Fields included in current scheme

The current scheme attempts to capture all the distinctions that users of the inventory will want to limit searches by, as well as include other information they would be interested in reading about. See the brief summary of how the inventory addresses each of the original oppositions in the charge.

There is a many-to-many relationship between projects and collections. So the database will also have keys between the two tables, associating projects and collections.

Sometimes, instead of digitizing or creating a collection, a project digitizes or creates just one resource. While "collection" seems like too specific a term, "resource" isn't appropriate when there are collections of individual resources. Furthermore, "collection" suggests selection principles, access points, and other principles of librarianship are at work.

General instructions

All values in Cyrillic must be transliterated according to the Library of Congress system but without connecting diacritics.

Any field whose value has an associated URL (such as a personal or institututional homepage, or a description of a project appearing online) should be recorded in a uniform way. The exact format for recording it must be decided; for now HTML hyperlinks are embedded in the sample records.


Field number	Field name	Repeatable	Data type	Notes
1	Name	yes	free text
2	URL	yes	URL
3	Creator	yes	Name(s): LC Name Authority File
4	Manager	yes	Name(s): LC Name Authority File	Use only if person/corporate body carrying on work is not original creator
5	Participating	yes	Name(s): LC Name Authority File
6	Funder	yes	Name(s): LC Name Authority File	Separate funding organization and grant number by " -- ". Separate grant numbers by "; "
7	Host	yes	Name(s):LC Name Authority File	Use only for websites, when the project is hosted by a person/corporate body other than creator or manager
8	Description	no	free text
9	Goal	yes	controlled vocabulary	access preservation scholarly analysis
10	Digital processes	yes	controlled vocabulary	scanning OCR keyboarding encoding NLP
11	Inception	no	yyyy/mm/dd or: yyyy/mm or: yyyy	Date project was begun
12	Future plans	no	free text	Plans to use new digital processes; plans to include wholy new collections
Collection
Field number	Field name	Repeatable	Data type	Notes
1	Title	yes	free text
2	URL	yes	URL
3	Creator	yes	Name(s): LC Name Authority File
4	Description	no	free text
5	Subject focus	yes	controlled vocabulary	LC headings
6	Geographical focus	yes	controlled vocabulary	Getty Thesaurus of Geographical Names
7	Chronological focus	yes	controlled vocabulary
8	Language of items	yes	controlled vocabulary	ISO 639-2
9	Size of collection		free text	Byte size, number of files, or number of cataloged items .
10	Format of original items	yes	controlled vocabulary	AACR General Media Designations
11	Source type	yes	controlled vocabulary	primary secondary bibliographic
12	Identifier for original items	yes	LC Name Authority File or form generated according to AACR2r2002r for the works, or OCLC numbers for the items being digitized	Not applicable for de novo collections.
13	Location of original items	yes	LC Name Authority File
14	Format of surrogate items	yes	Internet MIME types
15	Metadata/encoding scheme	yes	controlled vocabulary	Dublin Core TEI TEI headers EAD RDF Topic Maps
16	Medium of collection	yes	controlled vocabulary	CD-ROM no medium
17	Web services	yes	controlled vocabulary	OAI Z39.50
18	Access conditions, rights asserted	no	free text	A statement of any access restrictions placed on the digital collection; information about rights (copyrights, etc) held in and over the digital collection
19	Made available	no	yyyy/mm/dd	Date the collection was made publically available
20	Frequency of additions	no	free text	Frequency of additions to the digital collection (such as closed, irregular, daily, weekly, monthly, yearly)
21	Future plans	no	free text	Plans to include new materials in collections

Questions about the scheme

The project-collection (or project-resource) distinction remains problematic. Many digital projects do not distinguish the project from what it creates. If a collection or resource exists without a defined project, we could:

create a minimal project record to correspond to it, having the same name and creator as the collection. This is essentially what we've done so far, forcing a project-collection distinction even when there's no evidence of one.
don't create an associated project record. If we do that, we'll want the Participating, Funder, and Host fields in the collection table as well.

Another approach to the problematic distinction is to have more field values shared between projects and collections, with the values inherited from project to collection unless stated otherwise. For example, the Access conditions, rights asserted field would have one value for the project that would apply to all collections unless stated otherwise.

Do we give current affiliations of personal names or those at the time of creation or administration? If always current, it would be better to store personal and corporate names as a separate entity, so when a name is updated once, it's reflected in all projects and collections with which it's associated. Do we give affiliations for non-academics? Place of work even if a person creates the project outside of his/her job?

Should planned collections and parts of collections be listed in the inventory, without values for the Made available field? This will make searching to see if anyone else is planning to digitize an item easier.

When a personal name is affiliated with an institutuion, give it in form personal name | institution. Should we include all of institutional hierarchy (department, college/faculty/school, university)? Should hierarchy be in parentheses or separated by full stops? Just give whichever part of hierarchy LC gives?

Which fields should be required for every record?

Currently, the controlled vocabulary values are in English, or they're from LC authority files. Will we allow free text (such as for descriptions) to be in other languages? Which languages? When a field value is given in more than one language, should we explictly tag segments as being in one language or another?

Are we interested in recording particular software packages used in projects?