The Universal Locator Project was originally designed to provide a
more flexible interface and functionality to Web-based address and business
location services. The original prototype consisted of an Access Database
that contained roughly 6,000 business listings for the City of Berkeley,
California. In addition to that, this database was web enabled through
the Cold Fusion Application environment and Microsoft's Internet Informaiton
Server. For interface functionality outside of HTML and CFML java applets
were used, specifically the "Mapper" applet and the "CFTREE"
After the protoyping and testing phase was complete ULP demonstrated
the following benefits:
- Multiple domain based searching (i.e. by street, by industry category,
by name, by address)
- Map interface that dynamically plotted several listings as from
the user's selection
- Expandable tree structure interface for browsing listings without
the aid of a map.
In addition to benefits the testing and protoyping phase also revealed
the following drawback to ULP:
- Transaction times were to slow.
- Queries were too complex.
- Users are unfamiliar with SIC codes.
- After initial system optimization functionality seemed to drop to
the extent that did not comply 100% with the projects mission.
- Map preparation hard to streamline. Should be outsourced.
- Data access and integration was too rigid and should be outsourced.
These setback, however, do not mean that this project should be discontinued,
but rather that the implementation process should be revisited. One
major revision should be with data access to maps and business listings.
If a subscription to a service that provided this type of data were
established then ULP could focus more on the interface and its functions
and less on the process in which data is collected and prepared for
use in the applciation. If this were achieved ULP could be implemented
in a way much more in line with its mission of providing better use
and functionality with business listings.