Wrong. If you are introducing a 2-7% error rate into the texts (those are your figures) that is absolutely and completely unacceptable. You are making them WORSE than they were before. Josh ----- Original Message ----- From: Bowerbird@aol.com Date: Thu, 23 Sep 2004 15:49:31 EDT To: gutvol-d@lists.pglaf.org, Bowerbird@aol.com Subject: re: Re: [gutvol-d] New Online Reader

marcello said:

...

Bottom line: if information has been lost, never use guessing to recover it but go back to the source.

wrong.

in an ideal world, with unlimited resources, sure.

but in the real world of project gutenberg now, where no one seems even the least bit eager to _ever_ "go back to the source" and correct that practice of using all-caps to indicate emphasis?

for _that_, a more-practical strategy is called for.

if, by using guessing, the information that you "restore" is right in the vast majority of the cases (e.g., 93-98%), then you _should_ do that. you should also, of course, clearly _mark_ all of the changes that you have made, so the users know they need to be evaluated with care.

then, as i have argued here quite extensively in the past, you need to put into place a robust mechanism that will encourage users to report errors, and then act on them...

but you still want to go ahead and make those changes...

-bowerbird _______________________________________________ gutvol-d mailing list gutvol-d@lists.pglaf.org http://lists.pglaf.org/listinfo.cgi/gutvol-d