Nicholas Blackmer, MLIS

To replace an unwieldy and time-consuming card catalog, I created an electronic catalog for the corporate library. Once ready, the FileMaker database was mounted on the company's intranet to make it accessible to all employees.

To populate the catalog, I mostly relied on cataloging records from the Library of Congress, and sometimes from OCLC's WorldCat. On occasion, I cataloged items from scratch.

Aside from author, title, publisher, date and Library of Congress Subject Headings,… Show more

To replace an unwieldy and time-consuming card catalog, I created an electronic catalog for the corporate library. Once ready, the FileMaker database was mounted on the company's intranet to make it accessible to all employees.

To populate the catalog, I mostly relied on cataloging records from the Library of Congress, and sometimes from OCLC's WorldCat. On occasion, I cataloged items from scratch.

Aside from author, title, publisher, date and Library of Congress Subject Headings, etc., I added tables of contents to the catalog records when it was practical. Not only did this give an overview of what was in the book, but it also provided additional keywords. So, for instance, a search on the word "stress" would find not only results where "stress" was in the title or in a subject heading, but also where there was a potentially useful chapter about stress in a book about, e.g., interpersonal conflict, nursing, or health promotion. These titles wouldn't otherwise show up in a search, and users were unlikely to stumble upon these chapters while browsing the collection.

Later, I enriched tables of contents and subject headings with hidden synonyms. So, e.g., if a table of contents or a subject heading mentioned "hand washing", I could add the common variants "handwashing," "hand-washing," and "hand hygiene." Similarly with "sexually transmitted diseases" such variants as "STDs," "STIs," and "sexually transmitted infections." This saved users having to do multiple searches for related terms.

Where I could, I included links to Amazon's previews of books in our catalog, so that users could search within the books. Even for pages that weren't displayed in the preview, the search function would show which pages the search term was on. So, to return to a previous example, a user could find if a book had a discussion of stress, even where "stress" didn't appear in the table of contents, or elsewhere in the catalog record. Show less

I extracted the text of ~3,000 of the company's print publications to create an easily searchable collection of all text from the company’s stock, English-language products. After extracting the ASCII text from QuarkXPress, and later InDesign, files, I cleaned it up in a text editor to make searching easier.

While manually looking through scores or even hundreds of physical products could take days, searching through the text of the entire product line could be done in under a minute… Show more

I extracted the text of ~3,000 of the company's print publications to create an easily searchable collection of all text from the company’s stock, English-language products. After extracting the ASCII text from QuarkXPress, and later InDesign, files, I cleaned it up in a text editor to make searching easier.

While manually looking through scores or even hundreds of physical products could take days, searching through the text of the entire product line could be done in under a minute. Using Unix regular expressions in the search allowed for sophisticated searches that further narrowed down the results (e.g., only products in which “pregnancy|pregnant” and “vaccine|vaccination|immunize|immunization|shots” appeared in the same paragraph).

The enormous time savings made revision planning far more efficient, and allowed us to quickly estimate which products might potentially be affected by, e.g., a new guideline or a drug recall or labeling change.

I continued to maintain this corpus until 2019, updating it as new products and new editions were published. Show less