kurtk@dino (Kurt Klingbeil) (03/13/91)
I am looking for advice on how to approach the following problem: Need to catalogue/search/retrieve text files (business correspondence or gleanings from net traffic for example...) Document size can range from 200 to 64000 bytes Search capability must extend to whole body of text and cannot rely on keywords Performance is not a consideration (within reason... ) Solution should be useable by normal non-hacker-types Solution should be as close to out-of-the-box as possible and not be a major undertaking involving 3 months of coding and tuning. Near as I can tell, DBASE, REFLEX, PARADOX, et al would not take too kindly to this situation. If one wanted to catalogue abstracts/excerpts, the problem isn't too difficult. What I would propose is to roughly organize the documents into directories by 2-4 major divisions (i.e. year, source/dest., subject, sub-subject... ), and then use Magellan to search for relevant words. One could use grep in some semi-user friendly batch files, but Magellan is supposed to do a decent job of pre- and post-massaging the low-level text searches. Any comments? Is this too low-tech to actually work? Any better ideas? Someone suggested that a Concordance of the Bible on a disk, which indexes and cross-indexes each of mega-words would be a more appropriate model ? thanks, k(sup)2
ap@deimos.caltech.edu (Alain Picard) (03/14/91)
In article <1991Mar12.211436.21778@ee.ualberta.ca>, kurtk@dino (Kurt Klingbeil) writes... >I am looking for advice on how to approach the following problem: > > Need to catalogue/search/retrieve text files (business correspondence > or gleanings from net traffic for example...) > Document size can range from 200 to 64000 bytes > Search capability must extend to whole body of text and cannot rely > on keywords > Performance is not a consideration (within reason... ) > Solution should be useable by normal non-hacker-types > Solution should be as close to out-of-the-box as possible and not be a > major undertaking involving 3 months of coding and tuning. > Are we talking mega-amounts of data? If not, (say, 10-megs or less), take a look at Info-Select. It's a free form database, and let's you search by any word. It's a bit primitive, but it does the job, and is extremely easy to use and set up. (And it's cheap ~ $50) Alain Picard p.s. Just a satisfied customer!