Name/Version Reviewed: cns (Chew and Spit) Version 2.6.3

Reviewer / Date: Hugh Phillips / December 28, 2000

Date of last known release: 11/4/2004 (version 2.8.0)

Function: metadata format preprocessor for mp (CSDGM 1998 compliant)

Documentation

Get binaries and source code here

Platform: MS-Windows, Solaris, Linux,

Principal Contact: Peter Schweitzer, USGS, pschweitzer@usgs.gov

Status: Functional, general release

Metadata Storage Structure: not applicable

Brief Description:

cns is a metadata preprocessor which processes metadata with varying degrees of conformance to the CSDGM (1998) into an output which is more compatible with mp. This is especially useful for Clearinghouses which may collect metadata from source agencies producing metadata differing in format because of the metadata tool or template employed at the each agency, or because the agency metadata profile differs from the CSDGM.

Cost: none

Notable Plus: customizable for element name aliases and local extensions

Notable Minus:

cns gets confused by lines in text elements which begin with a string that is the same as an element name. Because it designed to recognize such strings amid noise it naturally jumps on them when they are found easily. This is a well documented 'feature' and the author of the tool is looking for a good way to turn it off.

Metadata Exchange

cns will import almost any ASCII document that has structure and element names such that it 'looks like' CSDGM metadata. With use of an alias file it will also import metadata whose element names need to be translated to FGDC element names.

cns exports an ASCII file with two space hierarchical indentation and metadata element names acceptable to mp.

cns acts as a preformatting filter for metadata before it is presented to mp. It will in some cases supply missing container elements, but does not itself enforce the production rules for the CSDGM on fundamentally flawed metadata.

Useability

cns is a command driven application, a 'conversion tool.' There is a documentation file for the tool.

Administrative

The tool will operate under Linux, Solaris, and MS Windows (in a DOS window). The source code is provided so it may be compiled on other platforms. It occupies approximately 8 Mb with documentation and source code (in a standard distribution package with mp and tkme). The tool (when obtained as binaries) requires only a decompression and dearchive step for installation. The tool is publicly available, free, and has maintenance performed promptly as needed by the author. Through use of a configuration file it is possible for cns to import metadata files incorporating extensions to the CSDGM or descriptive element names, and export a file formatted for mp with these extension elements intact.

Tool Reliability

On a bad command line, cns returns allowable command line parameters. cns will reject all lines after a line beginning with a reserved element name in a free text section (that's a bad thing, but it is consistent!).

Sample of Interface: Command driven

Sample documents processed with cns:

• NRCS metadata document Characteristics: Centered major headings, element names without joining underlines, only data entry elements have colons.
  cns results: Very good. Rejected free text block containing sentence that began with the words of an element name. Rejected a 'Time Period of Content' that should have been 'Time Period Information.'
• DRG metadata from 'Capitol Cities' CD
  Characteristics: Lead in explanatory paragraph before actual metadata. Element numbers as well element names. Asterisks and '--'s wrap some elements. Element names in all capitals or mixed case without connecting underlines. Most elements left justified.
  cns results: Very good. Put the lead in explanatory paragraph in leftovers file. Converted block following Source Citation Abbreviation (which immediately followed Source Citation without an intermediate Citation Information) to free text. Converted elements following 'Longitude at Central Meridian' (which should have been Longitude of Central Meridian) to free text. mp would catch further errors in restricted domain elements whose values were enclosed in quotes.
• Preliminary Dodge county control point metadata created with Corpsmet
  Characteristics: With the exception of the seven major headings only the actual data entry elements (and their values) are present. A <br> tag separates each element and its value from the next. Elements and their values are left justified.
  cns results: Poor, even after all the <br> tags were stripped out. There are just too many missing intermediate elements for cns to reconstruct the document.
  Peter Schweitzer took this as an example fixer-upper document and documented the step-by-step procedure (link broken 12/17/2004) to bring it to a form that would pass mp using cns and an editor.
• Wisconsin Catalog of Aerial Photography metadata created with Metamaker 1.1 Beta
  Characteristics: Elements are numbered and do not have connecting underlines. The elements are left justified.
  cns results: Fair. The Metamaker output report is missing some intermediate elements which causes cns to get lost, and some of the elements are misspelled so cns can't recognize them (although this could be fixed temporarily with use of an alias file).

Sample Output:

• First try output from cns applied to 'DRG metadata from 'Capitol Cities' CD'