern California and every time I look up a canyon entry there, a very specific latitude and longitude is given, but it's single point, it does not tell you where the canyon starts, where the canyon ends, and from what I can tell from the maps that I've looked at, the points given for canyons always seem to be at the mouth of the canyon. Is there some reason for that?
Roger - I'm a little perplexed. You'd have to give me the specific example because canyons are linear features. You should have at the very minimum two sets of coordinates, the mouth and the source, and you would have any number of what we call secondary coordinates, should that feature pass through other topographic maps. So I'm a little perplexed, I'm not sure why you're not getting them; they're definitely there. As a matter of fact, the data entry software will not let a canyon feature into the database without a minimum of mouth and source. So if you give me an example, we'll see what's going on.
[someone from the audience] - I was under the impression that the coordinates for each name are not where a feature actually exists on the map, but where the name occurs on the map.
Roger - Absolutely not. I appreciate you bringing that up so we can clarify it. It is the point that the annotator determined where the feature exists, not the labeling of the name. It is a point for the feature itself. Thank you for bringing that up.
How was this compiled? Just for your historical information, it's a 25 year project, started in 1976 and we're pretty much winding down with it. The first phase was to take the names off of the three, now four, primary federal graphics: U.S. Geological Survey topographic maps, U.S. Forest Service maps, Office of Coast Survey charts, and National Park Service maps. That we did and finished by 1984. We have since 1984 been compiling what we call phase 2: four years per state, contract with usually a state agency or a state university because there's no profit in it, it's really a cost resource-sharing contract where they examine all other known sources approved by us and input the information, and this is where the historical information comes in. That's the last thing done, but as much of the historical information as we can gather is put into the database. It's never finished, in four years they run out of money and time and they still have many other sources. Some states have as many as 400 sources or more checked.
You can look at a status map on our web page as to which states have completed extensive compilation and which states have not. You should be hopefully aware of our web site at http://mapping.usgs.gov/www/gnis. You can also click into Mr. Flynn's foreign names database from the introductory paragraph. You can also query our Antarctica file and you can also link to any other country's web site that has a search engine. Mr. Govoni?
Dave Govoni - We are upgrading our servers, and one of the changes is that we'll be moving GNIS to its own server within approximately a month, we hope, if all things go well, and I know that some agencies embed the GNIS service as an integral part of their web pages, and you need to be aware that that link will be broken when our server shift is done. So if you have that kind of situation, or you want to be notified, let Roger or me know and we'll notify you specifically when the cut-over occurs, and of course we'll have it on the web page as well.
Roger - It's never a good idea to change addresses, but in this case there's a positive outcome from it. We're using a non-standard port right now, and those of you with fire walls can't access it unless you program around our non-standard port, and when we switch it over to the new server, we're going to use the standard port, is that right Dave? Hallelujah. Another question? The question is - what about geographic coding for types of features? We're not in charge of FIPS codes per se; we are bound by federal law to carry FIPS codes in the database. For example, for every county we have its five-digit FIPS code counterpart of state and county digital code. We have to carry it by law. As it happens, it's more trouble than it's worth. We're also the responsible maintenance agency for FIPS 55, which is heavily used by the folks at Census; we're the maintenance agency for that one FIPS standard, but we also carry all of the FIPS codes because we have to. Nobody uses them but we have to carry them. They don't use them any more, they used to - you use them? OK.
I just want to take one more minute to say something about type of feature. We had a lot of discussion this morning about standards for feature types. We don't do that. You saw our 64 categories. Those were almost developed by Topsy, and maybe some are there that shouldn't be and some should be that aren't, but that's what we have. In the federal government there are no standard types of features, there are no codes for types of features that are standard, there are no standard definitions. We have encountered more than 2,200 different generic words applied to features using geographic names. We collapse those into 64 categories. I defy anybody to tell me the difference between a mountain and a hill. Tell me the difference between a creek and a river. It absolutely agonizes the hydrologists to know that just up the road from USGS headquarters Little River flows into Goose Creek - they can't stand it. But that's the way it is. The naming process is a perceptive process, and the use of names and the use of types of features and their definitions is totally application-driven. That's why we don't even try to do it. We tell you that we have these terms, developed by us and defined by us solely for the purpose of retrieving features with similar characteristics, period. That's why our definition of stream is so short: "linear overland flowing body of water", period, of which there are 154 generic terms that fit into that category, not the least of which are creek and river. The highest point in Ohio has "hill" as a generic term, with a mountain next to it that's smaller than it. Go figure, that's the naming process, it's perceptive.
Same thing with regions, we constantly get questions about regions.
No official federal definitions of regions: individual agencies may have
their definitions for their own in-house use, but no federal government
definitions. People can't even agree on New England, believe it or not.
I also listened this morning with interest about the association of a property
with place - I like that, “it's hot in New York.” Well, if you don't have
standardized names, it doesn't matter if it's hot or not, because you don't
know where you are, you're not talking about the same place. Believe it
or not, I can find somebody who won't agree with any name. There is a movement
to change the name of New York City back to its original name New Amsterdam.
Is it going to happen? Probably not, but there's somebody somewhere that
will not agree with you on any particular name. By the way, you can ask
questions - on our web site where it says GNIS Manager, guess who gets
a hundred of those a day? Thank you. I appreciate the time to talk about
standardization.
Mike Goodchild - We've got a few more minutes in this session, and
just to do a reality check here - this session was titled " Users of Gazetteer
Data and Services", so two questions before we wrap this session up. Do
we have an adequate view of the users of gazetteers at this point to drive
us into the subsequent discussions, and what are we missing? Are there
areas of gazetteer use that we've not covered in one way or another in
this session?
Pat McGlamery (Map Librarian at University of Connecticut) - I kind of liked it when Cliff did that fade and said, OK, I am the server and you are the client, and the question was asked of the server, and the server turned to other servers and said, yes, I can answer that and went back to the client and said, I can answer that. As a librarian, when I look at what my colleagues are doing with digital library initiatives, especially with coding text in SGML, and creating users that are data objects like the Diary of Jedadiah Hodgekiss or these expeditions texts, there are times when the user is really the text within a scanned and textual database - that's going to a gazetteer and linking up. Where the user's not really a person but in fact a number of textual databases that exist out there.
Mike Goodchild - Interesting perspectives. Any other comments?
Randy Flynn (U S Board on Geographic Names and NIMA) - Just looking back at the title of the session, the last phrase has to do with contributing gazetteer data, and I hope I didn't miss it but I'm not sure that I heard a whole lot about gathering geographic locations, and then contributing them to some second or third or fourth party. Perhaps the closest we got to the discussion was when some of the folks associated with museums doing the historical research - for example, finding the places cited in expedition logs, would do the research, apparently find the place, identify it but then what did you do with it? Did you make it available to a larger community, and if so, in what way?
Mike Goodchild - Good point.
Ray Larson (UC Berkeley) - I just wanted to comment on the previous comment. A few years back I had some students who worked on exactly that problem of using gazetteer information to georeference documents, and I'll be talking more about that later but I wanted to point out that that really is a use in the digital library world where you have no real georeferencing for a document, but you'd like to place it in its geographic context.
Tom Moritz (American Museum of Natural History) - I think that question of how we contribute is a really interesting one. We've actually kicked that around a little bit within the Natural History community, wondering how profitable it really is to concatenate or to collect the specific placename work that's being done, and particularly with our legacy data. Personally I think it is valuable, and I think it's an important element to include and I actually did address it in my talk. Sorry it was not more clearly expressed, but I like the idea of having a distributed system which allows us to collect the community's effort in the process, because these initiatives are going to go forward - as I mentioned, there are 6500 natural history museums.
Already these things are happening on an individual researcher basis
or sometimes a departmental basis in different museums. I think there's
a real benefit to come from collecting those in some coherent way. The
relative importance of that varies because sometimes there may be only
one or two other collections that exist outside of that immediate institution
where that site has actually been used. Nevertheless, I think
we should be sharing the information because there may be a need for return
visits to that site in a biodiversity context, or there just may be other
research purposes for having that site located. The Getty Thesaurus does
provide that dimension, I believe, it may be one of the only ones that
really has gone retrospective and has been trying to look at that problem.
So I just support that comment generally.