When writing or substantially rewriting an article, I would sometimes like a space where I could record certain "design decisions" made during creation of a paragraph. I think this could help future editors to sensibly continue my work.
What we now have is:
* The version log, which favors short decriptions of the changes, but not (perhaps longer) arguments *why* something was changed.
* /Talk can be used for my purpose, but convention has it that it is mainly for discussion, i.e. when someone (not knowing my thoughts) executes the exactly opposite decision I can bring this up on /Talk.
What I envision is that the edit page not only offered the option for a one-line log message, but also a multi-line argument section. As long as the change is diff-able (at least two weeks, hopefully forever), both should be viewable somehow. The Recent Changes should list the short What message as it does now.
Example:
Summary of changes: [rewrote the first paragraph] Why did I change it: [Many people ignorant about Pascal will probably be confused ] [by the terms 'parse' and 'token look ahead'. If necessary, this ] [information should come later in the article. I think my simplified ] [variant is enough, though. ]
I think this will replace some aspect of what /Talk pages are used for: discussion between various authors. /Talk (or talk:) will probably still have other uses, though. But the need will be somewhat lessened.
Advantages over just using /Talk as we do now:
* Some useful /Talk conventions are automatically ensured by the software: + Reverse date order (often forgotten by newbies and me) + Separation of arguments by different people (sometimes it is unclear in /Talk pages whether a new paragraph is by the same person as the last) + Signing of arguments + Significantly editing other's comments is faux pas. * If a change is no longer viewable, its supporting argument is automatically deleted. These often stay in /Talk forever. * The "Be bold in editing" dogma is fostered: I am much more comfortable with editing right away rather than debating first when I can state my point at the same time.