For what it's worth, I swap Backspace and Delete at the lowest possible level, usually one of: - linux keymap - keyboard-translate, which seems to be a no-op in 20.4 :-( - xmodmap depending on what kind of machine I'm on, where I'm on it from, etc. This always produced full recovery from new-style (read broken) keyboards that have fundamental keys like Esc, Ctrl, and Del in all the wrong places -- well, at least until Netscape came along. But at least in the case of Netscape, they didn't even pretend to keep sane key bindings -- half are from emacs and the rest are some sort of bastardized Windoze bindings, so I just put up with the annoyance of having it be the one Unix application that pretends its Delete key is running on another OS. I had no idea that XEmacs was planning on starting down this same misguided path -- I've been unable to read the list for the last two or three months. I mean, I can understand wanting to have Backspace delete backward, but who the hell ever thought that Delete should delete forward before Bill Gates came along? And why the hell is XEmacs changing to match the defaults of the single largest setback to computer science to date? I say, make the defaults be that Backspace and Delete both delete backward and that Control-H does something like a novice warning (`disabled' property) that forces the uncustomized user to choose whether to have it delete backward or run help. Then maybe add an ``I love Bill Gates'' thingy somewhere down in one of the 15th level submenus of customize that will make Delete act like Conrol-D. From: Karl Kleinpaste <karl(a)justresearch.com&gt; Date: 02 Jan 1999 22:14:27 -0500 Kyle Jones <kyle_jones(a)wonderworks.com&gt; writes: If this is the fundamental goal, then by all means, don't change the defaults to stop the Delete keysym from causing backward deletion. I know I'm not the only one that swaps these keys to deal with the last ten years' worth of keyboards. After swapping Control and Caps-Lock (I even found a Windoze hack for that one), it's got to be about the most common key swap around. As for "`C-h' is help" being something which threatened Stallman's health when introduced, all I can say is that you lived in a vastly different user population than I did when that came about -- where I was at the time, in an environment which included people using more HP than VT1xx terminals, nearly everyone I knew *loved* that mechanism. I gotta say, this is the first time that I've ever heard anyone claim that anyone thought that ``C-h is help'' was even a slightly good idea. Really, I'm not a big fan of the walking ego to start with, but that particular decision has got to be his dumbest one ever. The only thing that made it seem less galling was that most keyboards still had the DEL key where it belongs. The idea of seeming to dispose entirely of the availability of DEL in a tty, as well as the help binding, both in the name of keymap purity, disturbs me greatly. I don't care what it's in the name of, but making the default behavior for DEL be to delete forward is definitely the Wrong Thing in my book. Rick

Date: Sun, 3 Jan 1999 12:32:03 -0800 (PST) From: Martin Buchholz <martin(a)xemacs.org&gt; In tty mode, and by default in X mode, the block cursor covers the character after point. What could be more intuitive behavior than having the key labelled Del delete that character? Your question may be rhetorical, but my answer is ``leaving things alone.'' I type a key, it inserts a character. Then I type delete, it deletes the character. Pretty damned obvious if you ask me. Really, neither one is any more obvious than the other, but one is the way things have always been. You're arguing for a change, i. e. alienating your current user base. So far, the best argument that I've seen even hinted at is ``That's how Bill does it.'' Personally, I think it's a pretty sorry argument. This is not something Bill Gates got wrong. Maybe, maybe not. I think it is something he got wrong. You're the one arguing for a change. I'd hope that you'd come up with a better argument for it, or else leave it be. Rick

Rick Campbell writes: For better or worse, users are arriving with different expectations than the old UNIX community. It isn't the fault of these users that they think this way. DEC and minicomputers and dumb terminals were a mainstay of our indoctrination into the profession. Microsoft and desktop PCs are a mainstay of theirs. I agree with a later remark in your message that neither behavior of the DEL key is more obvious than the other. The problem is that we have two different communities that we are trying to serve. We could take the political stance than the Windows community is a closed, market driven community which has done nothing to improve our community, that being the free software community. And using this reasoning, we could refuse to support them and their UI assumptions. After all, we have limited amounts of resources and time and those resources should be spent building up our own community. This is, in a nutshell, my point of view. But with the broadening acceptance of open source products, I'm beginning to see that leaning a little toward the enemy might further our cause in the long term. We can obviously thank the explosion of the Internet for opening those corporate eyes, as no one could miss that the application layer of the Internet is thoroughly infested with open source software. But if we want to keep the attention of rest of the world, we have to bend a little. Fiddling with the DEL key is pretty minor in the grand scheme of things. If it will keep otherwise interested people from disgustedly closing their minds to open source solutions, then it is worth it. We don't have many advertising dollars to convince people, so we'll have to persaude them a few at a time. So yeah, I'm framing the issue in a way that makes us into noble Messiahs. It lets me sleep at night. :)

Karl> Kyle Jones <kyle_jones(a)wonderworks.com&gt; writes: Karl> Sorry if you think so, Kyle; I thought I was right on topic, regarding Karl> what to do about conflicting keymap desires. Karl> If the users deem insane the desire that keymaps behave identically Karl> under X and in a terminal, then I shall be pleased to be called Karl> insane, and I'll find my own solutions...though for the record, I'd Karl> still like to know how to coerce keyboard-translate into producing a Karl> DEL keysym for a literal \177, and why the seemingly obvious use of Karl> keyboard-translate didn't work. I just tested this. I tried xterm -e xemacs -vanilla -nw in my environment: Linux, with DISPLAY set to a Solaris console, and the key labelled Back Space deletes backwards, the key labelled Del deletes forwards the key labelled F1 is help-prefix the key labelled Help (!) invokes help-for-help This is pretty good. I congratulate the implementors (not sure who: Jan Vroonhof, Hrvoje Niksic and Gary Foster?) Probably f1 and help-for-help should work identically. There is probably a need to document better how to modify the current behavior for the geezer community (i.e. those who don't think naughty thoughts when they hear the word `tty'). There is still some work to be done before we achieve key-binding nirvana. If you want do hack on this, read comments at the top of src/event-stream.h. Martin