Small thing: keeping old books is a bit like a physical memory palace. Granted, it may be pointless/way to costly to cling to a collection of say outdated books read as an adolescent, but we tend to remember better if we can access at least the visual image of the cover.
So my advice is to take the pictures of book covers you have to get rid of. And organize them be it by the time period you have read them, topic or something else. One day you may be able to fish out memories you didn't think exist anymore.
Re science: yes and no. Meaning: the old way of compiling programs from sources on HPC systems while not dead is becoming rare. I can only speak about bioinformatics: the majority of programs one can get up and running in no time using either conda environments or apptainer containers. It may take few weeks for the brand new shiny version to get packaged by the maintainers, but this is seldom the problem.
Even if you want to compile some program you will save yourself the time and frustration using either conda env or a container with a required toolchain/libraries. Even on a box with root access you need to avoid conflicting system wide installed library versions where program A requires lib version X and program B lib version Y.
Well, call them lazy but once you have i.e. biocontainers in which individual bioinformatics programs are prepackaged, hardly any scientist in that field would be reinventing the wheel and often just waste te time trying to install all the requirements and compile a program already running "good enough" using downloaded SIF.
Sure, at times with say limited resources one can try to speed up some frequently used software creating SIF from scratch with say newer or more optimized Linux distro (if memory serves me right containers using Alpine Linux/musl library were a bit slower than containers using Ubuntu). But in the end splitting the input into smaller chunks, running i.e genome mapping on multiple nodes then combining the results, should be way faster than "turbo-charging" the genome mapping program run on a single node even with a big number of cores.
The whole idea of maintaining module systems in a perfect sync on several systems as compared to i.e. just rsync-ing SIFs sounds strange to me. Often HPC systems (or rather their admins) are fairly (and for a good reason) conservative, keeping old system and libraries versions. Your mileage may vary, but in a small benchmark of a bioinformatics program called samtools depending on the version the fastest binaries were either run in a conda environment or inside singularity container using Clear Linux distro. Binaries compiled using either system's GCC or from a module were slower.
One would have to repeat it throwing in at least Spack to see if this still holds water.
> maintaining module systems in a perfect sync on several systems
We use NFS. It's a fairly common solution. I don't think it's perfect but for this sort of thing you can turn on aggressive caching and then it works pretty well. In a previous company I believe all the modules were rsync'd nightly.
Frankly I have a problem with one of your problem as described. Who on the planet Earth creates a container with GCC, may I guess, to compile programs? and then complains about make being located in another container.
If you miss some utility just convert the apptainer container to a sandbox, install utilities you need and convert the sandbox to .sif as needed.
Also the whole point of building the program using compilers/libraries in the container is to use such container to run the aforementioned program in that very environment and not worry about libraryX or utility Y not installed in some box in a galaxy far, far away...
> Also the whole point of building the program using compilers/libraries in the container is to use such container to run the aforementioned program in that very environment
Says who? That's only because that's the only way it works, not because it's desirable to only work like that.
The only way Apptainers work well is if you put everything you need in one Apptainer, which makes them much less useful than you might imagine.
I have gave up on E. once they supported GWB over Gore. I can barely understand over the top devotion to neoliberalism and deregulation. But the shortcomings of GWB were sticking out in the campaign, so closing the eyes and singing "la la liberalism" was way too much for me.
You may check Europe. I can assure you that patients getting PD-1 inhibitors etc. while if I am not mistaken at least in part recruited for clinical trials pay zilch, nada for the drugs there are taking (in Spain).
I am not sure I want to have the same painting or a poster on the wall not even for the rest of my life but for 5-10 years. Apart from maybe a tattoo with one's blood group frankly I am puzzled by the idea of getting even the most artsy-fartsy tattoo anywhere on my body.
So my advice is to take the pictures of book covers you have to get rid of. And organize them be it by the time period you have read them, topic or something else. One day you may be able to fish out memories you didn't think exist anymore.