Henry Fountain's feature on uses of spider silk in the New York Times lauds the material but is pessimistic about its production:
There has been a big fly in the ointment, however: spiders cannot spin enough of the stuff. Although a typical spider can produce five types of silk, it does not make much of any of them. Obtaining commercial quantities is a practical impossibility -- spiders are loners and require a diet of live insects; some are cannibals. In other words, spider ranching is out of the question.
But is this right? When growing up in the midst of the industrial cornucopia that was Chicago in the 1950s I regularly saw from the elevated tracks a slide rule factory where silk from real spiders was used in making cursors. Maybe some of the more senior scientists interviewed by Mr. Fountain actually used them as students, since electronic calculators did not prevail until the 1970s.
Searching the other Web, I discover I missed an exhibition last year at the Hoboken Historical Museum about the Hoboken plant of the same company, Keuffel and Esser. According to the exhibition page:
One of the displays will explain a fascinating aspect of K&E's Hoboken operations: A spider ranch! Read about how spider web filaments were used as crosshairs for telescopic sights, and the "Spider Lady," Mary Pfeiffer, who ran K&E's spider ranch from 1889 to World War II.
So spider ranching can be done! There's an excellent piece by the historian of science Silvio Bedini here, from catch-and-release in English gardens to recent efforts (2005) to produce spider silk with transgenic goats.