3 ms·
The problem with keyboard or menu cut/copy/paste (hereafter c/c/p) is that it’s a mode (or at least a sort of quasi-mode), and modes are a fundamentally inhuman
by jacobolus 15y ago
The problem with keyboard or menu cut/copy/paste (hereafter c/c/p) is that it’s a mode (or at least a sort of quasi-mode), and modes are a fundamentally inhumane user interface approach that shift cognitive burden from computers to humans. The traditional c/c/p includes no visual feedback whatsoever, is an extremely overloaded concept (many types of objects can be copied, the result of pasting depends on context and there’s no way to tell precisely what will happen from past experience with other applications, sometimes a copy instruction only places a link to the content (for performance reasons) so that closing a document or application alters the paste behavior, c/c/p between applications is often brittle, especially when dealing with rich text), and is very limited and inflexible (no way or bad ways (depending on the implementation) to copy two things and then paste them back, copying another thing destroys the user intent of the previous copy operation and can’t be undone, no visual/spatial context is provided for the operation, etc. etc.).
C/c/p is better than what came before it (it was invented at PARC in the early 1970s, and the alternative was much more mode-ful, something like vim), but is still too unfriendly for novices and too limited for experienced users. Unfortunately, it is so ubiquitous that it has forced all users to learn it, forced all software to implement it, and crowded out alternative approaches, even in contexts like file managers where the concept doesn’t really make sense. The best we can typically hope for is extensions of the typical c/c/p interaction which work like other implementations in the simple case but have some extra functionality for experienced users. These are usually entirely undiscoverable, and still share most of the problems of the traditional c/c/p.
- lukeschlather 15y agoI'm not sure I follow where you're coming from. The biggest thing that I don't understand is what the alternative to c/c/p is. I feel like drag and drop is an even fuzzier metaphor than c/c/p. At best it has most of the issues you've outlined with c/c/p. But, I'm an Emacs user, so c/c/p is instinctual. If everyone grew up using Emacs people would probably be writing papers about how chaining keystrokes is some sort of deeply evolved human skill.
- jacobolus 15y agoThere are many possible alternatives to c/c/p in “idea space”, and the little details matter. A better system, though, would have some of these properties: it would persistently store cut/copied content in some kind of list, at least content of manageable size; it would have some way to see the clipboard visually/spatially; it would interact with drag/drop; it would give some visual indication that there was currently copied content, and perhaps the type as well; it would allow a user to interact with the history of cut/copied items in a more flexible way both from keyboard and mouse; it would allow easy conversion of copied chunks into “documents”/“files”/“pages” in their own right. Many of the problems here intersect with the problems of our current files + hierarchical filesystem + semi-spatial file browser + open/save dialog system for dealing with “content” in general. Ideally, a generally more intuitive and flexible basis for these problems would be built, and the replacement for c/c/p would integrate with these other tools. That’s sort of wishful thinking though. (Well, the idea of replacing c/c/p is also probably wishful thinking. Systemically different approaches like the Canon Cat word processor are all pretty much dead by now.)
- JackC 15y agoMany of your requirements are met by Ditto (on Windows) and ClipMenu (on the Mac). The basic model is: Ctrl-V pastes like usual, Ctrl-Shift-V pops up a menu of every previous paste. Previous items can be selected and pasted using either the keyboard or the mouse. This gives you persistence, visual display of the clipboard, and keyboard and mouse interaction with the history. And it turns out to be insanely useful. Oh man. I would never go back.
- robomartin 15y agoWell, drag and drop isn't any more discoverable than c/c/p. For example, when I first started to use a Mac I had to Google to figure out how to import ebooks into my iPhone. It turns out that you have to drag-and-drop into iTunes. Just like c/c/p, once someone tells you that this is the way we are going to play the game it all makes sense. The grab-something-and-move-it interface is OK for "normal" users. I would suspect that most techies might be far more inclined to figure out how to avoid miles of drag-and-drop action by using the keyboard. There's another element to this. Our systems have, at a minimum, two 24 inch monitors and most have three 24 inch monitors. Drag-and-drop gets pretty old once you start to have to cruise around so much real-estate. Beyond that c/c/p, per your own data point, has been around since the early 70's. That makes it a convention that's been around for possibly over 40 years. I think people get it. This, of course, extends to operating almost any software. If someone is using it with any regularity it'd be crazy to mouse all over the menu system to get things done. The keyboard is far more efficient. This is particularly the case when macros can be programmed and invoked. In general terms, I agree: For grandma, drag and drop is great and she is not likely to learn any keyboard shortcut at all. For anyone who is not a casual or lightweight computer user there's a whole other layer to the UI --via the keyboard-- that makes for better driving. Dragondrop is a great tool. I was sincere when I said that Apple should buy it and integrate it into the OS. Not for me and those like me, but for the benefit of others for whom the grab-and-do-something interface is the only option.
- jacobolus 15y agoI never said we shouldn’t try to make a system accessible from the keyboard, or that drag-and-drop as typically implemented is the best solution here. I merely said that c/c/p is at once too difficult/brittle and too limited. Personally, I wish we used some kind of mashup/spinoff/extension of DragThing, Quicksilver, drag-and-drop for this sort of thing: flexible, letting us do more with the stuff in it than just paste it back, accessible from both keyboard and mouse ideally from anywhere on the screen, persistent and non-destructive. It’s a non-trivial problem, because we need something that scales from novices up through experts. But there’s a very large design space here, and most of it has been ignored for 30 years because the existing non-ideal conventions have ossified. It’s exciting to see even tiny new ideas, like the utility being discussed in this conversation. To follow your thought here: with all that tremendous quantity of monitor space available, we really should be devoting some chunk of it to showing what’s on the clipboard, or at least that there’s something there.