On Wed, Mar 14, 2007 at 07:22:32PM +0200, Martin Vermeer wrote: > On Wed, 2007-03-14 at 15:39 +0100, Enrico Forestieri wrote: > > On Wed, Mar 14, 2007 at 03:18:54PM +0100, Jean-Marc Lasgouttes wrote: > > > >>>>> "Enrico" == Enrico Forestieri <[EMAIL PROTECTED]> writes: > > > > > > Enrico> One advantage is that you get numbered figures on screen. > > > > > > It would have been very easy to do even in 1.4; the only reason why I > > > did not do it is because (1) not enough time (2) procrastination about > > > the syntax to choose in the layout file. > > > > AFAIC the numbering was the only reason for preferring the inset. > > If this is also obtainable with the layout approach, I'm all for it. > > No, one reason was also that in the inset approach, it would be more > intuitive to the user where to put the caption text, and how _not_ to > put the figure in the caption. True, this would have required further > code to insert empty figure and caption upon caption inset creation, > which was never done.
I was only speaking about my preferences. However, I think that whatever we do, there will always be someone getting it wrong. That's a matter of experience. I would say that when choosing a float thing, it should come up empty, such that you have to explicitly put the caption in there. After all, I could well have a floating figure or table without any caption. -- Enrico
