Tommaso is right.
Please note that there is the possibility to add these back, but we need an extensive talk. Unfortunately, the thing is:
*) For Tahoe, I still have not found any reliable sources of its behavior. Wikipedia says one thing, the old ns-3 has done another. WIth a reliable source, adding it back should be easy (from wikipedia source, the difference is that after a fast retransmit it set the cwnd to 1, while Reno set it to ssthresh), but I still have some doubts (see later).
*) For Reno, the difference is in the Partial ACK management (they are ignored by Reno, but used by New Reno). In the actual code, that part is managed inside the TcpSocketBase and shared by all congestion control algorithms. Moreover, that part is standard since ages (more than 10 years). In my opinion, using Reno is going against the law, and demonstrating that the new law (NewReno) is better than Reno is a matter that has already been discussed in almost all environment (anyway we still miss the important grill-to-grill scenario, on a summer vacation with sands that introduces a variable packet error). Reintroducing Tahoe is, more or less, the same story.
have a nice day
Nat