Gitin. I can see the difficulty in understanding the argument between Rashi and the Rambam in Gitin page 63, side B. However Rav Shach offers an answer to this great problem that I fail to see really answers it. One says to to two people "write a get and give it to a messenger." The Gemara wonders if this means to write once,- or at least once. To the Rambam this refers to a case that the get [divorce] was found to be null. To Rashi the case is when the get [divorce] was lost. To Rav Shach the argument depends on the question if writing the get [divorce] requires being a messenger or simply command of the husband.
What Rav Shach means is: The difference is this: If writing the get [divorce] requires to be a messenger from the husband, then we can understand that after he has written a get [divorce] and it got lot it might be a doubt if he has fulfilled his mission. this might be the doubt in the Gemara on page 63 maybe the messenger-ship was filled since he wrote the get as commanded. this is how Rav Shach explains Rashi that holds the doubt of the gemara is if it was lost. but if simply found to be null, then of course he can write another. However even if we say that the Rambam holds the writing does not require messenger-ship rather only a command of the husband, still I can see that there is the exact same doubt about how far his command extends--to write it just once or at least once. Like a drill sergeant would say, "I told you to get this done. I did not say to TRY to get it done!!!". o to say the argument between Rashi and the Rambam does not seem to depend on the question of if the get requires "messenger-ship or only a command."
[The basic issue in Gitin is this: One says to two to write a get and give it to a messenger. The Gemara asks, "Does he mean to write it once only, or to write it until it gets done?" This is left as a doubt. The Rambam says the question is referring to a case when the get was found to be null To Rashi, the issue is if the get was lost--but if found null obviously one could write another"]
