home *** CD-ROM | disk | FTP | other *** search
- Comments: Gated by NETNEWS@AUVM.AMERICAN.EDU
- Path: sparky!uunet!gatech!paladin.american.edu!auvm!ALPHA.BLDR.NIST.GOV!KOONTZ
- X-Envelope-to: NOTABENE@TAUNIVM.bitnet
- Message-ID: <QADEC90C@pc-koontz>
- Date: Fri, 16 Oct 92 07:52:11 LCL
- Sender: Nota Bene List <NOTABENE@TAUNIVM.BITNET>
- From: "John E. Koontz" <koontz@ALPHA.BLDR.NIST.GOV>
- Subject: Re: Report on Indexing
- Newsgroups: bit.listserv.notabene
- In-Reply-To: <01GPZL2C1B80001WFT@ENH.NIST.GOV>
- Lines: 8
-
- In regard to Robert Lavenda's comments on indexing, it is typical of
- indexing routines to have a limit on the amount of material they can handle,
- or the number of distinct tokens. This is usually gotten around by
- requiring the user to run several indexing operations, and then a merge
- operation to combine the indexes produced. I suspect this has to do with
- the algorithms used in indexing, and does reflect, indirectly, memory
- constraints. But, if the index programmers would get smart enough to hide
- the multiple indexing and merge step from the users, we'd all be happier!
-