[Date Index] [Thread Index] [Author Index]
Re: v6: still no multiple undo?
Andrzej Kozlowski wrote: > On 13 Jun 2007, at 20:32, David wrote: > >> I'm amazed that anyone would argue against a >> multi-level undo for notebook input text corrections. Especially, that >> it would take processor time away from the CPU or development time >> away >> from the developers. I don't think anyone is suggesting multi-level >> undo >> for input and output. > > I'm amazed that anyone would join a thread without reading it. For > example, who is it who argued against "multi-level undo for notebook > text input corrections"? Could you post a quote? I myself wrote > explicitly: > >> Of course a >> simple kind of multiple undo, one that only undoes typing and not >> evaluation, might not present any problems, though it would hardly be >> worth making so much fuss about. > > Yes, this kind of undo is obviously such a trivial matter that I > would certainly not have wasted any time to argue against (or for it). > Then there is you assertion: > >> I don't think anyone is suggesting multi-level undo >> for input and output. > > Have you tried reading this thread before sending your response? Have > you looked at the posts by > David Reiss, Luc Barthelet and David Bailey? They all discuss more > than just text-input only undo. They don't necessarily argue clearly > in favour of it, but they all consider it a possibility, which means > that it was an issue considered in this thread. > I will just quote Luc: > >> While I strongly support that multiple undo is a feature we need to >> see >> soon, it will take some real work to get there. > > > Given that the issue of a more than text-input level undo had already > been raised it does not seem to me so amazing that some people, > (myself and Bobby Treat) would wish to point out their own attitude > to it. I clearly indicated that I was not referring to "text input > only" undo. > > I suggest to you (and others who have written in a similar vain) that > it is a very good practice to read all the posts in a thread you > decide to join particularly if you are intending to make sweeping > statements of the type "nobody is suggesting" etc. > > Andrzej Kozlowski > You're a bit over reactive about this! I did incorrectly make the generalization that "nobody is suggesting". My apologies for that mistake. I should have narrowed that to say that "those that want a reasonable implementation of multi-level undo is suggesting". My inference from the OP is that a typical multi-level undo in the "MS Windows or Apple application multi-level undo" paradigm was the desire. It's as simple as that. Others decided to make it much, much more complicated. I've followed the entire thread. A thread that turned into a mish-mash of well meaning, but "off the initial point", posts. Some posts are talking about autosave, which is not the undo that the OP refers to, in my opinion. Or, some have implied/suggested that the feature might need to undo kernel states. That just turns multi-level undo into a massive, unnecessary feature, again, in my opinion. You asked for a quote... one of your posts is one of those that would seem to argue against multi-level undo because it might slow down your computer or the WRI developers from more important tasks. If you're not arguing against, then I don't understand why you posted to this thread and why you made the following statements. ---Andjez Koslowski wrote:--- There are a few features that are useful to everyone, and there are many that are useful only to some but (unless they are made somehow optional), will slow down everyone, forcing people either to get new hardware or give up other features that they really need by having to stick with older versions. I agree that a multiple undo would occasionally be useful, but I certainly would not pay the price of having my computer paralyzed for a few minutes every time time I save. Besides, I can think of many other features, more directly relevant to the main purpose of Mathematica, that I would rather have than this one. Of course a simple kind of multiple undo, one that only undoes typing and not evaluation, might not present any problems, though it would hardly be worth making so much fuss about. Anything that would bring my computer to a halt is unacceptable to me, however nice it might be for people who alsways have the latest hardware. All I wanted to point out is that there is also this aspect to features like this one and I hope that WRI does not forget about it when designing new versions. Andrzej Kozlowski --- I will not reply to any future dissections of my posts by you.