Since someone is complaining about interaction with systems with ^H as an erase character I'll have to answer their objections.
There is no problem here. Non-neolithic systems will inherit the erase character across telnet connections; rlogin always has worked. If we implement my proposal then connecting to a Linux system from a system which uses ^H for <-- will still work correctly, because the stty erase will be set to ^H. We should arrange that the X resources which change the behaviour of xterm are not in the app-defaults, but loaded into the X server at startup of the X server. That way the behaviour of an xterm will correspond to the xmodmap settings, rather than depending on what system it happens to be running on. Christian, I think we have rough consensus that my proposal should be implemented. If you agree you should say so, and then we can (a) file bug reports against packages which do not conform and (b) put it in the policy manual. Ian.
