home *** CD-ROM | disk | FTP | other *** search
- Path: sparky!uunet!spool.mu.edu!agate!boom.CS.Berkeley.EDU!lazzaro
- From: lazzaro@boom.CS.Berkeley.EDU (John Lazzaro)
- Newsgroups: comp.robotics
- Subject: Re: Help for Technical Report from CMU
- Date: 5 Jan 1993 00:36:39 GMT
- Organization: University of California, Berkeley
- Lines: 21
- Distribution: usa
- Message-ID: <1ial6nINNk7j@agate.berkeley.edu>
- References: <58199@dime.cs.umass.edu> <C097xD.9y8.1@cs.cmu.edu> <58252@dime.cs.umass.edu>
- NNTP-Posting-Host: boom.cs.berkeley.edu
-
- In article <58252@dime.cs.umass.edu> connolly@piglet.cs.umass.edu (Christopher Ian Connolly) writes:
- >
- >
- >I have nothing against technical reports as a means of rapidly disseminating
- >information, but as you yourself point out, they're not peer-reviewed, and I
- >maintain that they're not consistently accessible enough to cite. Besides,
- >the report in question was from 1986, and if it were of high enough quality,
- >wouldn't it have been published elsewhere by now (see, e.g., Kevin Dowling's
- >post)?
- >
-
- Does someone lose priority on a research idea presented in a technical
- report, if it turns out they are unable or unwilling to publish in a
- journal or conference? This isn't an idle question by any means, I
- believe the original formulation of the backpropagation learning
- algorithm, in Paul Werbos's Ph.D. thesis in 1974, never appeared
- anywhere but in a TR-report thesis. This turn of events caused much
- controversy when the algorithm was independently developed by several
- different other groups in the late 1980's.
-
-
-