Quoth Siarzhuk Zharski <zharik@xxxxxx>, ... > So looks like I'm not alone who think this way. Sure, but there is so much confusion around this matter that every fool has plenty of company. (If you will pardon the expression - not aiming that at you, in fact I really appreciate your efforts with Terminal.) I agree with you over terminfo vs. termcap, but I think when the previous author said "The whole termcap system is broken by design", I don't think he meant to excuse terminfo from that. At some level it's immaterial which database you use, the system is equally lacking in robustness. One major problem with the scheme is that the TERM attribute is global, while the database is local. My terminal emulator may come with its own perfectly tuned terminfo entry, but that's useless on any remote host where that terminfo entry hasn't been installed. I used to carry around my own custom "beterm" and point TERMINFO at it, but that's far beyond where most users are willing and able to go. Web search for xterm-256color, and you will see hits for this problem. Add to this the high mutation rate of both terminal emulator implementations and terminfo descriptions, and the difficulty of testing an application with a representative sample, it's hard to take terminfo/termcap very seriously. For basic function, of course TTY applications should refer to it, but for advanced functionality I think it's a can of worms, and the little to be gained from trying to get to that functionality via TERM is outweighed by the problems. Is it better now than it was a decade or two ago when I had to deal with it on various UNIX platforms? Could be - but it could easily get worse again, for the same fundamental reasons it was bad then. Donn