The table at the end of this page is rather wide. It prints on European type A4 paper in landscape mode in IE4.

Introduction - An attempt to list the problems

A problem with FTW (and GEDCOM) source references is that a scheme that fits one type of reference, does not necessarily fit another so well. GEDCOM separates the citation from the source document description from the repository location. FTW merges the latter two (generating non-standard GEDCOM in the process) and printed reports treat the combined text with the citation as one source reference.

Sources can be categorised into three broad types—Public archive, Published and Unpublished.

With published works, there is a good document description but no repository though the field could be used to record where you found the book, particulary if it is rare.

With archive material the document description can be poor but the repository is known. Also each, of perhaps many, documents may need a separate Master Source because the call number for each is unique. Whether this is necessary will have to be judged by the user. The definition of a “source document” is vague, varying from a single will, a manuscript book or a whole archive such as a census. Some of what were true archive sources are now either published or available in multiple repositories. This can result in multiple Master Sources for one “document”.

A lot of genealogical data is unpublished and, initially at least, undocumented. It can include interviews, letters, emails etc. With the addition of online data, transcripts, indexes and third party reports the scheme can get out of hand if not properly planned.

A true citation scheme will quote the actual source that the information was received from i.e. if Mrs. X looked up a record and sent the details to you, then the source should be “Mrs. X quoting Record Source”. However, I find it more practical to cite it as “Record Source, via. Mrs. X”. Often a mixture would be appropriate, but the main requirement is that the reader should be able to locate the same reference, though with unpublished sources even this could be difficult as, in many cases, the archive is private to the author (you).

One of many possible schemes

The objective of this scheme is to come up with a workable plan whilst balancing the number of Master Source Titles aganst the number of citations per title. There is also a need to avoid duplication of information but to keep the Master Source Titles unique. Some titles may need to be subdivided (e.g. Census into states/counties) if sufficient citations warrant the expansion.

The example scheme here is the one I have designed for myself and is rather biased towards UK research. It should, however be adaptable for other locations and researchers.

Type Title Page Text Author Publication Facts Source Media Call No. Source Quality Source Location Comments
Book <Title> <Page> <Author> <Publisher> - <Date> Book
CD-ROM <Title> <File> <Author> <Publisher> - <Date> Electronic
Fiche <Title> <Frame> <Author> <Publisher> - <Date> Microfiche
Facsimiles of published works should cite the original with details of facsimile in comments.
Published indexes, facsimiles and transcripts of archive sources should be cited in their own right.
WFT Family Archive CDs could be listed here, but probably work better as unpublished data like Ancestral Files.
GRO Index General Register Office of England & Wales index of <type> <Quarter/ Year> Public Record Office Civil Registry Family Records Centre 1 Myddelton Street, London. EC1R 1UW (and widely available on microfiche elsewhere)
Newspaper, Journal and Periodical <Publication name> <Date> - <page> - <Article title> <byline> <Publisher> Newspaper <Archive>
For cuttings, show the source of the collection (unless a published work - see above) in the citation text.
IGI International Genealogical Index (R) <Batch/ Film numbers> The Church of Jesus Christ of Latter-day Saints Copyright

© 1980, 1997, data as of February 1997

Electronic/ Microfiche Family History Library 35 N West Temple Street Salt Lake City UT 84150 USA (and widely available online and on microfiche elsewhere)
Details as supplied by downloaded GEDCOM files. This doesn't quite fit the scheme, but is practical.
Church register <Type> register - <church name/location> Church record <name> Record Office
Will <Name> - will <Date written> <name> Probate Office
Admon <Name> - admon <Date granted> <name> Probate Office
Census GB Census GB - <year> <Piece/ Folio/ Page> <Place & Text> Public Record Office Microfiche/ film HO107 or RGx Family Records Centre 1 Myddelton Street, London. EC1R 1UW (and available elsewhere)
The 1881 census is available on CD-ROM & Microfiche by the LDS - treat this as a published transcript.
Census US Census US - <year> <Place & Text> Microfiche/ film National Archives & Records Administration
If reference was to the Soundex index only, this should be noted. Published facsimile and transcripts should be treated as such.
Census (other) Census - <Country> - <year> <Place & Text>
Certificate <Type> Registry & index details Rest of text Official document Private archive
Monumental/ Tombstone Inscription MI -<burial ground> <location> - <date seen> Tombstone <Address of burial ground>
A photograph is valuable corroborative evidence as stones can vanish without notice!
Research Archives often include unpublished research files and notes. Treat these as unpublished sources (see below)
Email <Author> - email <Date> <Author> Unpublished Electronic Private archive
Letter <Author> - letter <Date> <Author> Unpublished Letter Private archive
Extensive information from one document may require a dedicated master source. In this case, include the date in the title.
Personal knowledge <Name of informant> <Date> <Name> Unpublished Interview
The informant could be yourself!
Ancestral file Ancestral file <Submitter> - <Date> The Church of Jesus Christ of Latter-day Saints Electronic Unknown
WFT Family Archive CD WFT Family Archive <CD #><Tree #> <Submitter> Bröderbund/ Family Tree Maker Electronic Unknown
Website <Title of site or page> <URL> - <date extracted> <Author> Electronic Unknown
Official websites could be classified as published material, but in general they are so transient that this category suites them better.
Derived Derived, calculated and estimated dates should show the method of calculation in the citation text. Estimate
Misc. <Author> - <archive> or <Title> if given <Date/ Ref.> <Author> Unpublished Unknown <Archive>
This category includes unpublished research files found in archives, libraries etc.
For undocumented material, cite it under email/letter/personal knowledge headings.

Where multiple events are detailed in one citation, then I put the full text on the primary event and in the remaining events use the same title and “see <event>” in the citation text. In some cases “see <individual> <event>” is needed where a citation quotes a source promarily about another person. Some long citation texts work better as subsidiary notes, in which case a similar cross reference can be used. The end result must of necessity be tailored to a particular research requirement and style. This is up to you.

For material obtained from a third party, cite it under the original source but quote “via <Name> <media> <date>” in the citation text. The same applies to unpublished (or web-published) transcripts and indexes and to undocumented material obtained from a third party (hearsay)