Re: Using VC for change descriptions

[Joseph Myers]

On Wed, 10 Jan 2018, Richard Stallman wrote:

>   > I think we are going ever further into the wrong problem.  My hypothesis 
>   > is that for the vast bulk of changes, a list of changed entities is never 
>   > going to be of use in future development 
> 
> In my personal experience, this was very useful.  Your opinion will
> not trump my experience.

In my personal experience, this list is never useful.  The *inverse* 
problem often comes up - given an entity, listing the commits responsible 
for the present state of the code - but that's just "git blame".  (That 
git blame won't mention commits that only deleted lines of code without 
adding anything is not a problem in my experience, but of course there are 
other interfaces such as git log -S and -G that can be used to find when 
something was deleted.)  And if my attention has been drawn to a 
particular commit, "git show" to look at the diffs is sufficient for me 
without needing a list of entities.

What are the purposes for which you find that list of entities useful?  I 
strongly suspect it's a matter of different workflows people are used to - 
for any purpose for which the list is useful, there are other workflows 
that are just as good that do not need the list.

But if it's a matter of some projects tending more to have changes for 
which the list of entities is useful then the standards should reflect 
this, so that projects for which that list is not found useful do not need 
to maintain it.  Certainly typical changes in glibc or GCC are very 
different from the sort of changes for which example ChangeLog entries are 
given in the standards (larger, more complicated, involve more pieces that 
need changing together but aren't individually of particular 
significance).

-- 
Joseph S. Myers
address@hidden