I read that through a scotch laden haze... but it made sense...
Also, look at sigma and their issues with AF consistency. Canon doesn't give them their secret recipe, so Sigma reverse engineers it... and only somewhat effectively.
So if canon made lenses for Sony or Nikon or pentax, or even sigma... what happens if they don't share the Colonel's secret recipe... then the Canon brand is hurt as well as their reputation.
I have no issues with recommending either a Canon or a Nikon to a beginner... but it doesn't take much for someone to sour on a brand... because they heard a few bad things. I still haven't bought a Sigma lens even though I have had a boner for both the 35mm art and the 50 art... so there's that to consider.
When Canon adds a lens to their stable, they add characteristics of the lens to their firmware.
For example, let's say you are shooting with a 1DX and Canon comes out with a 200-600F5.6 lens. The lens would report itself as such to the body, and the body would say "I don't know you, but I will treat you like a 100-400" and the camera would focus the lens. Then Canon comes out with a firmware update for the 1DX and the focusing characteristics of the lens are in that firmware. Now the 1DX knows that if the AF sensor is X distance off that it has to drive the AF motor for time Y and it now focuses the lens faster with less hunting.
The problem with a Sigma (or Tamron) lens is that Canon does not put other manufacturer's characteristics into the Canon Firmware. Sigma (or Tamron) have to cheat and tell the camera body that their lens is whatever the closest Canon lens is to their characteristics. The AF can never be as good as a Canon Lens.
I thought Scotch only helped make things clearer
To focus a lens as fast as possible, you accelerate the focusing elements in the direction that they need to be moved until you get to a point where if you "slam on the brakes" the focusing elements will come to rest in the focused position. If you know how far you need to go, what the acceleration rate is, and what the braking rate is, you can calculate the time required to drive each step. Canon knows these values for their lenses and (presumably) they use those values in the bodies to make the focusing as fast as practical.
They have negative reason to put in other companies data. Say they put in data for a Tamron 24-70F2.8 lens... and now the Tamron lens focuses as quickly and as accurately as the Canon lens... that takes away one of the reasons to buy the more expensive Canon lens, and will financially hurt Canon. Why would they go to extra work to cost themselves money?
As to using Canon lenses on other mounts, we are dealing with the exact same problem, but now from the other side... Why would Nikon or Sony put in the firmware to enable top performance out of a Canon lens when it will cost them sales of their own lenses.
So much for the electronics, now on to the mechanicals...
Many of us who have been shooting for a long time will remember the Tamron Adapt-all mount. You would buy the Tamron Lens, buy the adaptor for your body, snap them together, and off to the races. (I still have my Tamron 90mm macro lens with it's Olympus OM mount and it's Canon EF mount) Slap in some simple electronics, and you could do the same thing with digital cameras. I seem to recall something about either Sigma or Tamron starting up a program where on certain lenses you could send them back for a mount change at a reasonable price.... same thing...