On Sat, Sep 08, 2001 at 10:58:34AM -0700, Seth Nickell wrote: > > Adam is here describing items which can have incorrect settings. These > > are things for which there is a specific reason not to make them > > instant-apply. > > No, he's describing things that can take a while to change. Window > Manager, for example, should have *no wrong settings*. We're concerned > about it because the current implementation is buggy, and because it can > take a while for the borders to change. In the long term, Window Manager > is a *preference*. My interpretation of what Adam is saying is as > follows: > > "Instant apply is preferable in cases X & Y, but because both X & Y are > currently buggy or slow to take effect, we shouldn't use instant apply > for them at the current date" This is a silly argument and Adam should step in and stop it, but you're not reading the bit I was quoting. The part I quoted gave the example of IP addresses. > Putting implementation issues aside, e.g. speaking theoretically, the > selection of Window Manager is a preference not a setting, and hence > probably should be instant applied. That there are wrong choices today > is an implementation artifact that we will hopefully eventually remedy. > But until we do, we can't use instant apply for it. Agreed, with all the usual reservations. > > I'm willing to expand the role of instant-apply dialogs, but only > > where I am sure that the framework is in place for them to co-exist > > with delayed-apply dialogs. > > Explain this further? What is a "framework...for them to co-exist with > delayed-apply dialogs". The technical framework here is pretty easy, do > you mean the logical framework, the labels choices of button etc? I > think Maciej's suggestion was pretty good in this area... Framework is perhaps too grand a term for something as simple as this. I just mean the system that will allow the user to distinguish between the two styles. Maciej's suggestion being that the dialog should have different buttons depending on how changes are applied, yes? I don't think this is adequate. People don't read those buttons until they wish to apply the changes (and they shouldn't have to), by which time it's too late. It needs to be more immediately obvious than that. Btw, you were saying on IRC that the transition from delayed to instant wouldn't cause problems, whereas the transition from instant to delayed would. But consider, when we have both styles of dialog, users who encounter the instant style more than the delayed style will have the difficulties you envisage when they eventually do encounter the delayed style. > > > and we should be pushing to try and make more > > > and more dialogues instant apply (fixing the issues that make that not > > > feasible today, for example if switching themes took half a second it > > > could be instant apply, etc). > > > > There will always be issues making it unfeasible. We have to deal with > > the fact that there will be both styles of dialog. > > All preferences should hypothetically be able to be instant apply > eventually. I'm not sure if there will always be issues. There will always be performance issues. The more our machine power increases the more we try to do with it. My BBC Model B (ah, nostalgia...) booted to a usable state in less than a second. It loaded programs at roughly the same speed as they load today. (From disk, at least. From tape it was another story.) > We will have to decide on a default to push, that's what's in contention > here. Opinion can be divided internally, but the final decision we reach > needs to have everyone behind it. We need guidelines for choosing when > to use instant-apply and when to use delayed apply. The big question > right now sort of is, should the guidelines read as: > > "You should use instant apply dialogues unless (...)" > > Or should they read as: > > "You should used delayed apply dialogues unless (...)" If both can co-exist well together we don't need to push either style. I think the guidelines should read "you may use either style, provided you follow the guidelines appropriate for the style you are using". If at some point down the road we all agree that these instant apply dialogs are the greatest thing since sliced bread then we can be more forcefully in favour of them. > > PS. Could someone do me a favour? I'd like to take a look at some of > > the instant-apply dialogs in Nautilus, but I don't have the program > > installed. It would be really great if someone could send me or put on > > the web some screenshots of specifically those dialogs which are > > instant-apply. Thanks. > > You can't really get a sense for instant apply dialogues without using > them. Is there some way I can get a copy of Nautilus to you? Nautilus > has no delayed-apply preferences, and almost no settings at all. It *IS* > possible. I'm perfectly well aware that instant-apply is a behavioural thing. Please, give me some credit. What I'm interested in is seeing if there are any visual cues that the settings will be applied instantly. I just have little desire to download a 10Mb program which I quite dislike and which I know I'll delete immediately afterwards because I'm permanently low on disk space. colin _____________________________ ____ rtnl http://rational.cjb.net c z robertson ndirect co uk icq 13294163
Attachment:
pgpWwXbqNinBr.pgp
Description: PGP signature