Film or Video?  That's certainly one of the most controversial questions in low budget production.  But the very fact that it is so controversial means that the answer might not be all that clear cut.  And that the best choice might depend more on your own circumstances than anything else. 
  
Quality 
  
DV video makes pretty good pictures, but film still sets the standard.  In comparison to film, DV is better than Super 8, but not quite as good as 16mm (if nothing else, the resolution isn't there).  Yet still, DV isn't too far from 16mm; especially if lit properly, taking into account video's lower contrast ratio.  (Lighting is always critical, no matter what medium you're shooting in; the book Film Lighting is a fascinating reference.) 

And yes, DV video can be successfully transferred to 35mm film for theatrical release.  It's been done - for very large sums of money.  Of course, no one is going to mistake the result for something originally shot on 35mm; that's for sure.  But it can still be quite watchable. 

If you're going to shoot on video, it might be best to approach it TV style - concentrating on closeups, where it's at its best, and avoiding a lot of wide shots.  On the other hand, if you want big vistas for your sequel to Lawrence of Arabia, you'd most likely want to shoot on film. 
  
Ease of Use  
 
One advantage of DV is that it's a complete system.  From shooting through editing, everything remains digital, at broadcast quality.  You can shoot it on your camcorder, input it to an ordinary desktop computer, edit it, and then output again as digital video, all without loss of quality along the way.  Direct and simple. 

In contrast, film is more like a diva - one with brilliant qualities, but also a bit cranky about getting along with anything else.  To edit film on a computer (which is now the usual approach), you have to shoot it, have it developed properly, then transfered to videotape and digitized.  At least a couple of extra steps right there.  But film doesn't get along well with sound, either, be it analog or digital - you have to have a synch sound camera and a separate audio recorder, or go through tedious post-production looping or manual synching of wild sound, to finally get picture and dialog back together properly. 
  
Cost of Footage  

Cost of 10 Panasonic 60 minute Mini DV cassettes from B&H Photo, with shipping to Southern California - $82.35. 

Or $8.24 per hour of footage. 

Cost of 400 ft. roll of 16mm color negative film stock, plus developing, plus the cost of transfering the negative directly to videotape for editing - a minimum of around $250.00 or so.  A 400 ft. roll runs for 11 min. 7 sec. 

Or around $1350.00 per hour of footage. 

Now, in terms of how much actually has to be shot for, say, a 90 min. feature, let's assume a minimal 3:1 shooting ratio for film; that works out to 4.5 hours of footage.  For video, however, we will increase the shooting ratio to 10:1 (to make use of this natural advantage of video).  The final result for the cost of shooting stock for our 90 min. feature: 

16mm Film: $6075.00 

DV Video: $123.53 

Plus, the DV camera can also record excellent quality synch sound as part of the package.  Whereas for film we would have to shell out yet more. 
  
Getting It On Film  

Of course, if you want your movie shown it theaters, you have to get it on film, sooner or later.  And having video transfered to film isn't cheap.  It reportedly cost $58,000 to have the DV footage for The Cruise put on 35mm at the Sony High Definition Center.  And currently, this is a real advantage that 16mm has at film festivals; these generally still seem to require that submissions be on film.  (Though this is starting is change - with Sundance, no less, apparently about to lead the way at its year 2000 festival.) 

Then again, blowing 16mm film up to 35mm for commercial presentation isn't real cheap either.  And a video to 16mm transfer could be made for less money than putting it on 35mm. 

Still, all this is really missing the point.  Which is simply that, as was pointed out before, any such fantastically expensive transfer doesn't have to be made until after the feature is completed.  In other words, you can make your movie for a modest sum, see how it turns out ... and if it's any good, get investors to finance the transfer later. 

And though theatrical exhibition may be the holy grail the movie maker most seeks, there's always the "direct to video" market, too.  It's an alternative for getting at least some money (and with micro-budget production costs you don't actually need much to turn a profit).  In fact, this was the original idea for El Mariachi - it was intended to be sold to the Mexican home video industry!  And got sidetracked to the big screen along the way... 
  
Two for the Price of One 

Even El Mariachi cost $7000 for film and processing.  Once money is spent on film, it's just plain gone.  And El Mariachi was an exceptional bargain, because all the other projects shot on 16mm that I've heard of cost at least double that.  After which most have never been heard of again. 

On the other hand, if you spend $7000 on a DV camera and a computer for editing, and you make a first feature ... well, no matter how it turns out, you can then make a second feature.  Or a third.  Or even more.  For very little more money. 

(And at the end, even if it all turns out to be a total catastrophe, you can still sell your DV equipment back and get a fair amount of money for it.  Just try to do that with used film!  Or maybe you'll be a big success, and just want to upgrade to better equipment - somehow that sounds better!) 
 
Detailed Breakdown of Costs
 
Brian Boyl has worked out a (considerably) more detailed breakdown of film and video costs than the one I've presented above.  It could be of considerable value to anyone making their own production.  To take a look at it, just consult this Appendix.