This thread has become one of those arguments about which it may be said that "both sides are right" and, with equal force, that "neither side is right."
For the edification of any who might be confused, I humbly offer this attempt to clarify.
Let us start by defining some terms. Note that, since I'm using terms for my exegesis, I get to define them how I like, so let's not start by challenging my definitions:
"Single Action" refers to a trigger mechanism whose function is limited to releasing a firing mechanism whose motive spring has already been compressed by some other means.
"Double Action" refers to a trigger mechanism that combines two separate functions: compressing a firing mechanism spring that was previously at rest (that is, in the same configuration in which it lies when a round is fired) and then releasing that compression.
"Short Trigger Action" refers to a trigger whose travel from at rest (i.e., finger all the way off) to fire (i.e., gun goes bang) is very short, on the order of 0.1" or less; essentially, no longer than necessary to displace sear engagement.
"Long Trigger Action" refers to a trigger whose travel between the same two points is longer.
The classic "double action" firearm is the Smith & Wesson double action revolver. It so happens that, when a Smith & Wesson double action revolver is thumb cocked (turning it into a "single action"), the trigger is also changed from a long action trigger to a short action trigger.
As a result, people have tended to conflate the single action/double action dichotomy with the short trigger/long trigger dichotomy. Such a conflation was reasonable back when, in fact, the two distinct phenomena were always related. Today they are not.
The risk of an unintended discharge of a single action revolver (including a thumb cocked Smith & Wesson double action revolver) has two sources: one, is an errant finger press of the trigger, and the other is a drop induced jar that causes the trigger's momentum during the fall to move it into firing position when the fall stops. In a Smith & Wesson double action revolver, the drop induced jar risk is mitigated by the rebound slide, but the errant finger press risk remains.
On the other hand, a Smith & Wesson double action revolver with the hammer down has long been considered "safe" without any manual safety because the length of the finger press required to discharge a round is far too long to occur unintentionally.
Enter the Glock and the M&P. The M&P is a single action firing mechanism, despite what S&W claims and despite those who claim that a miniscule "camming effect" changes things. In fact, I defy anyone to set up an indicator and measure a camming effect greater than 0.001", if any. However, unlike the cocked revolver, the M&P employs a long trigger, pretty obviously by design. (In such a pistol, the length of the trigger pull for a second shot is governed, in part, by the design of the disconnector reset function and in part by the design of the striker block design.)
The Glock is a tad different: the firing mechanism is neither single action nor double action, because the cycling of the slide (coupled with disconnector function) compresses the striker spring by only a fraction of its full compression length. (I have read estimates of 80% to 90% and have not taken the time to devise and implement a measurement scheme to detect a more precise value.) At that point, further trigger travel further compresses the striker spring for the distance required to "trigger" (sorry about that) the down travel of the trigger bar by the disconnector slope. And, at the same time, and for the same reasons as the M&P, the Glock employs a long trigger.
Whew! Now what is the purpose of wading through all of this minutiae? Well, the philosophical issue that seems to have generated the discussion seems to be whether the "single action" striker-fired pistols, such as the M&P, are "safe enough" without a manual safety, and this is an appropriate query. However, because "safe enough" neither can be quantified numerically nor measured against a numerical standard, it is an issue the resolution of which is a matter of subjective judgment and, therefore, is up to the user in question. Here's what I can tell you:
I have decades of experience carrying, handling and firing Smith & Wesson double action revolvers and, at least in years past, instructing others to do so. Today, I have years of experience doing the same with M&P pistols (and, to a lesser degree, Glocks). I am satisfied that the M&P's long trigger is equally resistant to an errant finger press as an uncocked Smith & Wesson double action revolver, both in factory configuration and after installation of an Apex DCAEK. My only hesitation in reaching the same conclusion about the Glock derives from the fact that I've handled fewer of them and one or two had been so heavily modified (for competition use) as to lead me to believe that an errant thought might set them off. An M&P pistol with no thumb safety is one of the several firearms I carry on a regular basis.
Postscript: Someone will ask me, what then about the reported frequency of Glock UDs upon reholstering? Doesn't this mean that the Glock trigger is unsafe?
I think not. Of course, to be authoritative on the root cause of Glock reholstering UDs, I'd have to study the details, hopefully including instant replay video focusing on the holstering itself, of a large number of such incidents. Which I haven't done. However, surprise of surprises, I do have an opinion. First, while Glock gets the apparent award for frequency of such discharges, I suspect that that is largely because of the early prevalence of Glocks in the police market. Give the other makes a chance to catch up as they become more popular.
No, what I really think is at issue is training and user carelessness. I hate to say this, but my observation is that recent generations of police recruits bring zero firearms experience to the range; many (most?) have never handled firearms before and did not learn anything about them from their fathers, uncles, and other forebears. And, candidly, a lot of these recruits show up for firearms training only because they are under orders and getting time-an-a-half, not because they are eager to learn and develop a skill set. As a result, I've seen a lot of fingers inside trigger guards when they shouldn't be.
And I'd point to one other culprit: the modern holster for semi-auto pistols, which everyone insists must cover the trigger of the holstered gun. In my day, revolver holsters had exposed triggers, and as a result there was no unmoving bit of leather to press an errant finger against the trigger should such finger be where it wasn't supposed to be. I've heard folks talk about cover garments being trapped when holstering, but frankly I've never seen it where it could have caused a trigger movement. I have seen a lot of fingers where they shouldn't be that have had that potential.