3 ms·
There are a few estimates of total books ever published. Those range from the 40 million or so within the US Library of Congress collection to a Google estimat
by dredmorbius 5y ago
There are a few estimates of total books ever published. Those range from the 40 million or so within the US Library of Congress collection to a Google estimate from a few years back of 140 million books existing, with about 2--5 million new ones published per year.
Bowker, the US issuer of ISBNs, has registered roughly 300k new titles annually, since at least the 1950s, and beginning in the aughts, about 1--2 million "nontraditional" (self-published) titles.
In straight text, assuming 250 pages of 250 words at 8 bytes/word, 40 million books works out to about 20 TB. This is a likely close to the lower bound as there are texts with image and other content which bump this value up. It works out to about 0.5 MB/book, which in my experience is too low.
At 150 million books, we'd get 75 TB, and at 200 million, 100 TB. Nice round numbers....
A more realistic per-book size is 5 MB. This is typical of a mostly-text, formatted PDF. At 40, 150, and 200 million texts: 250, 750, and 1,000 TB.
A data-heavy / graphics-heavy book might weigh in at 20--100 MB, though those tend to be less common.
Scanned-in books with both page images and OCR text typically run about 5--10 MB. The largest I've seen (a copy of Lyell's Geography was >300 MB.
Apple has offered iPads with > 1 TB onboard storage for at least several years. This seems to be the higher end of available tablet storage. (For some reason, the Cloud-and-Surveillance-oriented Google skimps with many devices still at 16--32 MB. Remarkable, which also went Cloud, tops out at 16 MB.) If fully devoted to texts at 5 MB each, that would offer storage for up to 200,000 documents, roughly the size of a well-stocked municipal public library.
As to how many public-domain works might exist, the 1927 Report of the Librarian of Congress notes that the Library had assigned 4,503,164 copyrights over the preceding 57 years (presumably 1869--1926), the time in which such registrations were the work of the Library. Those works should all presently be in the Public Domain, and would total 22.5 TB at 5 MB/work.
Annual report of the Librarian of Congress yr.1927, Washington, Library of Congress; for sale by the Supt. of Docs., U.S. Govt. Print. Off., p. 22.
https://babel.hathitrust.org/cgi/pt?id=uc1.a0013520523&view=1up&seq=42 https://babel.hathitrust.org/cgi/pt?id=uc1.a0013520523&view=...
The Library of Congress's total holdings was 3,420,345 million as of end of FY 1926 (p. 23 ibid). Presently, works published prior to 1 January 1927 in the United States are in the public domain.
An additional roughly 1 million maps, 1 million volumes and/or pieces of music, and ~500k prints were also in the library's collection as of 1926.
A full set of the Librarian's reports, dating to the 1860s, is available at Hathi Trust. From at least the 1950s, annual holdings and aquisitions are broken out by Library of Congress classification, which makes for some interesting data, for those who find that sort of thing interesting.
http://catalog.hathitrust.org/Record/000072049 http://catalog.hathitrust.org/Record/000072049
- wolverine876 5y agoThank you! We are not far in time (10 years?) from being able to provide a complete set with every computing device, though we would need them all converted and formatted, etc., and many will (unfortunately) think we should just use the cloud instead. I will add this study published in 2010, which estimated 129 million 'editions' published in all languages. Any text can have multiple editions (I don't recall how 'edition' was defined, but IIRC the study was of published objects not of texts), and "includes serials and sets but excludes kits, mixed media, and periodicals such as newspapers". It also contains a useful description of methods to bulk process bibliographic data. (Though maybe this is the Google study you talked about, because IIRC their data is involved.) J-B Michel, et al, "Quantitative Analysis of Culture Using Millions of Digitized Books", Science (16 Dec 2010) https://www.science.org/doi/10.1126/science.1199644 https://www.science.org/doi/10.1126/science.1199644 By the way, I greatly appreciate your informed comments. It's some of the most substantive and interestin reading on HN.