I have been wrestling with these issues for a good while now and would summarise as follows.
It may seem sacreligious to say so but in my opinion (despite a very long track record);
The English tying joint does not work that well! There; I've said it! And this from an Englishman too....
Ken says it all with his accurate observation that about half of the historic examples have been reinforced with iron bands. Now this could be down to other factors found in historic frames; such as subsidence of foundations (what foundations?) placing undue stress on areas of the frame. However you think about it, the shrinkage of the dovetail will losen the joint and compromise it's ability to tie.
If you are dead set on authenticity, one way to deal with the shrinkage that I have learned from historic frames is to cut the dovetail at very steep angles (tapering from full width of the tie beam to say 1 1/2" or even nothing. This will reduce the amount of spread at the plates for any given amount of shrinkage on the tie beam but it will still move a bit.
One should consider the forces at work within the frame when deciding whether to use the dovetail joint. For example; if the design of your frame puts significant thrust on the wall plates then you need a strong tying joint. If however you have purlins taking much of the rafter thrust them you don't need to be as cautious.
Then there's the question of whether to leave a gap at the teazle tenon shoulder to allow the tie beam to settle when the plate shrinks. The theory is good but what about the top shoulder of the brace? If you leave a gap here then you have a loose brace and a wobbly frame.
If you don't leave gaps at the shoulders then the plate shrinks away from the underside of the tie beam and is left fairly loose and at the mercy of what ever the rafters want to do to it.
You also end up with all of the loads from the tie beam (and in a traditional English purlin roof frame this generally means all of the roof load) bearing on the teazle tenon shoulder. Hence the need for good swept grain in a jowl post and the large number of historic examples which have splits runing down the length of the jowl. OK; I know that this is also due to heart shakes at a very inconvenient location but you really have to pick & saw your logs well to avoid this.
Even a cog joint instead of a dovetail doesn't get you away from these issues.
When choosing this joint much will depend on your customers expectations about tightness of joinery over time. It is unlikely that the frame will fall down due to your choice of dovetails or cogs.
So there you have it; a character assasination of the English tying joint. It's certainly not a perfect joint but it does stand the test of time sometimes (with a little help).
Having said all of this though, I agree with Kens approach. This is house carpentry in green wood we are talking about and it is possible to obsess about details.
Best wishes
Steve