I’ve been re-reading the Space Odyssey novels and there is a part in 2061: Odyssey Three that I found weird.
Today the whole concept seems absurd. Even in the early days of the Internet, search engines were free and could turn out results in less than a second. So what was Clarke having in mind when he wrote this? Was there an era when computer searches were time consuming and cost real money?
Maybe they had to pay for processing power? Time the computers spent on a particular query?
It was written before 1987, which I think was before any search engines existed. Clarke’s experience with computer searching likely consisted of pay services that searched through private data archives, the creation and maintenance of which was likely somewhat costly.
Well, before my time as a librarian, you had to form your Dialog searches very carefully because they were fairly expensive. Nobody’s around at the desk today who can tell me how much, though.
Remember, the “free” searches you are used to are actually subsidized by advertising. In Clarke’s future, maybe that model didn’t succeed.
I haven’t read the book, but I can say that IT in some organizations does try to charge departments for processing time. Sort of like you might charge back staff time from other support services like maintenance or whatever. It’s not very common, but it is done.
Also, although search engines are wonderfully fast, other kinds of database queries can still take quite a while to run. At my current workplace, we have massive reports that can take several hours to grind. I’m only a database dilletante, so I can’t really explain why that is the case, but a query that runs for 2.5 hours is not unheard of even today.
I don’t know how much equipment Google owns. I don’t know how Google works.
But maybe the author thought that “database search” services could become commercialised and available for hire. He also may have assumed that the sheer number and sizes of the various databases connected was huge, much larger than we have now.
I know that if my single desktop PC, alone, was to try and communcate with every other computer attached to the internet to see if there is the info I am looking for on it (let’s say I want to know which production model F4U Corsair was the first to use a four bladed propeller), it would take an extremely long time.
However, if I can get other computers to look for this info for me at the same time, the work can be divided up amongst them. The more computers working, the faster the search goes.
Possibly, the author evisioned these “search service” computers being avialable for hire. If time is not a factor, but money is, you hire only a few machines. If time is an issue, and money is not, you can rent more computers, and have the requested info faster.
Although in Imperial Earth, published in 1975, Clarke gives a pretty accurate prediction of the Internet, almost casually, when he describes Makenzie using a “communications console” in his hotel room to do research. The search engine isn’t Google, though - it’s a service called the “Yellow Pages” for some reason that’s been lost in antiquity. 
Free to the user, yes. Free to construct, configure, power, support, connect to the internet and maintain? I think not. I believe each of the dozen or so data centres powering google costs about $600million to set up, and a very large amount to run. Fortunately, advertisers fund the whole thing for you.
Well, anytime computers have existed. If you want to do a meaty search through a decent amount of data, you need a decent size system. You’re not going to e.g. scan through a fifteen-terabyte database of retailing data on a typical desktop PC - at least not in any sensible timescale.
Within companies there is always a cost line somewhere for ‘data processing’ or the equivalent. Whether it is charged out to the individual user or department, or as overhead, someone always has to pay for this stuff. Hence the joy and cuddles around things like SETI@home and folding@home - the boffins have figured out a way of getting the public to give them billions of dollars worth of processing for free.
A library search is probably not the best illustration, because it’s a relatively trivial task these days - Clarke was caught out by the speed of advances in computing. Something like that would take seconds and cost fractions of a cent - but someone would still have to pay for it.
When I was working on a mainframe each job card had account information in it, and at the end there was a statement of CPU time used. The account was charged for the amount of CPU time used. (Still is, actually; but I’m not there anymore.) So a job that took .001 seconds CPU was cheap. One that took 2:05:37.338 was expensive.
BRS (Bibliographic Retrieval Services) wasn’t exactly free either, but they at least offered BRS After Dark so students could conduct searches at night at a lesser cost. “Lesser” being about $10-20 an hour, depending on which database you needed, and curiously, the time of day that you printed a citation…
Actually, online searches 20 years ago could be wildly expensive:
(From here.)
CompuServe was also far from free. It was not difficult to run up a couple hundred dollars a month. IIRC, they charged $12.95 per hour for 9600-baud connections in the mid-80’s, leading people to run searches at 2400 or even <gasp> 300 baud, then swoop back in at 9600 to download results.
You still need to pay to search the good archives. Lexus-Nexus charges to search their stuff.
Ahh, kids today. When I was an undergrad you had to pay for computing time in “funny money”. Nobody used computers for free. Everything cost, somehow.
When I first did my computer searches as a grad student I did them through a librarian, and I had to shave my requested keywords dowmn to a carefully-selected few that returned a reasonable number of entries. This took place when I wasn’t around, and resulted in a paper printout that made a fair stack, that I then reviewed carefully. Going back and refining the search, if done, was a similarly slow and painstaking process, and you could only do it as long as the budget held out.
Indeed. Most university time sharing systems charged for compute time and disk space, and you got some funny money at the beginning of the term.
I don’t know how Lexis/Nexis works, but the IEEE paper database seems to have a single charge for unlimited use. I use it through work, which has a site license. If you search for academic papers, you’ll often run into sites which require you to pay for access.
Clarke might have messed this one up, but he did, in Voices from the Sky, predict that in the future there would be no long distance, and that any call would cost the same. (That’s from about 1960.) He got that one right.
The Gopher protocol and Veronica and Archie search engines that used it came into being in the early 'Nineties; before that, online “search engines” were limited to proprietary databases that were generally for-pay and often restricted to subscribers. (The BITNET LISTSERV system had some kind of searchable metadatabase for lists that kept archives or digests, IIRC, but that’s basically like going to a message board.)
Clarke was hardly the first to conceive of a searchable, world-accessible, hypertext style database; in a practical implementation the concept goes back to Vannevar Bush’s Memex, although the Memex was microfilm-based, not networked. (Technically, Paul Otlet’s Permanent Encyclopedia and H.G. Wells conceptual “World Brain” presaged the Memex, but neither were practical implementations of adaptable searchable hypertext data.) And the fact that Clarke’s concept was overtaken by advances in technology is scarcely a unique flaw, either. Heinlein and Pohl (among many others) had spacefarers taking sightings with sextants and performing astronavigational calculations via mechanical calculators or slide rules and logarithms (actually a key plot point in Starman Jones).
Computer and communications technology has advanced so rapidly that even futurist predictions about capability made a decade ago often come up short; who would have imagined in 1990 that it would be possible to carry the equivalent of a massive record collection–enough to fill a large closet–in a device small enough to fit in your breast pocket, or stream high quality digital video to millions of viewers simultaneously on demand. (To be fair, there are many people who have dreamed of it, but few who conceived of a plausible implementation of such technology based upon a linear extrapolation of the capabilities of the time.) I remember when Mosaic (the first mature HTML web browser) came out, and many people said, “Well, it’s neat, but what good is it? Nobody can make money off of this.” Hmmm…
Clarke was actually fairly good–better than most, really–at interpreting the implications and impacts of future technology, even if his guesses about the details of the implementation came up short.
Stranger
When I was in law school, some shady characters came around all the time pushing this crack-like substance–for free. Free and unlimited access to Westlaw and Lexus databases. During summer work and after graduation, few were able to kick the habit, and to this day pay thousands of dollars per case/subject for a few hours of searching. Legal databases. Lawyer crack.
(I don’t practice, so I can’t tell the post - 2005 situation)
In short, their databases are fee-based, either by type of search or type of results (there are a slew of contractual arrangements available). The fee covers a lot of exclusive access.