Re: [Gnu-arch-users] Re: Online book for usability

[Juliusz Chroboczek]

JB> The reason that reverting changes to 'a' single file(s) yet is that
JB> _you_ haven't provided the code to do it -- yet.

True, but perhaps not quite the point.

JB> With regards to switching to a different version, why wouldn't you just
JB> get a new working copy?

I do, when it's convenient.  I switch when it's not.

JB>   2. The capability that the user wants is already there, and (s)he
JB>   doesn't know about it yet.

That's not the case I'm describing: it's about the capability already
being there, but being distributed across half a dozen primitive
commands.  This is both annoying and intimidating: quite often the
individual commands make no sense to the user, so you have to either
learn the sequence by rote or write a script.

Other examples include:

  archive-mirror, update
  register-archive, archive-mirror, update
  archive-setup, tag, set-tree-version

On the other hand, there are many cases in which useful high-level
commands exist although the same result could be achieved with
multiple low-level commands.  For example mv, archive-setup, heck,
even update and replay are not primitive.

I hope that clarifies what it is that annoys me with tla.

                                        Juliusz