fvwm, twm, openbox, Blackbox, were a few other light window managers that have been popular in the past.  Fvwm-crystal was something that blurred the boundaries of a window manager and a desktop environment.

Openbox became more useful when it was used with a light desktop - and that's what happened with LXDE.

Enlightenment is another interesting case; you could argue that it was one of the earliest environments that began to implement what we call a desktop environment today, though it is classified as a compositing window manager, can become a full desktop environment when implemented with"Enlightenment Foundation Libraries" (EFL), and it's with these libraries that a really light window manager can be integrated with other lightweight components to provide an integrated system experience.

The Enlighenment/EFL is probably more than what a true minimalist would want, but it does demonstrate that the ecosystem is mature with many possible routes along the pathway - consistent with the flexibility and choice that are foundations of the antiX principles.

min-jwm [...] it used only 68 Mb of RAM, compared to Space-IceWm's 89 Mb

"it used" is a misinformed statement. Most of that reported total is not attributable to the window manager.
ref:
2014 article"A Memory Comparison of Light Linux Desktops"
========= SCRAPER REMOVED AN EMBEDDED LINK HERE ===========
url was:"https://l3net.wordpress.com/2014/02/15/a-memory-comparison-of-light-linux-desktops-part-3/"
linktext was:"https://l3net.wordpress.com/2014/02/15/ ... ps-part-3/"
====================================


note: the chart & associated article are unclear which"KDE" version was tested (and what configuration? which widgets/components were enabled?)
On a separate page, netblue reported
========= SCRAPER REMOVED AN EMBEDDED LINK HERE ===========
url was:"https://l3net.wordpress.com/2014/03/03/lightweight-debian-lxde-desktop-from-scratch-part-2/"
linktext was:"https://l3net.wordpress.com/2014/03/03/ ... ch-part-2/"
====================================
"KDE 4.8.4" @329Mb

icons and pictures = bloat

My testing showed that each desktop icon, and each toolbar tray icon, typically adds 1Mb -- per item -- to the session overhead.
Even though the images are rendered small onscreen, each individual pixbuf buffer consumes 1Mb.

with only minimal services enabled

Yes, disabling autostart of unneeded services is an important detail that many folks probably overlook (and/or misattribute, blaming window manager for the incurred overhead).

The one limitation to both MX and antiX systems is that they are based on processors in the Intel and AMD families only - fortunately most of the equipment commonly used in commodity hardware comes from Intel and AMD, so MX and antiX run on the hardware that the vast majority of us use.

? ? ?

not sure whether my shoephone qualifies as"commodity hardware" but, yeah, I guess I'm fortunate that it uses an AMD processor...

That's interesting, I used to use Blackbox until I found Fluxbox, which is what I've been using for years now, with some Openbox usage.

I see that I ought to be using JWM or IceWM instead....... or maybe I should just go back to TWM. :lol:

fatmac wrote: That's interesting, I used to use Blackbox until I found Fluxbox, which is what I've been using for years now, with some Openbox usage.

I see that I ought to be using JWM or IceWM instead....... or maybe I should just go back to TWM. :lol:

Well, I don't know if you"ought" to do anything other than what you are doing.  Any of these window managers can get the job done, and they are all"reasonably light" in footprint size and reasonably capable; it all depends on the way you work,  and you know that best!

This is a very helpful discussion.  Thanks

skidoo wrote:

min-jwm [...] it used only 68 Mb of RAM, compared to Space-IceWm's 89 Mb

"it used" is a misinformed statement. Most of that reported total is not attributable to the window manager.
ref:
2014 article"A Memory Comparison of Light Linux Desktops"
========= SCRAPER REMOVED AN EMBEDDED LINK HERE ===========
url was:"https://l3net.wordpress.com/2014/02/15/a-memory-comparison-of-light-linux-desktops-part-3/"
linktext was:"https://l3net.wordpress.com/2014/02/15/ ... ps-part-3/"
====================================


My testing showed that each desktop icon, and each toolbar tray icon, typically adds 1Mb -- per item -- to the session overhead.
Even though the images are rendered small onscreen, each individual pixbuf buffer consumes 1Mb.

Yes, disabling autostart of unneeded services is an important detail that many folks probably overlook (and/or misattribute, blaming window manager for the incurred overhead).

Many years ago I did some"seat of the pants" research, using a variety of"lightweight" window managers to examine the memory usage and response time of several of them.  I also compared them to LXDE and XFCE.

I found that even with a few MB difference in the size of the memory footprint, there was very little difference in the perceived responsiveness of each of them.

In the case of LXDE and XFCE, they took a few seconds longer to initialize and make their environments available, but once logged in, there was very little difference between them and the lightweight window managers.

Back when I had a Dell Dimension 4100, when I first acquired the system, I'd use desktop environments, but as it aged, it took progressively longer to login to a desktop.  KDE was a common desktop I used from 2001 until around 2006, but by 2007, I was using XFCE, and after that, when I used the 4100, I'd use only a window manager.  But I do recall going back to KDE to compare overall response.  I discovered that the main limitation was the 256 MB of memory, which would swap if I ran more than a couple of applications.  Even then, the response would stabilize as long as I wasn't constantly swapping in and out of applications and causing the active applications to get swapped out.

My conclusion at that time was that memory was the primary limitation, followed by processor speed and context switching when both CPU and memory were constraining performance.  Memory, however, was definitely the number one constraint.  Therefore, if you have older hardware, the best thing you can do to keep it running well for as long as possible is to acquire as much memory as you can afford that can be supported by the hardware that you are using, and this investment, along with good care of the equipment, will allow 5-10 years of well-maintained use of any decent hardware.

Yes, I remember when I upgraded my netbooks from 1GB to 2GB, Firefox wasn't as slow as it had been & they became usable again.
Ram is the 'most bang for the buck' regarding upgrading, without a doubt.