Timescale issues

[Owen Taylor <otaylor redhat com>]

As Havoc and Tim brought up on the 'GObject' thread, moving GtkObject
to a separate library would be a significant amount of extra work
added onto the 1.4 timeframe. It probably would mean that we wouldn't
get 1.4 any sooner than then next April or so.

However, some of the other things for 1.4 are also on a fairly slow
schedule ... for instance, at this point, I don't expect that the
heavy-duty Unicode support I have been planning for 1.4 is going
to be ready until at least then around anyways.

So, as I see it, we have two options:

 - Put out a small-scale 1.4 soon and put off the rest of the 
   features for a 2.0 targeted around the middle of next year.
   1.4 would then include:

    - the Win32 port (and possibly the BeOS port)
    - My drawing changes for double-buffering.
    - Possibly, light-weight Unicode support (basically a Unicode
      veneer over the current i18n support ... this would give
      us the equivalent of what Qt 2.0 has)
    - Miscellaneous fixes as time permits

 - Do a single 2.0 (or we could still call it 1.4) release in April or May
   with all the new features.

The first option has the advantage of getting some of the neat stuff
we have out currently as soon as possible, but would require more
coordination/maintanence work (That is, we'd have to go into release
mode within the next month or so), and also presents a less stable
target to developers.

Regards,
                                        Owen