In the subsequent stages of transcription and building of a computer database, the opportunity for additional deviations arise. Even though there may be an attempt to correct the "errors" of the previous stages during this process by relying on knowledge and experience outside of the individual source record itself, there is also the possibility that new errors will be introduced.

The database used in this webGED DBQuery system was contructed from the base of a previously existing database that has been widely distributed. Its origins are obscure, although it has been said that it was contructed by students funded by a cultural grant from the Province of Nova Scotia. While researchers must be extremely grateful for the work that made this information more widely available, there is every indication that little attention was paid to data integrity throughout the process.

Over 10% of the records were found to have "errors" of some sort. Many were typographical errors that, although they would be relatively easy for an experienced Arichat researcher to spot in the context of the individual record, might cause the record to be missed completely in a computerized search. Others were even more damaging; e.g. the copying of some names from the previous entry, accidental insertion of irrelevant material in the middle of fields, etc.

In the preparation of this webGED DBQuery database, considerable effort was made to identify and correct these errors. This, of course, added yet another layer of interpration to the final product. In spite of the effort involved, there should be no doubt that many such errors remain unedited and that some new ones were probably introduced in the process. On balance, however, it is believed that the file integrity has been improved considerably.