The first feature that makes gold so easily solderable is that it does not oxidize. There is no worry of surface contamination that is the bane of SMT engineers the world over. Whether soldering to Electroless Nickel / Immersion Gold (ENIG) or gold cable seals, it should remain solderable for years. Gold has remained a mainstay of the surface mount technology (SMT) world as the thin-film layer that keeps nickel solderable. Which is where the second feature of gold comes to light; that it is readily dissolvable in other metals. From my quick calculations, standard Tin-Lead eutectic solder dissolves gold at 35 micro inches (35μ") per second at 200°C. For comparison, molten tin dissolves gold faster, and Tin-Lead solder solders to nickel about 60 times slower. This is why when soldering to an ENIG circuit board, it is important to form the strong Tin-Nickel intermetallics with a relatively longer Time Above Liquidus (TAL) and higher Peak Temperature when reflowing when compared to copper.
Of course, there are some worries… I believe the biggest worry is the amount of gold that ends up in the final solder joint. If soldering to thick-film gold using a tin-based alloy, right around 10% gold in the final solder joint is where the brittle Gold-Tin intermetallics start to become a hindrance to reliability. Even worse, the tin may continue to dissolve the gold, even at room temperature, to form Kirkendall voids. Using thin-film gold may help in this area. Thin-film or "flash" of gold is typically 3-5μ". But, I have seen "thin-film" as thick as 15μ", which based on experience, I still consider a safe level of gold to be dissolved into the solder joint. On the other side, 50μ" and above is what I consider thick-film, where tin-based alloys are strictly prohibited. The gray area in between of 15-50μ" is when I normally recommend using Accelerated Life Testing (ALT) to determine whether issues will develop with the stated life-span of the product.