An XML-based bird image cataloging system

Here is the opening comment block for the RNC schema.

# slidecat.rnc: Relax NG schema for bird slide 
#   Documentation: http://www.nmt.edu/~john/scans/slides/
#----------------------------------------------------------------

The root element for the entire XML document type is image-catalog:

start = image-catalog

image-catalog = element image-catalog
{ attribute date { text }?,       
  attribute revision { text }?,   
  original*                       
}

	The date attribute is intended for an RCS Date tag or other content indicating when the file was last touched.
	Intended for an RCS Revision tag or the equivalent.
	Each primary entry in the catalog is an original element.

Each original element describes one exposure, whether on film or digital media.

original = element original
{ attribute cat-no { cat-no-pattern },     
  attribute ab6 { text },          
  attribute state                  
  { xsd:string { pattern='[a-z]{2,3}' }},
  attribute qual                   
  { xsd:string { pattern='[abcdf]' } }?,
  attribute scan { xsd:positiveInteger }?, 
  attribute old { old-pattern }?,  
  attribute arch-no { text }?,     
  info                             
}

The cat-no-pattern pattern defines the format of a catalog number.

cat-no-pattern = xsd:string
{ pattern='[0-9]{4}-[0-9]{2}-[0-9]{2}[\-a-z][0-9]{4}'
}

The old-pattern pattern defines the format of an old-style catalog number.

old-pattern = xsd:string
{ pattern='[0-9]+\.[0-9]{2}'
}

	The catalog number is required and it must be unique within the file.
	A list of one or more six-letter bird codes representing species in the image; each code may be followed by a ? character if the identity is not certain. For the bird code system, see A system for representing taxonomic nomenclature. Examples: ab6='amerob' ab6='larbun?' ab6='eargre amecoo' ab6='dowwoo|haiwoo' ab6='amewig^eurwig?'
	U.S. postal code or the foreign equivalent. Two or three lowercase letters.
	Optional quality indicator. Quality a is reserved for really stunning photography, suitable for contests or commercial sale; b is a pretty good picture; c is decent; d is poor (the default); and quality code f is reserved only for really horrible pictures that are in the catalog only because they are the only documentation there is.
	Catalog number under the old system. Originally I assigned serial numbers to each roll, so for example “#59.38” would mean roll 58, frame 38. There are a few undated originals with these notations, so I can sometimes use the old catalog numbers to approximate their dates.
	This attribute is obsolete. Originally I was going to record in the image catalog which archive directory contained the reference image. Later I decided to put this information somewhere else; see archx: A program to index a photo archive. Formerly, if this image were contained in one of the standard PNG archive CDs, this attribute is the three-digit archive number. For example, arch-no='003' means the image was contained on the PNG-003 CD. This has been abandoned; bird images are now in www/scans/bird/bird-NNN.
	If the original is film, this attribute specifies the dot pitch, in dots per inch, of the scanner used to produce the digital image. For example, for a scanner that produces 4000dpi, the attribute would be scan='4000'. Omitted for digital originals.
	The children of this element can be any of the elements in the info pattern.

Several elements describing the image can occur as children of the original element.

info = ( loc? & note? & film? & light? & beh? & desc? & pose? )
film  = element film
{ attribute iso { iso-pattern }?,
  text
}
iso-pattern = xsd:positiveInteger
loc   = element loc { text }
note  = element note { text }
light = element light { text }
beh   = element beh { text }
desc  = element desc { text }
pose  = element pose { text }

The content of the film element may be either a filmstock description such as “KR” for Kodachrome, or a digital camera type.

Each of the remaining choices is just a container for text. Their meanings are: