Just to be clear on the terms, "constant directivity" refers to similar response at different angles off axis, and we should throw in "controlled directivity" which refers to diminished intensity off axis.  This is the sort of pattern controlled by horns/waveguides (like the Thoress), or the result of a large driver playing frequencies of wavelength less than the diameter of the driver (like the Haigner).  So both speakers you linked to would be beamy.  The Thoress looks like it has a fighting chance of producing smooth polars (constant directivity), but the Haigner looks like a hopeless case (no way the dispersion of the woof and tweet could possibly match at crossover frequencies, making it a fail in terms of constant directivity).

Also, it's a misconception that narrow dispersion speakers don't provide a big sweet spot.  They can if you set them up correctly, as described in the following link:

http://libinst.com/PublicArticles/Setup%20of%20WG%20Speakers.pdf

Regarding the Haigner, just because there is a directivity mismatch at 6 khz does not mean the speaker sounds bad.  The designer discussed the method to his madness in the 6moons Tech Talk article  (link on the product page, which includes some decent discussion on his take on CD speaker design too, check it out).  He also, not coincidentally, suggests cross-fire orientation as described in the link I posted.  

As far as the Thoress, I didn't realize the tweet was upward firing, but like the Haigner, it would exhibit controlled directivity at least from the midrange horn.  The woof's pattern probably matches the mid at the crossover point (well, who knows, but it really should!).  So kind of like the Haigner, controlled directivity over certain frequencies including the mids/lower treble, with the upper harmonics sprayed around more liberally.