[Date Index] [Thread Index] [Author Index]
Re: v6: still no multiple undo?
On Jun 11, 5:37 pm, DrMajorBob <drmajor... at bigfoot.com> wrote: > As computers get faster and hard-disks get larger, bloated code is the > least of our problems, and slow code in the front end is not much bigger. > Multiple undo is a poor use of WRI's time (in the complicated notebook > environment), for a different reason (I'll get to it). Wow. Multiple undo is bloat. As opposed to a visual representation of a notebook as it has changed over time, but which can't be used to actually roll back to previous states. > What if the sequence was bad change, good change, bad, good, good. Won't > multiple undo lose more good changes than bad? What if the sequence was good bad bad bad bad? The thing is, you want to give people the option -- your example plays to the workflow that you're suggesting, keeping multiple documents and cells lying around just to save the progress of your work. The complaints about implementing undo being complicated because of the kernel miss the point. I don't really care about rolling back the state of the kernel. It's cell input *that I make* that I want to be able to remove. Like the case when I make a mistake and go to hit Ctrl- Z but type "Z" instead. Bam, there goes my precious undo. I concede that this would be possibly confusing if you undo an assignment and then forget about it. But that wouldn't outweight the convenience of being able to do so in the first place. Best regards, Will