Apollo stuff

Jules> Ahh, that was what really prompted me to ask about the wall Jules> boxes - I worked for a software company for several years who Jules> favoured IBM TR over Ethernet, and recalled things not being Jules> as simple as just joining machines together as there was a Jules> rack of IBM electronics in the corner of each room. From what Jules> others have said about Apollo TR though it really is simple Jules> and the wall boxes are purely for cable routing. I don't think IBM wall boxes did anything (other than act as a passive bypass if you unplug the drop cable). The active machinery was in the wiring closet, to bypass the cable to the wallbox if the workstation attached to it was turned off. With token ring you had a choice: 1. Just wire the wallboxes in a ring and plug in the workstations. Simple -- but if the workstation was off, the ring was down. (Fixable with a bypass relay -- in which case, if several stations were off, you'd exceed the cable segment limit.) 2. "Star" wire to a concentrator in a wiring closet, which would electronically bypass any stations that were off. This would make token ring fairly reliable, at the expense of a complex piece of hardware. FDDI had the same issues only more so, which is why TR failed eventually and FDDI failed almost immediately. FDDI also suffered severely from standards committee politics and committee-induced delay (much of it intentional). paul