[oslc-scm] Updated scenario on wiki

[Frank Schophuizen]

On Wed, Nov 4, 2009 at 5:22 AM, Eric Minick wrote:

> For me, the changes between baselines that I'm often most interested in are
> [...]
> between the state of a branch at a given date/time and the state of that
> same branch at an earlier date/time.


I think we should support the scenario to (1) identify a date/time based
configuration for any arbitrary date/time in the past, and (2) be able to
compare two date/time based configurations.


> Besides, in many SCMs, baselines are movable which can render them less
> applicable for build diffing than branch / time combinations.
>

Baselines are per definition immutable. Therefore, "movable baselines" are
not baselines.


> A variant of date/time configuration identifiers would be changeset based
> identifiers. In systems like SVN or Peforce, a global commit ID uniquely
> identifies the state of the source for a project.


I'd rather not call them "changeset based identifiers" because a changeset
identifies only the changes to a configuration, not the entire
configuration. A better name could be "commit identifiers". In some SCM
systems (e.g. SVN), a commit itself may be regarded as a single changeset,
but in many SCM systems one can commit/deliver multiple changesets, which
can be dealt with as individual changesets lateron (e.g. excluding a
changeset from a configuration).

Regards,
Frank Schophuizen
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://open-services.net/pipermail/oslc-scm_open-services.net/attachments/20091104/0d564a37/attachment-0003.html>