From owner-freebsd-hackers Tue Jan 17 17:46:07 1995 Return-Path: hackers-owner Received: (from root@localhost) by freefall.cdrom.com (8.6.9/8.6.6) id RAA02867 for hackers-outgoing; Tue, 17 Jan 1995 17:46:07 -0800 Received: from is1.hk.super.net (jbeukema@is1.hk.super.net [202.14.67.232]) by freefall.cdrom.com (8.6.9/8.6.6) with SMTP id RAA02859 for ; Tue, 17 Jan 1995 17:46:04 -0800 Received: by is1.hk.super.net id AA17513 (5.67b/IDA-1.5 for freebsd-hackers@freefall.cdrom.com); Wed, 18 Jan 1995 09:45:43 +0800 Date: Wed, 18 Jan 1995 09:45:43 +0800 (HKT) From: John Beukema To: Terry Lambert Cc: Garrett Wollman , freebsd-hackers@freefall.cdrom.com Subject: Re: CVS stuff In-Reply-To: <9501172230.AA28575@cs.weber.edu> Message-Id: Mime-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII Sender: hackers-owner@FreeBSD.org Precedence: bulk On Tue, 17 Jan 1995, Terry Lambert wrote: > > > What does siting message by number have to do with not citing message > > > by English Message Text? !a => b? That's a logical fallacy. > > > > Because if you're trying to figure out what the !@#$%^ is wrong when > > somebody complains, you need to know what the REAL message is. You > > can either put the messages in English, so that everybody understands, > > or you can stick on message numbers. > > Maybe I can put it in all caps, with spaces between the letters, because > E V E R Y O N E U N D E R S T A N D S E N G L I S H I F I T > I S S P O K E N L O U D L Y A N D S L O W L Y E N O U G H. > > Personally, I don't have any intrinsic objection to message numbers, if > they are accompanied by text, if the intent is to index a catalog for > a program maintainer. I'm probably in a minority here, even though this > is exactly what exit codes are. > > > > > It makes no sense to be able to display Russian, for instance, if none > > > of the data on the system is Russian text! > > > > If you want some Russian text, you go into your editor and generate > > it. I'm certain that Andrew has lots of Russian text. This does not > > mean that it is a good idea to have system programs printing out > > messages in languages which are not understood by the program > > maintainers. > > It's no skin off my nose, but, I just want to be perfectly clear on what's > being said here, so that I don't misunderstand. > > > There are several implied assumptions in your statements: > > 1) The people having the problems will not be able to diagnose > the problem from the error messages, even if they are in the > user's native language (pretty shitty error messages, if you > ask me). > > 2) The people having the problems have some way of contacting > and communicating with the authors about the problems they > are having (Stahlman's "everyone will be on the net", a > demonstration of the failure of modern academia to live in > the real world, where software is distributed on CDROM to > people who don't even own modems). > > > Now, these have several correlaries: > > 1) If Pierro, for instance, takes over maintenance of csh, > then it is acceptable for an English spekear to type the > command > > moose > > at a prompt, and instead of getting back the message > > moose: Command not found. > > he or she should get back the Italian equivalent, since it > is more important that Pierro be able to resolve the problem > for the person instead of the person being able to resolve > the problem themselves. Same thing for: > > moose: bus error, core dumped. > > > 2) If Andrew or some other Russian language speaker takes over > maintenance of getty, then it's OK to see a KOI-8 version of: > > Login incorrect > > When most users are probably running ISO8859-1 fonts anyway, > since as the maintainer he should be able to understand when a > user is providing an invalid login/password so he can tell them, > and KOI-8 encoded Russian "login failure" messages in their > system log files. > > > It seems to me that this embedding of translation facilities in the > program maintainers is wrong. If the maintainer can't understand the > error, then the maintainer should ask the person reporting the error > to use a particular locale and cause the error to occur again. Or > even better, put "Notice! use the C locale to generate error messages" > in the same place you put the maintainer's email address. > > > The whole point of an error message is to provide sufficient information > to the user that they don't *have* to contact the maintainer, IMO. An > error message is a mechanism for reporting a condition over which the > user has control but the program does not. An error like: > > Error: PI is 3.1415926 > > Is useless, since what the hell is the user supposed to be able to do > about that? > > > > If the "Lingua Franca" of BSD is to be English, fine, but don't cloud > the air with BS and hand-waving instead of just stating the decision, > and if that *is* the decision, be prepared to back it. > > > Terry Lambert > terry@cs.weber.edu > --- > Any opinions in this posting are my own and not those of my present > or previous employers. > It would make life a great deal easier for those of us overseas if the English string messages could be isolated in a string table. Source code could be commented and/or meaningful int #defines provided for debugging (#define 11 BUSERR_COREDMP. err_msg[11] = "Bus err core dumped") The set of English strings could be translated by anyone wishinbg to internationalize a version just by setting an alternate string table. Messages could be improved (in any language) by updating the string table. jbeukema