Are there any man pages for any of the system error messages?
in [General Linux]
|
Prev: Live CD to scan windows pc for a virus
Next: where is CentOS storing my hostname that it resets it on each reboot?
From: Baho Utot on 7 Apr 2010 19:52 Darren Salt wrote: > I demand that Mark Hobley may or may not have written... > > [snip] >> To get an explanation of what went wrong, the user will now simply type: >> man e c2 > >> This will pull up the man page for the c2 error message. > > No; one or both of e and c2 or, if neither is found, e-c2. > Wasn't it e=mc2?
From: Mark Hobley on 7 Apr 2010 19:37 unruh <unruh(a)wormhole.physics.ubc.ca> wrote: > Someone who knows so little that they do not know that in the command > "ls fred" the word fred is supposed to be that of some file or directory > is not going to know enough to know that "man" even exists. No but they only need to ask once :) The name fred may have come from a utility program or script. They may not have provided this name themselves, which is another reason why they may not know that fred is a filename. Look at this error: OPENGL_DEBUG GL_ERROR (1280) 'invalid enumerant' in: Universe before doing anything Is Universe the name of a file? What about this one?: No template file exists for templating "user" with template "forgotten" Is user the name of file? Presumable "forgotten" is a filename. Here is another one: iptables v1.2.5: can't initialize iptables table `filter': Table does not exist (do you need to insmod?) Is filter the name of a file? Mark. -- Mark Hobley Linux User: #370818 http://markhobley.yi.org/
From: despen on 7 Apr 2010 21:12 markhobley(a)hotpop.donottypethisbit.com (Mark Hobley) writes: > Robert Riches <spamtrap42(a)verizon.net> wrote: >> On Mandriva 2010.0, this command >> >> ls fred >> >> produces this output >> >> ls: cannot access fred: No such file or directory >> >> A user should know that the name of the file in the middle of the >> message is a dynamic part of the message--that's just common >> sense. > > A use may not know that fred is the name of a file. And they are even less > likely to know that fred is the name of a file that does not exist. > > On a system utilizing Universal Error Messaging Infrastructure, > the message will probably appear as something like: > > Error: C2: No such file or directory: fred So where is the message number from "ls"? You have perror producing a C2 prefix, but "ls" just bleeping out "Error". What happened to "/bin/ls cannot access"? When IBM embarked on it's mission to number all it's messages, they realized that applications running on their systems needed to pump out messages too. So they started out reserving messages starting with "I" to IBM. IBM knew they wouldn't get far with just "Ix" so they started out with 3 letters. For example "IOS001". Of course, IBM long ago started using prefixes all over the place and random error message number formats. Today, IBM pretty much clings to numbered messages, but one of their best products dispensed with numbers entirely. In the primary development environment for IBM mainframes, ISPF is designed to first produce a short (20 character or so) message, then if you don't understand, you press PF1 (the help key) and get a long message (usually 80 characters but it can be more), then if you press help once more and you get the actual full documentation. Now, that's my idea of a solution.
From: Mark Hobley on 7 Apr 2010 22:46 unruh <unruh(a)wormhole.physics.ubc.ca> wrote: > I think he is worried that the error messages are too terse and would > like longer error messages which would, I suppose, give the user a > complete lesson in the operation of computers. Not necessarily. Take the pango example. Some people might know what that means, so the message may be sufficient. But other users may need to refer to technical documentation to determine the specifics. By tying the error message to documentation, we can accomodate all users. It also means that if someone else writes a document. > That he would go through the 100000 programs and find and classify all their > error messages I find ludicrous to contemplate. I will just do packages that are of concern to me. I hope other developers will do the rest. Obviously the more people using the infrastructure, the faster this project will grow. > That writers of software would apply to him for error message numbers so > they could report error messages with numbers I also find ludicrous to > contemplate. If the author cannot write a readable error message they sure > are not going to take the extra 10-15 hours to apply to him for an error > message number, write an whole description of the error message and submit > it back to him for inclusion in his man page. I think I would be surprized > if more than 5 developers would do so. That's up to the individual authors. I am not in command. I think we should make software as good as it can be, and the documentation as good as it can be. (Even Microsoft got their act together recently, and their error codes are now available through MSDN, together with associated documentation). This project provides a facility to share documentation, because an author may find that an existing reference document is already available that describes how to overcome the error that his software is producing. > If he were to try to persuade writers to produce better error messages, > that might be worthwhile. If he could persuade developers to even write > documentation ( eg man pages) that would be worthwhile. Indeed. I would love to see this happen. I am making the infrastructure available for this. > But this project is an exercise in futility, and I think he has not the > ghost of an idea of what it would involve. I know what it involves. I have developed and used such mechanisms in the past. FWIW, this project is a by-product for me. I need to output some error messages from some software that I am modifying (Currently the author has hardcoded messages, that do not have any documentation relating to them). I also know that I may need to output error messages from future software that I write, so I will need appropriate infrastructure for this. I also have a load of notes that I have made from software problems and issues that I have had with various software packages in the past. These need to be organized and placed into manual pages. These notes will help to form part of this infrastructure. I actually developed similar infrastructure for this over 20 years ago, so this is just rework for me. I would like developers to start writing man pages though. Mark. -- Mark Hobley Linux User: #370818 http://markhobley.yi.org/
From: Robert Riches on 7 Apr 2010 23:51
On 2010-04-07, Curt <curty(a)free.fr> wrote: > On 2010-04-07, Robert Riches <spamtrap42(a)verizon.net> wrote: >> >> On Mandriva 2010.0, this command >> >> ls fred >> >> produces this output >> >> ls: cannot access fred: No such file or directory >> >> So, I put this in Google's search term box: >> >> "ls: cannot access" "No such file or directory" >> >> It got 82,700 hits. >> > > I googled the entire quote, fred included, and got 53,800 hits. Searching for the whole phrase/sentence/message (including fred), with quotation marks around it, Google returns _FOUR_ hits. Searching for the individual words in any order yields 53,300 hits. I suspect you forgot to put quotation marks around the phrase. -- Robert Riches spamtrap42(a)verizon.net (Yes, that is one of my email addresses.) |