I could be wrong about this, but I vaguely recall reading that Rudolf Diesel himself had done some experiments with using finely powdered coal for his newly-invented diesel engine.

As pointed out, the ash content of the coal is a major stumbling block.

But, in times of crisis people come up with all sorts of half-assed yet workable ideas. I've read that in Sweden during WW II, some cars were equipped with wood gasifiers that essentially pyrolyzed wood and fed the pyrolysis gas into the engine's carburetor. Or course, it also sent all  manner of gums and goos into the engine as well, and thus caused engine damage and resulted in the need to do frequent engine overhauls, but the point is that people were able to drive their cars without petroleum-based fuel, albeit rather poorly.

Guys, there is no need to speculate about the viability of coal to gasoline or diesel.  The process (Fischer-Tropsch) is commercial, with four large plants in operation world-wide.  One coal-based plant in Secunda, South Africa, supplies 30% of the country's gasoline and more for diesel.  

All of us peak-oil students need to familiarize ourselves with F-T since it is a proven and least-disruptive route to the post-peak economy. (This is especially the case for the U.S. which is the Saudi Arabia of coal).

So for all who are interested, these are the steps in the F-T process:

(1) Pulvurized coal is gasified--partially oxidized--to CO and H2 (if it was completely oxidized, i.e. combusted, we would have CO2 and H2O as products instead); the CO/H2 mixture is called synthesis gas or "syngas" for short.  Note that the heavy metals like mercury mostly end up in the ash and not in the syngas.

(2) The sygas can be cleaned using a number of different technologies, such as scrubbers and/or hydrodesulfurizers.  The Sulfur content in the syngas can be lowered down to 1 ppm without any exotic technologies.  

(3) The clean syngas enters the F-T reactor where it comes into contact with an iron or cobalt catalyst.  The desired reaction is given in Eq 1 below:

CO + 2 H2 --> -(-CH2-)- + H2O     (1)

The -(-CH2-)- is the hydrocarbon chain which represents the liquid product.  If the chain is 10 to 20 carbons long, we have diesel.  If it is 9 to 14 we have kerosine/jet fuel, and so on.

(4) Note that since coal has a high carbon-to-hydrogen ratio, the syngas will not have the desired 2:1 H2:CO ratio for Eq 1.  So a "shift" reactor is used upstream to convert some of the CO to H2 through the "gas-water shift reaction" of Eq 2.

CO + H2O --> CO2 + H2            (2)

(5) Depending on the F-T reactor operating conditions, the product will need to be refined differently to make gasoline, diesel, kerosine, or chemicals.  Overall, the refining section is nothing new and many of the existing crude refineries can be used to convert the products of Eq 1 into fuels that today's cars, buses, trucks and planes can run on.  

(6) The fuels are low in sulfur (because of the ease of gas cleanup in Step 2), and in the case of diesel, better performing (the highly linear chains lead to higher cetane numbers).  Low sulfur means better catalytic converter performance.  There are no aromatic compounds (alkyl benzenes or polynuclear aromatics)-- these lead to particulates in the tail gas, and can cause cancer.  

So in terms of product quality, F-T synthesis makes a much more environmentally friendly fuel than the highly aromatic hydrocarbon that is recovered from tar sands and oil shales.

On the negative side, the overall thermal efficiency (or EROEI) of the coal based process is about 60%.  Clearly this extra inefficiency translates to incrementally higher green-house gas generation (cf Eq 2).  

The reason I am giving this "lesson" is because I have noticed some degree of skepticism on this web site regarding the viability of coal-to-liquids.  Sure, it will increase our CO2 generation; but for an economy so dependent on the internal combustion engine and so blessed with coal, it is the most viable medium-term post-peak strategy I can think of.

and CO2 from a coal gasification plant in Beulah, ND is being shipped via pipeline to Canada where it is being injected underground into the Weyburn oil field.
but for an economy so dependent on the internal combustion engine and so blessed with coal, it is the most viable medium-term post-peak strategy I can think of.
Up to now, crude oil is used for transportation and coal for electricity generation. A widespread use of fuel from coal-to-liquids for transportation  could put even more pressure on our coal consumption already soaring to satisfy our growing need for electricity.

Other question: What quantity of coal is required to produce one gallon of fuel?

Coal is essentially CH, oil is CH2. Coal is variable in composition, much more than oil, in terms of C, H, S, and solids. Oil also has large amounts of water in it's makeup, if it's not anthacite. I don't think any oil in the world has more than 1% solids. Indonesian coal (from Borneo?) is probably the lowest sulfur and solids around.
A good rule of thumb is that you burn two CH to get CH2 and CO2 and power, that is, one ton of dry coal will make one half ton of oil and two tons of CO2. Very roughly.
In real life the plant will produce power and oil at whatever ratio is profitable, at roughly two cents a kilowatt hour baseload, ten cents a kilowatt hour for peaking (using a turbine to burn stored CO from the burner side of the plant) and two dollars a gallon for methanol, which will sell retail for three because of taxes, transport, etc.
Unless the Saudis really so have all the oil they say they have, in which case oil will go back to ten dollars a barrel and bankrupt all the synfuel plants.
There is another problem. Technological advances could abolish liquid fuels completely. In the case of carbon nanotubes we could have energy densities that would give us not merely flywheels that will drive our car across the country before requiring a "fillup", but enough power density to run an aircraft, probably a flying saucer ducted fan design.
How lucky do you feel?
I'd go with some syfuel plants, just in case.
Depending on the quality of the coal, the F-T liquid fuel recovery is 1.5 to 2.5 barrel per ton of coal (1 Bbl=42 gal).  The higher the moiture and ash content of the coal, the lower the liquids yield.  High volatile bituminous coals are considered best for F-T synthesis.

I have seen pictures.  Not many cars need to have a chimney.
Rudolf Diesel used gasified coal in his early engines. Many european and american cities of the late 19th century gasified coal on a large scale for gaslights, cooking, and as boiler fuel. Diesel simply compressed this gas to about 500psi and injected it into the cylinder. M.A.N. sold early diesel engines which included a coal gasifier.
It was what we call syngas, they called it water gas.