[svn.haxx.se] · SVN Dev · SVN Users · SVN Org · TSVN Dev · TSVN Users · Subclipse Dev · Subclipse Users · this month's index

Re: Command-line verbosity

From: Matthew Braithwaite <matt_at_braithwaite.net>
Date: 2000-10-27 20:09:18 CEST

On Fri, Oct 27, 2000 at 10:29:51AM -0500, Karl Fogel wrote:
> >
> > But that aside, the argument that ``verbosity makes it easier for the
> > user to get bitten in the ass'' is more important than ``verbosity
> > makes the user happy'' IMHO.
> I don't see how verbosity makes it easier for the user to get bitten
> in the ass. Can you describe a scenario? My whole point in
> supporting the progress reports is that I think they're the best
> default.

I understood Jim to be saying that, by producing status information as part
of `cvs update', CVS encourages screwing up by (a) increasing the likelihood
that important stuff will scroll off screen and (b) increasing the likelihood
that important stuff will be concealed amidst unimportant stuff.

An example value of ``important stuff'' is a conflict or merge, an example
value of ``unimportant stuff'' is ``Updating in foo/bar/baz''. An example
problem might be, as happened to me the other day, not realizing that a bug
was due to a merge of someone else's code, because I was unaware that a
merge had happened.

Whether stuff like successful merges should be part of the default output
can be sensibly debated, but I definitely think that output that serves
only to remind the user that the program is doing its job is detrimental --
it decreases the attention that the user pays to the output overall, which
increases the chances of missing something important.

I think it's a terrible thing to have a program that always produces
output. Nothing draws attention to a program's output so much as the
knowledge that it usually runs silently. I'm definitely in the ``if
cvs update modifies no files it should produce no output'' camp.
Received on Sat Oct 21 14:36:13 2006

This is an archived mail posted to the Subversion Dev mailing list.