Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP Posting-Version: version B 2.10.1 6/24/83; site ucbvax.UUCP Path: utzoo!linus!decvax!genrad!grkermit!masscomp!clyde!floyd!vax135!cornell!uw-beaver!tektronix!ucbcad!ucbvax!upstill From: upstill@ucbvax.UUCP Newsgroups: net.cooks Subject: Recipe Indexing Database wanted Message-ID: <182@ucbvax.UUCP> Date: Sun, 15-Jan-84 00:30:48 EST Article-I.D.: ucbvax.182 Posted: Sun Jan 15 00:30:48 1984 Date-Received: Fri, 13-Jan-84 05:26:00 EST Organization: U.C. Berkeley Lines: 48 As the owner of a large cookbook collection, I am continually frustrated by the process of finding a recipe which uses ingredients I have on hand. What this usually amounts to is a linear search through my current favorite book looking for recipes that contain ingredients I have and none that I don't have. This is always time-consuming, endlesssly aggravating, and is generally unsuccessful i.e. I still have to go to the store for something. Does this sound familiar? Does this sound like a computable problem? Sounds like it to me. I fact I have a pretty good idea of how to solve it. Those of you who have dealt with bibliographic databases know what I mean: there, a reference is found by applying a list of keywords; all references containing all of the keywords are offered forth. Substitute 'recipe' for 'reference' and 'ingredient' for 'keyword' and you begin to get the idea. Let us visualize a conversational system which queries the user as to what hse feels like cooking (genre, cooking time, other imponderable considerations) and then asks for a list of ingredients (omitting, of course, staples which can be assumed to be in stock). The system then spits out a list of recipes tailor-made to the mood of the moment, and the user runs off to the cutting board with nary a thought of the store. Primary design considerations would have to include: 1) Similarity of ingredients (i.e. fresh vs. canned tomatoes) 2) Distinguishing (assumable) staples (salt, flour) from key ingredients (those not in continuous supply or those likely to be special-requested, like sweet red peppers) from wierd stuff one has in the freezer that would get used up if you thought to include it in the query. The trouble with this fantasy is, it needs to be executed, and I am two ways lazy: too lazy to build the system (right) myself, and unwilling to type in all the thousands of recipes (or at least sufficient indexing information) I have on hand. Thus my query: Is anyone else in netland interested in such a system? Does anyone know if one exists? If it came into being, would it be too much to hope that people would individually index one cookbook apiece and share the commonly-formatted results with the rest of us? I am certainly willing to help with the system itself and will definitely be typing in several books if it all comes to pass. (Please note that there are no copyright problems at all with publishing an index of this nature) Reply to me or, in case of epiphany, to the net. From the Anxious Breadboard of Steve Upstill