class 'intent' documentation

[Erik Bågfors]

2005/12/23, John Yates <jyates at netezza.com>:
> Thursday, December 22, 2005 11:19 PM Robert Collins wrote:
>
> > I've started some documentation about the key classes in bzrlib at
> > http://bazaar.canonical.com/Classes.
>
> The Classes/Branch page opens with this description
>
>   A Branch is a key user concept - its a single line of history
>   that one or more people have been committing to. A single CVS
>   modules' MAIN is a branch, as is 'trunk/module' in svn.
>
> This crystallized for me my vague discomfort with the term "branch".
> To me branch suggests development that is subordinate to some trunk,
> a flavor of fork.  It is a term that focuses on the relationship
> of some work to other work.  The misleading connotations are
> especially jarring when starting a new project or when taking
> about the principal branch in a centralized development model.
>
> While this may often (even very often) be an interesting attribute
> of a bzr branch it is not its essence.  That essence is suggested
> in Robert's definition.  A branch is a line of history or evolution.
>
> Would it be too late to seek more felicitous terminology?  I am not
> sure that I have an ideal candidate.  I do like "line", similar to a
> cell line.  It resonates with the term "mainline".  Is the potential
> for confusion when we are writing lines of code too great?
>


I got the same reaction when showing ppl bzr.  If you want to follow a
'branch'(using bzr branch and bzr pull), at least one person got
confused by the command 'bzr branch', because he claimed that he
didn't want to create a new branch, he just wanted to 'get' the branch
to follow it.

In general, I don't think that there is anything wrong with the
therminology, but it has to be clearly defined.  I think no matter how
you twist and turn branch/repository/archives/etc, there is a chance
for confusion.


/Erik