Why fuel cells?

Why not just burn hydrocarbons as has been the case since the beginning of the Industrial Revolution?  One may ask that question if the need to obtain hydrogen cheaply seems to lead you back to hydrocarbons in fossil fuels.

It seems that hydrogen fuel cells are more efficient in terms of energy use.  A fuel cell powered auto can get 60 miles to a kilogram of hydrogen.  If the cost of a kilogram of hydrogen is twice that of a gallon of gas, it can still be economically competitive.  Most cars get less than 30 miles a gallon for gas, thus the comparison. 

It turns out that a kilogram of hydrogen produced from ammonia is less than a dollar per kilogram.  Now, if ammonia can be synthesized from hydrocarbon feedstocks, the end cost of a kilogram of hydrogen could be competitive with traditional fuels such as gasoline.  Let's say that refineries can shift production from gasoline to ammonia.  The crude oil would go further economically if the price of a kilogram of hydrogen can be held under twice that of a gallon of gasoline.  The cheaper the production, the greater the benefit.

Not being an expert in the matter, it appears that this isn't out of the realm of feasibility.

Update:

It also came to mind that natural gas is cheaper than oil these days.  One could use the hydrogen in natural gas to synthesize ammonia.  The chances that one could come up with a liquid fuel that is cheaper than gasoline and will go twice as far should not be too difficult to imagine.

As for the price of fuel cells, which are not cheap, these can also come down if the economies of scale can be applied.  This will occur if popular acceptance reaches a sufficient number to begin large scale manufacturing.  In addition, if an extraterrestrial source of platinum group metals is found, the price of could be further reduced.

Update:

I bought the pdf and read it.  The cost of hydrogen from this method is about 2 dollars per kg.  That was a few years ago, but it was in a time of high fuel prices, so the comparison may still be good.  If so, that would mean about 1 dollar per gallon equivalent assuming the 60 mp kg  holds up for this configuration.   The analysis shows a considerable expense in setting up an electrolyzer.  One would presume that if these were mass produced along with the fuel cells, the costs could be brought down to economic viability.  This is my guesstimate at the moment.  If there are other factors that aren't being considered here, I am unaware of them.

So, why not do this?   It was done in the case of the Tesla Roadster, where a battery powered car was produced.  Perhaps it would take a similar effort by someone who could pull off the same feat with hydrogen fuel cells.  If you used the same approach, with a high end vehicle, such as a luxury car, you could get a foot in the door, so to speak.