Date: Thu, 17 May 2001 08:44:01 +0200 From: Roelof Osinga <roelof@nisser.com> To: Mike Smith <msmith@freebsd.org> Cc: stable@freebsd.org Subject: Re: Administrative tag a possibility? Message-ID: <3B037331.7850BE61@nisser.com> References: <200105170543.f4H5hvZ01232@mass.dis.org>
next in thread | previous in thread | raw e-mail | index | archive | help
Mike Smith wrote: > > > ... > > Say introducing a convention like '$!$' to denote an always to be > > acceptable change? > > No, not acceptable. > > In this case, mergemaster should be smart enough to notice that the file > in question was unchanged before the merge, and just not irritate you in > the first place. You've just won my vote! Nonetheless I think a simple pattern match like "$!$" could've been implemented long ago - thus saving me uncounted q+i keypresses, whereas introducing smartness into a program might take a tad longer. In the end it amounts to the same. In fact, if mergemaster where truly that smart it would stand up and point it out to you. Noticing a file to be unchanged when in fact it has been changed is a contradiction in terms. What you would like it to do is noticing that the change in question is restricted to an insignificant, read administrative, aspect of the file at hand. Yet how would it recognize which areas are to be deemed insignificant, not to say administrative? It would need knowledge to base its decision upon, that's how. So how to teach it what is knowledge and what not? How to recognize a thing for what it is when it stands up and kicks its petoeyee? My suggestion was To Keep It Simple, ehm, Silly! However, if you think that too simple, I would be interested in hearing your solution. Pointing to neural networks I think mere obfuscates the issue. It hides the answer into a distributed knowledge weight mesh, if it were. Not by assignment but by indirect implication. Like genetic algorithms hide or distribute knowledge into indirectly determined algorithms. You have no control. You get presented with an answer. Ok, using neural networks you can, if so inclined, deduce what's going on by analyzing the matrix network. But with genetic programming, well, I have not read enough to say. But I suppose it'll boil down to backtracking one heck of a lot of algorithmic manipulations of algorithmic manipulations. Personally I'm inclined to go with Sowa's Conceptual Structures, even though that means accumulating a vast body of knowledge, not to mention some procedural means of manipulating it (+++). At least it provides you with a direct means of grasping the issue! But let's hear your side of the story. Roelof -- _______________________________________________________________________ eBOAź est. 1982 http://eBOA.com/ tel. +31-58-2123014 mailto:info@eBOA.com?subject=Information_request fax. +31-58-2160293 To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-stable" in the body of the message
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?3B037331.7850BE61>