DDC, code and architecture

In the chapter “Reuse Versus Compression” of his book Patterns of Software, Richard Gabriel talks about how full-scale reuse in software development—within organizations as well as between them—would require librarians:

[T]here has to be a means of locating the right piece of code, which usually requires a good classification scheme. It does no good to have the right piece of code if no one can find it. Classification in the world of books, reports, magazines, and the like is a profession, called cataloging. Librarians help people find the book. But few software organizations can afford a software cataloger, let alone a librarian to help find the software for its developers. This is because when a development manager has the choice of hiring another developer or a software librarian, the manager will always hire the developer.

A Dewey Decimal Classification (DDC) system for reusable code. This reminded me of the idea of the software architect as librarian.

The DDC seems to me to be something similar to a hash table: in trying to find a particular book, the segment of the entire library is identified, by deciding which category the book belongs to, and then one goes there to search for it. However, not all books obviously belong to a single category.

Today, I guess that the classification systems used in libraries are at a lower level of abstraction: Visitors to a library would generally use computer terminals to search for the book, then use the category key to look it up. Other systems of locating books can be layered on top of this sometimes ambiguous system. How the books are organized at this lower level is no longer that important: what matters is that the subcategories aren’t too big, so that once one finds the shelves for the subcategory, one would search sequentially on the author’s last name or the title.

Got to go.