Innovations in window management (NewsForge) [LWN.net]

Innovations in window management (NewsForge)

Posted Jan 25, 2005 3:28 UTC (Tue) by louie (subscriber, #3285)
Parent article: Innovations in window management (NewsForge)

Skippy (linked to from the article) is a lot of fun... despite the slowness of XComposite, I'll still probably drop alt-tab from my mental map pretty quickly.

(BTW, am I the only one who finds it amusing that the author equates scriptability with innovation, when Sawfish was completely, 100% scriptable 6 years ago and I'm sure others were farther back? The only real WM innovation I've seen in my computer-using time is Expose, and sadly, we weren't the ones to come up with it...)

(Log in to post comments)

Innovations in window management (NewsForge)

Posted Jan 25, 2005 9:26 UTC (Tue) by cgray4 (guest, #11599) [Link]

It's like they stole lisp and gave it crappy syntax.

That's a slightly paraphrased quote, but it's pretty much true. Lisp and python are extremely similar -- python is about as verbose as lisp, and it is object-oriented (though I don't see why you would want object-orientation when scripting a window manager).

As far as innovations go, things like ion, ratpoison, and stumpwm are something innovative that I have noticed in the past few years. These are keyboard-driven, tiling window managers. While they may seem like steps backward, I am pretty sure that they are things that have not been done before.

Tiling WMs

Posted Jan 25, 2005 10:18 UTC (Tue) by wt (guest, #11793) [Link]

Didn't early versions of MS Windows have only tiling since windows
couldn't overlap?

wt

Tiling WMs

Posted Jan 25, 2005 10:23 UTC (Tue) by nix (subscriber, #2304) [Link]

And, of course, the thing which Ion and suchlike are visibly modelled on: Emacs.

Tiling WMs

Posted Jan 25, 2005 10:56 UTC (Tue) by cgray4 (guest, #11599) [Link]

Is that true? I figured that since the things that they stole from -- er "borrowed from" -- at PARC could have overlapping windows, MS would have been able to replicate that.

Tiling WMs

Posted Jan 25, 2005 15:19 UTC (Tue) by tjc (guest, #137) [Link]

> Is that true?

MSFT Windows 1.01 had modal dialog windows, but AFAIK top-level client windows could not overlap.

http://www.infosatellite.com/news/2001/10/a251001windowsh...

Tiling WMs

Posted Jan 25, 2005 19:44 UTC (Tue) by AJWM (guest, #15888) [Link]

I could be wrong, but I seem to recall that overlapping, vs tiled, were one of the (more than most people give credit for) innovations that Apple added over and above the original Xerox PARC work.

(Mind, Apple had quite a few of the original PARC developers working for them by that point.)

(Or maybe overlapping was just too hard for Microsoft in round one. Do you really want to try dealing with exposure events on DOS?)

Innovations in window management (NewsForge)

Posted Jan 26, 2005 7:42 UTC (Wed) by nix (subscriber, #2304) [Link]

Er, FVWM is flexible, but even adherents would admit that its programming language is a dog's dinner. I doubt it's even Turing-complete without using chunks of shell in it.

There's just no comparison with Sawfish: a real pity that Sawfish is so nearly maintenance-dead. (an unconfirmed rumour which reached my ears long ago has it that direct instructions from Apple are responsible for the effective activity-ceasing of its author, but I've not even tried to confirm this and it could be purest moonshine.)

Innovations in window management (NewsForge)

Posted Jan 26, 2005 17:11 UTC (Wed) by larryr (guest, #4030) [Link]

Er, FVWM is flexible, but even adherents would admit that its programming language is a dog's dinner.

Ive been programming the behavior of the FVWM managing my desktop through the standard FvwmCommand interface using C, Python, and shell programs for 7+ years. I didnt even know FVWM had its own programming language.

Larry

Innovations in window management (NewsForge)

Posted Jan 26, 2005 22:28 UTC (Wed) by darthmdh (guest, #8032) [Link]

Sawfish's author was hired by Apple, and he got to working on more fun stuff. Sawfish had long been seen as "feature-complete" in his eyes so there wasn't much wrt. maintenance to do.

Then the GNOME folks did their usual backflip and threw away sawfish for the almost completely featureless metacity and many users (read: users as in real users, not fanboys, people simply running whatever GNOME provided) switched.

AFAIK FVWM doesn't have a programming language. It has a configuration "language" but you certainly can't program in it.

There's not much innovation as such to have in an X11 window manager. The real trick is in providing a) a useful & pleasant interface and b) extra-wm functionality (such as menus, pagers, whatever). We have so many choices because everyone's idea of best usefulness and functionality varies so widely. I believe (from experience) the real gap that needs to be crossed is dealing with X11, writing a WM is fairly low-level and most programmers with the ability don't have the time. There's a lot of vision out there in the community and for the most part we're waiting to see it realised in implementation. Things like Expose (and even better stuff) have been in people's minds for decades, but its either been impractical to implement or its had to be pushed back in favor of more pressing needs.

One day I should go back through my email archives and pull out all the ideas we ever came up with and publish them somewhere. I've got a wealth of information stored there from the minds of the top WM developers and power users from back when the community seemed more united.