Formal metadata: information and software
Formal metadata information and software
Reply by Peter Schweitzer on 6 Feb 2001:
Because geographic information can be used in a variety of ways by many different people, it's important to know more details of the data's origin, history, and characteristics than you would need for a library book. So metadata tend to be long descriptions of the data. But it's hard to read long descriptions, especially if the descriptions made by one organization look radically different from those produced by another. Most geospatial data have a lot in common; using standard structures and formats for documenting them makes it easier for people to find the characteristics of the data that will help them understand it quickly. The same structures and formats help us write software to search for key characteristics of the data and to present the documentation in a consistent manner.
The Clearinghouse is a way for USGS and other organizations to make it easier for people to find, get, and use our information effectively.
But that technological capability does not bring understanding by magic. People still need to figure out what they have, what they need, and how any data that are offered to them might help meet those needs. The most dramatic reversal of recent time is that in order to work well, regular people (including regular scientists) have to become data managers. They need to use some systems for keeping track of the data they have and the data they need, in the context of the problems they want to solve. USGS can help with that, but not if our solution is to always be different. We cannot expect people to use our information just because it's available, or because it's free, or because we work for the government.
In the early days of the web, an organization could craft a web page using any graphic design, without considering a systematic method for arranging its information. Each web site was like a jewel, cut and polished, and each jewel sparkled when a bright light shone on it. Now those web sites are like grains of sand on a vast beach. If you pick up any one and look at it under a microscope, you can see its beauty, but your understanding will never be able to encompass all of the sand grains on the beach. To be effective now and in the future, web sites must be part of a larger knowledge-organizing system, more systematic than type-in-the-box search engines, returning information that is easier to scan and evaluate than artistic web designs.
That said, the exercise of creating thorough metadata will often uncover problems in the data that you will want to fix. Attribute values are misspelled or out of range, longitude values sometimes have the wrong sign, and bibliographic references for sources are sometimes incorrect in published texts. These errors are often caught when you make good metadata.
It doesn't always work. Our publications people are given conflicting instructions (do it cheap, get it out the door vs. do it right), and I fear that despite my entreaties the managers of those groups and of their putative clients, the Team Chief Scientists, do not value consistency in structure and form for geospatial data. The best work, as you might expect, is done by people who care. I do too, so I often find myself making significant revisions to metadata that people send me or that appear on our web sites. For that reason the metadata on geo-nsdi will often be different than the metadata that are provided on other web sites.