|
�
Metadata
Harvest
Although Metadata provides a good solution
to organize various digital resources, and improve the
quality of the information retrieval in the giant digital
world, its concept is still new to most information users;
let alone how to apply it in the traditional information
environment, especially there are so many Metadata standards
and schemas exist. Also, it's inconvenient and uneconomic
if we use human-generated metadata for each digital entity.
The best way for metadata application is letting machine
generate metadata according to the preset schema or standard.
But with current technology, automatically generated metadata
could not provide enough precision, which restricts the
implementation of metadata in the real world of information
retrieval.
Lots of projects have been implemented trying
to solve this problem. They can be grouped into several
components: Metadata Creation, Metadata Mapping,
and Metadata Registry.
There are lots of tools or software designed
to create metadata for specific digital resources. These
include Metadata templates and harvests. Normally templates
require user inputting the descriptions of a resource's
major properties such as title, author and other information
to generate the metadata elements according to the predefined
metadata schema.
Nordic
Metadata Project has a very famous Metadata
template for Dublin Core Metadata Element Set creation.
The U.S.
Geological Survey (USGS) also provides several software
to parse or edit Formal Metadata according to the FGDC
Content Standard for Digital Geospatial Metadata. This
type of software can be simple that only uses fifteen
DC elements. It also can be very complicated that use
several levels of Metadata elements set, such as GEMCat
from U.S. Department of Education Initiative for the
Gateway to Education Materials project. It adopts
DC elements and qualifiers to describe electronic educational
materials on the Internet.
Metadata
harvest are those tools that extract metadata description
of Internet Website from HTML Meta tags according to Metadata
schema chosen (most of them are using Dublin Core). Some
simple harvest tools only require user typing the URL
of that Website. They can be seen as a simple web search
engine only index Meta tags in HTML headers. They are
useful only when most Websites they index using Meta tags
for site description. One example is DC-dot
from UKOLN.
There are some software that combine both
template and generator functions together. One example
is Reggie, the
Metadata editor from the Resource
Discovery Unit (RDU) project operated by Distributed
Systems Technology Centre (DSTC). It can create eight
types of metadata format using Java Applet including DC,
GLIS, IMS, GEM. User also can input one Website URL to
generate specific metadata (if applicable).
�
| 
