This limitation is not really relevant for RSTP and MST.To make the whole story short, the root of an STP network advertises its information and can only "hope" that everybody has received it. When I say hope, I mean that the root is going to wait long enough so that it is sure that the information has had enough time to propagate to the leaves of the network. To compute this time, it has to take into account that the bridges may be slow, and may loses seconds while relaying the BPDUs, and also that the links may be unreliable, and that up to 3 BPDUs can be lost. In these condition, you can see that the time required to put a port in forwarding state will depend on the distance from the root to its farther leaf. In order to be very conservative, and considering that the root can be randomly located in the network (STP is supposedly plug and play, manually locating the root is an option), this restriction has been converted into a rule based on the diameter of the network.
Anyway, RSTP does not rely on timers to put a port in forwarding state but instead use an explicit feedback mechanism. I've described a little bit more in detail this in the following document: http://www.cisco.com/warp/public/473/146.html
All this to say is that if your network is entirely RSTP, you can use basically the diameter you want as long as it is a little be under "max age". You can extend max-age at will in order to accomodate networks with larger diameters, and there will be no performance impact.
BTW, MST uses a hop count inside the regions. In the latest IOS images, you can set max-hops up to 255... which allows you to create really huge networks;-)
Regards,
Francois
