User Feedback :: Smaller, nearly apps-less DSL?



One option would be something similar to what was mentioned a few days ago.  If a large portion of the "non-essential" applications were included on the ISO as UCI files in the new mydsl directory, it might benefit both the people who want the full desktop and those who want it minimal (or want to replace existing software).  For the first group the behavior of DSL would appear to be the same, and for the second group the removal of applications would be as simple as removing the packages from the mydsl directory.  I say UCI because if the packages were not within the base system then .dsl and .tar.gz apps would increase the amount of ram required to run DSL.  One drawback to this is that UCI apps are not in $PATH, so things such as piping or running apps from scripts might be a little more complicated...unless they are only applications that people run interactively.

This could also allow for a more dynamic set of apps in use at runtime.  For example, a user could add a list of apps in his home or in the kernel line (e.g. xmydsl=firefox,beaver).  If that list is empty or missing, all apps are installed.  If the list contains "firefox", it would exclude mounting the firefox application at boottime.

Just some think food.

I'd like to echo the above and for a little suggestion (something I think user Henk
raised a while ago)...how about a half way house option: put the majority of the
apps into a CI (but not a UCI) and hard symbolic link them, a la mkwriteable but
instead to the cloop mount point of the CI, if required. It'll look the same, take little
RAM and those who want a core system can just detach the apps CI.... Hopefully
it circumvents all PATH and wrapper issues but I think its a one-way until reboot
if you mount it.

A barebones version would be nice for those with minimal hardware, such as many of the old laptops and destops people are resurecting, but I think for newbies its fine just the way it is. You can add and subtract as you please and in the end master your own version if you desire.

That said, a minimal version that would appeal to me would be small enough to fit on a few floppies, run on a machine with 16-32M or less of RAM and a small hard drive, have enough drivers to get me on the net, have a minimal gui, and allow me to do a net install of as much as I wanted from a giant repository.

Quote (nucpc @ Mar. 28 2006,08:55)
I'd like to echo the above and for a little suggestion (something I think user Henk
raised a while ago)...how about a half way house option: put the majority of the
apps into a CI (but not a UCI) and hard symbolic link them, a la mkwriteable but
instead to the cloop mount point of the CI, if required. It'll look the same, take little
RAM and those who want a core system can just detach the apps CI.... Hopefully
it circumvents all PATH and wrapper issues but I think its a one-way until reboot
if you mount it.

Symlinks are very expensive.
On some machines just doing a mkwriteable nearly depletes the available ram.
Whereas our current UCI does not.
I could fiddle with FAT but not sure that is the way to go.
Also, I don't really want YADSLIT ( Yet Another DSL Install Type) to try to support across all other install types.

My focus for DSL 2.4 version is to try to enhance UCI.
My personal ultimate goal would be to have an "I Love UCI" edition of DSL.
No promises only progress. I have some ideas that I will soon be prototyping for v2.4.

I think that base DSL should be more than a kernel + XWindows.  DSL should continue to strive to be a livecd DESKTOP os.

If people want a better remastering experience, have them use MORPHIX.  That is what MORPHIX is designed to do.  Use the right tool for the right job.

In my opinion, the only thing that should trump usability (as many desktop apps as reasonably possible) is hardware support.

With the MyDSL system, it is easier for the user to correct any desktop application gaps than it is for the user to get some hardware up and running.

So tracking the KNOPPIX project is great for hardware support purposes, even if it means growing in size (64MB?) or reducing functionality (Ted goes back to the MyDSL repository instead of base ISO?) as a price.

Next Page...
original here.