Now, before this direct-drive love-fest gets too out of control, let's talk about some other issues.
First, the "belt-drive crowd" does get it, and there are turntables out there that do not succumb to any of the things that Moncrieff points out in his turntable article.
For example, a quality ironless-core, non-cogging DC motor provides a perfectly smooth output. So using a motor of that type takes you a good part of the way there. Next, using a non-suspended turntable takes the subchassis interaction out of the equation. Third, having a non-stretch belt, combined with the above 2 things, eliminates the "stretch-release-stretch-release" syndrome that Moncrieff talks about. So no cog, no platter/belt interaction, no RC-tank effect interaction with belt and subchassis. What's left? Motor speed control. And heavy platters provide the momentum needed to reduce/eliminate the effects of stylus drag, making speed adjustment very infrequent, and maybe not even needed during the play of the LP. If speed adjustment is required, slow application of this adjustment will make it nearly or totally unnoticeable. A quartz-lock mechanism will not allow slow applications of change, and make immediate changes in a very small variation tolerance, so speed does not vary alot in amplitude, but it does vary alot in frequency. It "hunts" for speed. Up and down and up and down. This is true in either belt-drive or direct drive. Some may say it is perfect because of tight specs, some may disagree. So, as you can see, there are belt systems which do not suffer from these "demons" that Moncrieff waxes so epically about.
Now, direct-drive motors are inherently directly connected to the platter by their shafts, and all motors vibrate. ALL. When the motor vibrates, this vibration is directly coupled to the platter, and the platter vibrates. The vibration of the platter during play can and will cause information loss or distortion, since the record is moving microscopically under the stylus in directions other than the time axis. Better main bearings can minimize this effect. So can heavy, well damped platters.
Belt-drives on the other hand, have the motor somewhat isolated from the platter by the belt, and much of the motor vibrations are damped by the belt, and gone by the time they reach the platter. Since they invariably have some reasonably heavy platter weight, any vibrations coming into the platter will be of low magnitude, and easy for the platter to damp.
As I tried to show here, a well thought out belt drive table has much to recommend it, including good speed control and high vibration isolation, and there are a number of belt drive turntables that perform at extremely high levels. Poor belt drive tables are not an accurate representation of what belt-drive can do, just as poor direct drive tables aren't a good representation of what direct drive can do.
The absolute worst combination possible, is a cheap direct drive, cogging motor, quartz-locked, light platter turntable. These were typical in the late 70s mass-market units. They vibrate, cog, and hunt, with very little compensation from platter weight. Next up is a a cheap belt drive with a poorly designed subchassis, rubber belt, AC cogging motor, and cheap light platter. Neither of these is a very good choice. These are primarily what Moncrieff was talking about. IMO.
At the highest levels, both technologies may be good, but the execution of each design will determine how great it is.
Speed control is very important, but it does not exist in a vacuum. Vibration induced into the platter is also an important aspect of the design. They should not be considered as separate from each other, since the overall performance of the turntable is dependent upon BOTH of these parameters, and not just one or the other. This is why bearing design, platter design, motor and speed control, and vibration control and damping, are included in turntable design. Failure at any one of these criteria will result in a poor sounding turntable. The steps up the scale of performance reflect the proper addressing of all these points together.