I have Skype for voice plus a shared tabletop app (Maptool being my favorite).  I was challenged some time ago to design a game specifically for this environment, and while I'm well aware of the shortcomings, I don't know how to fix them.

Problems:
1) The die roller is fairly primitive, so physical grouping of dice is highly inconvenient, as is several people rolling at once.  Also, cards aren't supported, and if they are they can't do most of what cards really do.
2) I don't have the other players present to provide cues as to who's engaged, who's bored and who's been forgotten about.
3) Communication is only mostly synchronous.  Because of bandwidth problems, often you've got a 1 or 2 second delay which can lead to people talking over each other (this can exacerbate problem 1).
4) It's socially much easier to blow off this kind of game, so you get a lot of no-shows and cancellations, especially when playing with strangers.  People usually claim their internet went down, and there's not a lot you can say to that.
5) If you lose a player's attention, there's instant messengers, google, e-mail, and news out there.  This is why I have Maptool as part of the parameters, but it's got to be something the players are really using.
6) Players don't meet in advance so it's easy to have miscommunications about expectations.
7) Everyone and their dog seems to want to play D&D 4e - which is to say, a game that supports tactical combat with some fantasy window-dressing.

I think of these as the parameters restricting a truly internet-friendly design, but that's as far as I know what to do with them (basically nothing).  Does anyone have brilliant insights on how to get around any of these?

Are these unsolvable problems? 

Can anyone recommend games that suffer less from these problems?

Here's a readout for In A Wicked Age (solves, fails to solve):
1) F
2) F
3) F
4) S
5) F
6) S
7) F - but who cares about problem 7, honestly.  If I can solve 1-6 simultaneously I'll be overjoyed.

Howdy, Ryan.

When we did IaWA in GameTable (similar to MapTool), we had all of the Best Interests on the map, along with lines connecting characters via conflict points.  This really kept us engaged in what was going on overall. 

But I think IaWA would do better at your point #5 if the current scene were set up visually.  It's harder for me to imagine how to do this with IaWA, since the game doesn't demand a physical representation.  At least label some sort of physical location, and what important characters are present.  We all may already know this, but I think you build focus when it's up there on the screen visually.

This is really where MapTool shines for the type of game it seems to have been built for (i.e., D&D).  I think you've got to have some type of visual representation of the game, or MapTool will just be a distraction.  You'd be better off with a chat window and an online dice roller.  A visual representation, of course, doesn't have to mean miniatures, though.

Oh yeah, the dice roller in MapTool would have worked just fine for our purposes.  Just program a button with each of your character's forms and particular strength.  Click on the buttons that apply in any given roll.  I'd put #1 as an "S" as well; at least for IaWA more than any other MapTool game I've played.

Your points #2 and #3 are tougher.  Keeping a smaller group (3-4 players total) seems to help a bit.  Originally I had a house rule that people who thought they were being talked over could type into the chat box, and that would take precedence over the verbal conversation.  Of course, things don't sound the same in text.  Players tend to want to speak in-character over Skype.

Now, having all the characters on the screen, either in Filip's IaWA R-map-ish form with Best Interests or in D&D form with avatars helps when you're worried you've forgotten about someone.  For Engaged vs. Bored, in D&D I try to know what the NPCs/Monsters are going to do in advance so that the players feel that most of the time is spent on them.  Don't let a scene go on too long if it's not including one or more players.  Hit just one character flag of a player who is unengaged, and they'll pay attention.  Or perhaps I'm just stating the obvious. 

A point #8 might be MapTool slowdown.  In v1.3.b41 (or something like that) I was running D&D4 with several dungeon or battle maps necessary for the campaign loaded up.  MapTool didn't like the info overload, and seemed to be spending most of its time trying to autosave the mess, and locking up while we were trying to move units around the field.  One lesson here is, don't put a whole D&D campaign into a MapTool campaign file.

So, in summary, for the game you're designing:
1) Take advantage of what the MapTool roller can do (read the tutorials, etc.).
2-3) Small groups of 3-4 players, and timing or scene-framing rules that don't leave a player out for too long a period of time.  And see point #5.
4) IaWA really hits the nail on the head here.  Making something more ongoing like D&D only makes you suffer more from this.
5) Keep a visual representation of the game going and updated, that is evocative of Situation.  It might also help to require the players to interact with the visual representation somewhat.
6) Again with #4, just having a good pick-up-and-play game helps.
7) If you're trying to play with a specific group, well, you've got to play what they want to play.  Otherwise, I've found that there are sufficient people online who would like to play something else.  They're not incredibly hard to find.

-Jason