On Fri, 2007-11-09 at 13:41 +0100, Bernd Eilers wrote: > Hi there! > > Kohei Yoshida wrote: > > On Thu, 2007-11-08 at 16:47 +0100, Bernd Eilers wrote: > > > >> Anyhow you certainly don´t want to break that semi-automatic process to > >> generate Release Notes and volunteer to offer to parse a few hundred > >> specifications in a few hundred different plain text based formats > > > > So, is that true that we have a few hundred new features going in to > > each release? If that's true, I'd agree for the need of automation but > > I have my doubt that that is really the case. Correct me if I'm wrong. > > > > Just have a look at > http://development.openoffice.org/releases/2.3.0.html to get an > impression, these are the actucal release notes for the 2.3.0 user Release. > > And that´s the result of what we currently semi-automatically generated > for the 2.0.3 Release by parsing (or not being able to parse) > specification documents. > > Note that everywhere where there is a "feature-info:" in the "Spec. > abstract" column of the Release Notes the process used a dirty fallback > to use information from the feature announcement mail instead of using > information from the specification because either a specification was > not available at all or was not usable because it was not conform to one > of the specification templates.
I wouldn't call that a "dirty fallback". It's a very useful fallback IMO. :-) > Our aim should IMNHO be to keep those > "feature-info:" fallbacks small and use the stuff the specification > writers wrote instead. Currently that´s the first paragraph of the > abstract which will be copied to the Release Notes interim document. > > The generation of the interim document checks which issues have been > fixed for all ChildWorkspaces that have been integrated between this > Release and the last Release and which corresponding > feature-announcements and specifications we have. Keeping track about > all that without some kind of automation is kind of hard, don´t you > think so? Sure, I'm all for keeping track of new features. But I'm against using a means of public humiliation to make us comply with that requirement. I believe in entrusting the developers to do the right thing, and if one of the developers overlook that requirement (maybe he/she was being swamped with tasks, the feature is not yet complete, etc), you can always ask that developer privately. There is usually an explanation for that. Of course, I don't know the culture inside Hamburg, so things may be different on your end. But I'm just speaking for myself. Kohei --------------------------------------------------------------------- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]