This may be useful background information about building a shower with Kerdi or any other waterproofing membrane. It is from the UPC (Universal Plumbing Code) that was written and has been adopted by many areas in the USA. I'd be surprised if the rules in Canada are not similar or nearly identical, if not identical. Now, this is from an older copy of the UPC standard, but it is my understanding it has not changed in this area (2009 version) to the present.
From section P2709 - Shower Receptors
The lining material shall extend not less than 3 inches (76mm) beyond or around the rough jambs and not less than 3 inches (76mm) above the finished thresholds. Sheet-applied load bearing, bonded waterproof membranes shall be applied in accordance with the manufacturer's installation instructions.
The North American Schluter Shower Handbook states the following (i.e., the version that has the US and Canadian certification and thus the one referenced in the UPC code extracted above):
Seams in the membrane are constructed by overlapping the edges by 2" (50mm) with unmodified thin-set mortar or by abutting adjacent sheets and installing KERDI-BAND with the unmodified thin-set mortar, centered over the joint.
The code approval agencies require the 'system' to be tested by a recognized agency, and in this case it is the ICC. This certification gets redone on an approximately 2-year cycle, and over the many years, the Kerdi shower system has always passed. The current approval is in report ESR-2467
http://www.icc-es.org/Reports/pdf_files/ESR-2467.pdf. Any other installation method does not meet code, does not meet the North American installation instructions.
FWIW, when Schluter deemed it prudent to require unmodified thinsets for use on their uncoupling membranes and sheet membranes, they had first determined that it worked, and that in these particular circumstances, it had significant benefits. Here is what the TCNA says about the use of modified thinsets under these circumstances:
When latex/polymer modified Portland cement mortar is used to install ceramic, glass, and natural stone tiles in an area the may not thoroughly dry out in use (e.g., swimming pools and gang showers, etc.) or where initial drying is inhibited (between tile and impervious substances), it is recommended that the completed installation be allowed to dry out thoroughly before exposure to water. This drying period can range from 14 to more than 60 days depending on the temperature and humidity and other climatic conditions, and whether the installation is interior or exterior. Consult the material manufacturer for minimum set times before grouting tile or allowing traffic, water exposure, or submersion.
Testing by Schluter has shown wet mortar on tile can extend to 90-days when installed over either Kerdi or Ditra. A dry set mortar only needs to cure (it gets stronger during the cure process when it is kept wet and able to attain the theoretically best possible strength), and you can get a reliable, repeatable, known result in a short time. When using a dryset with Schluter's membranes, they want it to be reliable, in all cases, and testing has proven way above the minimum requirements for a successful bond when you follow their instructions and guidelines with good workmanship. When I specifically asked the question about why their European instructions on building a Kerdi shower differ, I was told this: In the USA, when they proposed using a dryset mortar over their membranes, the industry raised a major hell storm. While true that bonding some types of tiles with a dryset over rigid surfaces (cbu, a cement slab, etc.), a modified provides some significant advantages, that isn't always the case when used with a waterproof membrane. But, consider that 'typical' substrates like cbu, a slab, or plywood all allow the modified mortar to dry - IOW, it is not trapped between what may be a nearly impervious tile and a waterproof membrane. Backed up by recognized testing agencies results, it became the North American approved method of installation. Schluter tried to do the same thing in Europe, but while some countries agreed, there was by no means a consensus and trying to get approvals in all of them meant they could not have a common Euro installation method which was and is important to them, and complies with only having to get one approval rather than dozens in their market place. So, they used their backup plan, and thus, the installation instructions differ. But, like the saying 'When in Rome, do as the Romans do', in the USA, it's do as the USA (and Canada) say, and as a result, to get the code approval and the company's backing, you have to follow their installation instructions as written for this market, something that John Whipple refuses to do and one reason why Schluter and he have some disagreements.
A shower built with conventional methods will see the structure reach a relatively high moisture content - a mudbed is porous, but water does take some time to migrate through it. It rarely will ever be totally dry in a shower used regularly. On the other hand, one constructed with one of the approved bonded sheet membranes, the membrane, after the tile is installed will rarely become wet or if it does, remain wet between normal uses. The tile and grout, on the sloped bed, shed the vast majority of it directly to the drain. On a mudbed shower and a membrane one, a very small percentage of moisture does make it through with each use. On a mudbed, it tends to be drawn down into the porous substrate. On a bonded sheet membrane, it tends to evaporate to the surface since it cannot penetrate the membrane immediately below the grout and tile with a thin layer of much less porous thinset verses a mudbed. Close proximity of the fleece between layers of Kerdi inhibits wicking very far into or through the joint, but it will absorb some moisture. That's the reason why the overlap and workmanship is critical on the seams. Submerged while in a flood test (the worst situation, and the only time there will be continuous coverage of water on the membrane) can push that moisture in on average up to about 1/4" depending on workmanship and how long the mortar has had to cure (it's less when fully cured). But, in the process, that moisture is going somewhere. IF there's a bunch of excess thinset on top of the naked (i.e., untiled) membrane, some of that moisture WILL penetrate that mortar, and the mortar without the fleece's small fibers from two sides to break the capillary action, can extend for a fair distance - all of which will remove water from the pan. All water has surface tension, and thus holds a curved edge. That slope of the curve will change depending on the surface and water content of the adjacent material. Thus, it can be misleading measuring the exact height of the water in the pan at the beginning and the end. A conventional, smooth, solid pan liner as used in a conventional mudbed shower is easier to check the exact height.
Bottom line, if you follow the instructions and have anywhere near decent skill, a shower built with Kerdi works, and works well. It has been doing that for many years in North America, and will continue to do it while abiding by all of the building codes and certification tests applied to it.