Wow. I'm not exactly a professional comic artist, either - more of an avocation . . . but that's quite the offer.
Certainly the file thing has been something I've been struggling with - I've been told that 150 kb is about all that should be expected from a comic image of reasonable size. I've played with jpegs (they get blocky), pngs (which don't seem to display properly on several test pages I've done), and gifs, which don't seem to reduce quite enough at high quality, and get kind of iffy at lower qualities. I've just started producing my work online, and it's kind of a pain. In a way, I guess it depends on how much time it would take - if it's easy, then great. Otherwise, I can play with my images some more. They almost always come out okay with enough experimentation.
As for other things . . . ideally, I think the presentation engine would vary for each comic. If we're really talking about design to enhance, and not just support the art, then it needs to be flexible enough to suit a lot of different possibilities, packaged into a pretty clean set of options. Naturally, simplicity wars with options. I have a hard time envisioning a 3D engine without more artwork than I can really create right now.
Some things I'd really like to have as I've designed comics in the past:
Images that change/melt with mouse passover. Especially useful if you're taking a panel from far shot to close up.
Images that would change with time (transparency enabled). Nice for having that skull gradually fade into the back of the panel during the horror plot-layout. (Or whatever.)
Easy add-ins for sound, animation, darken image, contrast image, blur, zoom, fade, etc. tied to time or location of page or mouse position. (These could also be layovers - create two images, with a "switch" between them, but the less work, the happier I am.) Especially nice would be a color bleed in option, where the image starts black and white and gradually adds color information. God bless you if you can get this limited to just part of an image. (A panel, for example.)
Simple transition animations in place - page turn simulation from one page to another, or fades. For one story, I'd really like to have seen an animation that treated my images like they were on adjoining sides of a cube. As the reader finished, he or she could click, and the cube would visually rotate, and the next "side" of the cube would be revealed. Extra brownie points for other geometric shapes - spheres, dodecahedrons, etc.
I've been playing with ideas for two almost totally opposite ideas - hands-off reading and additional reader involvement. With hands-off, it would be nice if the screen could "follow" the flow of the comic without needing clicks or scrolling on the part of the reader. This is tricky, since it could be more distracting, and people read at such different speeds. Maybe some other people have more thoughts on this . . . all I can come up with is the "timed" interface, with perhaps different selections that the reader can make, from "superspeed reader" to "snail paced."
I've also wanted to play with "Choose your own adventure" style comic, where the reader chooses which storyline to follow, or what decision the character makes (between two or more options). Unfortunately, that's just been really hard for me to implement without making a physical flowsheet covered with complicated pointers. It's been so ugly that I haven't even tried. Off the top of my head, my "helper" engine would need to:
Help me organize storylines that branch off, and where they re-attach. (Not essential, but really nice.)
Help create the pathways to these various pages (i.e. choice 1 leads to page 3a, choice 2 leads to page 6)
Make it clear when I've 'forgotten' threads, and make those easy to go back and fix.
Make it easy to modify 'paths' between choices.
Make it easy to add choices.
'Save' the reader's place at the last choice, and be willing to 'resurrect' them at the last branch that they picked so they don't have to push the back button lots of times or restart the entire comic. (Or last major branch, or whatever I choose.)
More stuff, I'm sure, that I haven't thought of.
Wow. That was probably way more information than you wanted. This is all purely a wish list, of course - I certainly don't expect this software. But hey - you asked, and it's something I've been kind of brooding over for a while.
emoticon. Images containing both are either small but with visible jpg artefacts around brutal color changes, or far too big if image quality is set too high, just for the sake of those few brutal color changes that will anyway not look perfect. What most people don't know because no tool to my knowlege allows this is that the jpg format splits the image into small 8x8 pixel squares, then apply a lossy color changing function on it (how lossy depends on the quality setting when saving), then compresses the sequence of modified squares with a lossless compression and stores the result in a file. I can write a program where you could open a TIFF file for exemple, then select areas of 8x8 tiles and choose the quality for the given area and have as a result a jpg that would both be small and have few arefacts around choosen important areas like bubbles.