Some people report of problems with coffee based developers and I get some inquieries every once in a while, but not very frequently. You can get very reliable and reproducible results with Caffenol developers if you know what you are doing.
First of all the basic darkroom techniques are exactly the same as with any other developer. If you are a newbie to film development, look for instructions you find a lot of in the web or very often in public libraries. It doesn't matter if the book is 50 years old. If you can develop a film with proper results in Rodinal f.e., you also can with Caffenol. And vice versa :-)
What are the special items to notice for a proper Caffenol development? It's the kind of "agents" you use and the kind of measuring.
---------------------------
Different coffee brands aren't so different as some people might think. Even expensive coffees work, I got reports f.e. for Nescafe Gold, but why wasting money? There may be very few exeptions, but if any other step of your procedure is OK you will get something and may eventually have adjust the amount of coffee slightly. By the cheapest "strong" or "classic" instant coffee available. Don't use deacidified coffees that are frequently branded "mild" and don't use decaffeinated coffee at all. Use pure instant coffee, no mixes with surrogates etc.
Vitamin-C (ascorbic acid) doesn't make any problems if you get pure one. Crushed pills or something like that may be another thing, try to get pure ascorbic acid. Here in Germany we are lucky to get it in any pharmacy at a very low price.
-------------------------
Soda definately can make probs if you don't know exactly(!) which kind of soda you have in hands. Please read this post first:
http://caffenol.blogspot.com/2010/03/soda-myth-and-truth_07.html
I asked in some discussion groups for veryfied data concerning the quality of soda they use. I didn't get too much feedback and I must believe that many users don't know what they have. But that's important to know, and demanding to know for developers containing a low amount of soda like Caffenol-C-L.
So dry your washing soda in the oven and meter the weight before and after. The loss of weight is the water that was in the soda before drying. All my recipes and most others are based upon anhydrous, waterfree soda. When you know the water content, it's easy to recalculate the amount of soda.
Verified data i have got for different brands of soda:
"Holsten reine calzinierte Soda" (and most other brands in Germany) is anhydrous/waterfree, containing neglectible 2 % of water. So far I didn't hear from a hydrated soda sold in Germany.
"Arm & Hammer washing soda" (US) is also anhydrous/waterfree with 2 % of water.
Update: Jeff reports that his recently bought A&H soda contains about 9 % water, so better check yourself if using low pH mixes like Caffenol-C-L
Senzora Soda (Netherlands, Ukraine) 20 % water. Monohydrate as stated by the manufacturers lab.
"Soude cristeaux" (France, Belgium) contains 50 % water. You need the doubled amont measured in gramms! It seems to be neither a monohydrate nor a decahydrate but something in between.
These are only examples showing that it's important that you determine the water content if not absolutely 100 % sure. There are different kinds of hydrated soda with different water content. Store the soda in an airtight box, otherwise it will attract water from the air. BTW, anhydrous, waterfree soda is a very fine white powder.
If you can't get washing soda (sodium carbonate, Na2CO3) you can use baking soda (NaHCO3, also known as Natron) but you
must heat it in the oven or a pan and not only water but also CO2 gas will evaporate in quite large quantities. No warrenty for your safety if you inhale a lot of CO2 gas!!! You will end up with waterfree washing soda. How long it takes? No clue. Heat above 120 °C until it doesn't loose any more weight as you do when drying a hydrated washing soda.
You can use hydrated washing soda once you know how much water it contains. You can't use baking without transforming to washing soda, it will fail.
Don't use any other washing additives like bleech.
Dilute always in given order, first the soda, then Vit-C (let bubbles clear), coffee at last, let stand for some minutes and stir gently.
------------------------
------------------------
Last and winning hands down as a source for trouble is measuring. I highly recommend using a scale/balance. Digital kitchen scales are fine for measuring 100 gramms or so, but may have a big deviation at low weights like f.e. 10 gramms. My digital kitchen scale is unusable below 50 gramms but precise above 100 gramms. I bought a cheap lab scale for about 12 Eur with a resolution of 0.01 gramms. If you have no scale precise enough you may use premixes with larger quantities that are easier to measure. Ezzie explains how to do this on his fine, fine blog:
Thanks a lot Ezzie for your invaluable support. And maybe you find the idea of premixed dilutions to be very charming.
If you stick to volumetric measuring it's important to determine the density of the agents you have in hands Scientific density data are absolutely useless. Believe me. You must do it yourself. That's why recipes based on volume are so unreliable. They are only suitable for the agents you use. Different brands of coffee, Vit-C and Soda can vary extremely regarding the volume, but not the weight. 1 gramm is 1 gramm. Always. But also pay attention to the water content of your washing soda. Once you determined the densities for your agents you may use your now individually calibrated teaspoons.
-----------------------
-----------------------
-----------------------
So as an extract : Know which kind of soda you use. Use a scale/balance. Or at least calibrate your teaspoons for each agent. A scale is better ;-)
Happy developings - Reinhold
after Wolf got fine results with the Ilford HP5+ at EI 1600 in Caffenol-C-L and I had splendid shadow details in Caffenol-C-H at about EI 800 I gave HP5+ another try as 35 mm film in Caffenol-C-L because I'm still looking for a replacement of the Neopan1600. HP5+ could be the film I will stick with. I could verify exactly Wolfs results, nice fine grain and still acceptable tonal range. 
Caffenol-C-L with 1 g/l potassium bromide, 5 minutes presoak, semi stand development 75 minutes at 20 °C, agitation first minute constantly, 3 times at 2, 4, 8, 16 and 32 minutes. Very low base fog level, very easy to scan from negative (Canoscan 8800F + Vuescan Pro) with only minor postprocessing. 
