08:16 < abentley> Better phrasing: 'what circumstances should cause us to produce a working tree in a repository branch'?

[Denys Duchier]

Aaron Bentley <aaron.bentley at utoronto.ca> writes:

>> I don't see why you believe that separate commands are easier than separate
>> options.
>
> It's harder to type.  The help is harder to read.  It means one command
> can do radically different things, including pointless things like "bzr
> create --repository --checkout"

I don't believe that

	bzr create --repository --clone ../FOO

is so much harder to type than:

	bzr repository
        bzr clone ../FOO

or whatever the commands might be; especially if we make short options
available:

	bzr create -rc ../FOO

as for the non-sensical combinations, you'd still have to detect them
regardless.

Just imagine that under the hood, you are indeed using different commands.  I
just don't think that multiplying the commands at the UI-level is a good idea.
It seems to me that advertising "bzr create" as the command to create things
makes it cognitively simpler for the user.

>> I think it is easier for a user to remember that there is one command to
>> create things, and that makes it also easier to lookup the copious --help that
>> should be associated with it.
>
> People don't like copious help.  They like direct, to-the-point help.

I think you are overgeneralizing on your personal typical use case.  People like
terse help when they only wish to be reminded of something they already know,
and more extensive help (especially including examples) when they either don't
know or can't remember.  Ok, I am not "people", but I know that's the case for
me.  So I'd probably like "bzr create -h" to print the terse help and "bzr
create -h -l" to print the more extensive help.

--Denys