I think you might have misunderstood my statement.
I think that is an understatement.
I feel like I'm in the Abott & Costello who's on first skit.
If you put a 100x100 image onto a 500x500 image, it will take 1/5 of the area, regardless of zoom. that's what I mean by native size.
Now you're talking about a different issue.
I'm not talking about putting images on images as in stamps on stamps. I'm talking about dropping objects into a map, which may or may not be a single bitmapped image. Like quickmaps for example.
The GM has a created a quick map with the grass texture and has a grid shown. The view is at 1:1 and there are exactly - no more, no less - 40 display pixels per grid. The PCs are accompanying a wagon train when some attack occurs. The GM places the tokens for the PCs, horses and attackers. Then he picks a PNG file of a wagon. This PNG file is 200 pixels by 400 pixels. He puts it on the display. He makes no adjustment what-so-ever to the graphic.
How many display pixels will the wagon occupy?
How many grids will it cover?
The answers better be 200 x 400 pixels and 5 x 10 grids. (Really freaking big wagon.)
Now if MapTool, when loading the image, got the DPI property from the image and saw that it was 200 DPI and then interpreted that to mean 200 Pixels Per Grid, then MapTools would load it and scale it to 40 x 80 pixels (1 x 2 grids) for the 1:1 view and all would be right with the world.
We'll have to figure out a way to make dropping stamps very simple, resize and rotate very intuitive and trivial.
Which is why there must must be a defined native resolution (pixels per grid or whatever) so that people can create/use graphics knowing that they won't have to resize every single one they bring in.
Using the DPI property of images to determine how big an image was meant to be removes this issue for the majority of graphics that people may want to use. It also makes it trivial to correct within the graphics themselves without requiring custom tools or other tricks.
Being able to resize and/or rotate images as you bring them in is a good thing.
Having to do it for every image you bring in is a bad thing.