The Canal Street connector is a good idea. It's also one that's pretty easy to see in the schematics. As such, my guess is that it's not there on account of highway geometry.
Elevating out of the tunnel requires a safe grade to maintain visibility and prevent slowdowns from drivers being unable to see more than a few hundred feet in front of them. This is particularly important going into a massive interchange where two of the roadways will be engaging in a hard turn.
The engineers know better than I do, but it appears that the crown of the road will be 56' below grade by GRB, per last year's schematics. The Dallas High Five tops out at 140 ft, and I would expect the 45 lanes to be of similar height when turning. So, there's 196 ft of elevation in ~4,200 ft if the rise starts from Commerce, or ~3,600 if it starts from Canal.
Starting at Commerce would give you a 4.6% grade - steep, but not excessively so, and allows for good sight lines. Canal, on the other hand, would require a 5.4% grade. This is steep enough to usually be restricted to hilly or mountainous areas, and close to the Interstate design maximum of 6%.
Along with this, there's the issue of vertical clearance in the depressed section once the cap is built. If the rise must start at Commerce for geometry's sake, and the clearance is 18 feet including fans, there is no way to construct a cap over Canal that will not interfere with the roadway - the rise over the 590 feet between Commerce and Canal would be 28 feet.
TL;DR: the numbers don't work for a cap over Canal, IMO.
I don't think they looked too far outside of their ROW, especially on the east side.
A Canal street connection would go a long way towards creating vital connections that have been eroded over time by the convention center, the convention center expansion, the basketball arena, the baseball stadium, and the soccer stadium.
and soon to be a 550' wide freeway ROW further eroding connections.
I wish there were a solution for the Polk Street freeway crossing.
I am assuming things will move faster once they are in a rhythm, but I just watched a crew trying to put the first forms around the rebar for the columns. I'm unsure if they were successful or not, I ran out of lunch, but I watched them try to thread that needle for almost 30 minutes.
Here's the corresponding image to help with my unarticulate description of events: