BUDDENBERGRA@USC-ISI.ARPA (Rex Buddenberg) (05/11/85)
Charlie Strom mentions an overhaul of MODEM/YAM apparantly in the offing. First, let me compliment all the folks that made MODEM7 et al operate as well as it does -- Ward Christiansen and conspiracy have done well. Now some suggestions when the new model is worked up: -*-*- Larger blocks will improve throughput on quiet lines, and that capability is a worthwhile pursuit. But some of us live with noisy lines, at least some of the time. The option of reverting back to smaller blocks should be preserved for that and compatability reasons. -*-*- Has anyone considered adding error correcting coding? There exists a CCITT standard. Since processing power is cheaper than communications power, this might be worthwhile. -*-*- A little systems integration might be in order here. To do business, I need MODEM7 to work with my micro friends and Kermit to talk to the net. Could we all get together? -*-*- Suggest writing the code in as high a level language a practical to facilitate porting. -------
JALBERS@SIMTEL20.ARPA (Jon Albers) (05/11/85)
I agree that we should keep short as well as long blocks in the code. It makes more sense for some people. The idea of Kermit and modem7 getting together sounds VERY exciting. They are both very popular in their areas (Modem7 is more micro-micro, less micro-mini/mainframe, where Kermit was devloped to go between the larger systems, as well as micros). I think it would be a great step forward if someone could merge the two protocols into one program. Jon Albers JALBERS@SIMTEL20 -------
CSTROM@SIMTEL20.ARPA (05/11/85)
While I see your points, there seems to be a consensus among those who are planning to actually implement the changes to the protocol that if things are made significantly more complex we will lose the universality of Christensen protocol. Anyone is of course free to add whatever bells and whistles he likes (Pro-YAM for example supports Christensen, YAM [1K blocks], Kermit and Compuserve B protocols) but the basic definition of the protocol should not be made needlessly complex. By the way, YAM's protocol allows an arbitrary mixture of 128 and 1024 byte blocks, and SOH signaling the former wile an STX the latter. -Charlie