FreeBSD's future related to the software ecosystem

RetroComputingCollector

RetroComputingCollector

Using FreeBSD (or any other BSD for that matter) makes you too dependent on Linux libraries and Linux software that is poorly designed, bloated, and frequently without support for *BSD because it was dropped by LinuxBros. I actually think it would be easier from a maintenance standpoint, as well as a usability standpoint, and purely for the sake of "debloatifying" to develop new software and libraries (or just fork the already existing ones before they drift past the interoperability event horizon), so the BSDs don't have to deal with software written for a horribly-designed series of OSes that drift further away from UNIX every day. If the BSDs could work together on a BSD Alliance (a FreeDesktop type group for BSD that focuses on sanity rather than chasing stupidity) the BSD software ecosystem would be even better than it already is. Also, if the FreeBSD xorg port is based off of NetBSD's xorg impementation so it doesn't pull in all of that Linux crap, that would be great.
 
RetroComputingCollector said:
If the BSDs could work together on a BSD Alliance (a FreeDesktop type group for BSD that focuses on sanity rather than chasing stupidity) the BSD software ecosystem would be even better than it already is.
Click to expand...
I feel that might re-invent Linux's wheel :p https://xkcd.com/927/

I'm more a fan of decentralized approaches to breed creativity within spaces, instead of aligning spaces to a standard.


FreeBSD, OpenBSD, NetBSD, even DragonflyBSD all seem to be good BSD OSs, but all doing their own thing to all be unique in their offerings! OpenBSD is cool for security (no Wine), NetBSD is lean (runs on Wii even), FreeBSD is cool general-purpose (I like Wine :p), and DragonflyBSD has an interesting graphics stack (they had newer AMDGPU support years ago). I can pick technical reasons first.

Meanwhile, my desktop launchers worked as-is Fedora -> Ubuntu with ptyxis or a xfce4-terminal swap Debian and openSUSE: All 4 Linux distros barely differ at the base, while tailoring to corporations and standards (systemd, Wayland, XDG). All mainstream distros with enterprise/Server editions use a standards template, so what else is there to pick for a reason to use a Linux distro? Fads are cachy :cool: (while still entertaining the standards)
 
I understand your concern, but I think it only applies when using FreeBSD on a desktop. Yes, I think it is not the ideal situation, but it is what we have today.
 
RetroComputingCollector said:
Using FreeBSD (or any other BSD for that matter) makes you too dependent on Linux libraries and Linux software that is poorly designed, bloated, and frequently without support for *BSD because it was dropped by LinuxBros. I actually think it would be easier from a maintenance standpoint, as well as a usability standpoint, and purely for the sake of "debloatifying" to develop new software and libraries (or just fork the already existing ones before they drift past the interoperability event horizon), so the BSDs don't have to deal with software written for a horribly-designed series of OSes that drift further away from UNIX every day. If the BSDs could work together on a BSD Alliance (a FreeDesktop type group for BSD that focuses on sanity rather than chasing stupidity) the BSD software ecosystem would be even better than it already is. Also, if the FreeBSD xorg port is based off of NetBSD's xorg impementation so it doesn't pull in all of that Linux crap, that would be great.
Click to expand...

There are plenty of Linux users who don't like bloat either. Lightweight desktop is something many over there want.

Problem is, people who hate bloated desktops are usually in that category of people who don't code on desktops either. So the bloaties who are comfortable with thicker desktops win in the code department.

Not to mention sponsorship (paid developers) from companies that *really* like bloat.
 
Sometimes bloated works better because for one reason or another they have a larger user base & developer base.
So more resources.

One example i used artix linux.( arch clone, no systemd). One day it just died.
I tried other arch clones which also suddenly died. (manjaro did not install & endavouros died).
Now i use garuda linux (another arch clone, with systemd & no zfs support ). Why because it did not died until now. I also use MX-linux. (debian clone, 3-th place on distrowatch)
 
RetroComputingCollector said:
Using FreeBSD (or any other BSD for that matter) makes you too dependent on Linux libraries and Linux software that is poorly designed, bloated
[...]
so the BSDs don't have to deal with software written for a horribly-designed series of OSes that drift further away from UNIX every day.
Click to expand...
I notice it is mainly desktop / graphics-centric stuff that tends to be boated and poor quality. If you do a clean (non-graphical) install of BSD or Linux; both are pretty light. As soon as you want any kind of graphics, it becomes a mess (which is why I started a small side-project to make graphics on BSD a bit more fun again).

RetroComputingCollector said:
Also, if the FreeBSD xorg port is based off of NetBSD's xorg impementation so it doesn't pull in all of that Linux crap, that would be great.
Click to expand...
I think Linux going towards their own OS-specific display system (Wayland) might clean some of this stuff up. Xorg (probably Xlibre) can get fixed up and streamlined going forwards. I really do feel that going back to a monolithic display server will ensure the cruft and bloat doesn't creep in again. It was a bad idea when they made it modular (similar to my concerns of PkgBase), the inevitable cruft that will accrue wasn't immediately obvious to many.
 
Resources are limited. One should use them best.
But then different people have different interests.
For me valuable resources are wasted on a bloated browser like chromium, i don't now how many lines of code , or kde/plasma or gnome.
 
In my experience development of any tool actually takes a lot of time. -e.g., designing the code, setting up tests, creating examples, writing documentation, creating (un)install, etc... And unfortunately, also in my experience there is zero interest by others to even give feedback--or even "run a test" in most cases. The complexities of a window manager aside, try to create a simple program that "removes extra spaces in a plain text file" (with 0 dependencies so, a user wouldn't have do anything extra to test your program) and see if you can get any feedback whatsoever--or even someone to "test".

I, myself, have sort of switched to a more selfish approach; I design/create for myself. If I think something I write has some value to others, I'll push to github but for the most part, now-a-days, I just write software for myself (I've given up caring if Ricky, or Tommy, likes my stuff or even finds it useful). -e.g. I'll state in doc "history" this is a personal project and even most times I'll omit a license.

Maybe it was my ignorance at the time, but the BSDs used to be more "open" and "friendly" (more like a group of people with something in common). *meh* *shrug*
 
Alain De Vos said:
Resources are limited. One should use them best.
But then different people have different interests.
For me valuable resources are wasted on a bloated browser like chromium, i don't now how many lines of code , or kde/plasma or gnome.
Click to expand...
Interesting. I feel like the "unix mindset" has diminished. It used to be "use an existing thing where you can" and this had a bit of genius and bug-a-boo behind it; the 'existing thing' was often (more) stable and subject to less cruft but also a moving target in some respects. However, at the same time, I prefer less dependency in my designs. -e.g. if I can get by writing a simple JSON reader/writer function vs using an external lib, I will (if the JSON string is static and/or minimal, or used in internal communication, or...) just to minimize footprint (I prefer KISS).
 
JohnK I think perhaps the unix mindset may have changed? In the beginning it was about small tools, doing one thing and doing it correctly with ability of using "stdin and stdout" for chaining tools together. How many times have you (that's a generic you) cat file | awk | sed | grep > output? I know I've done that more times than I can count.
The concept of "tool" seems to have changed; hence the rise of bloated applications. Why do we "need" a Desktop Environment when a WindowManager works just fine?
Browsers: do they really need to be as big as they are? In theory a browser just renders HTM; perhaps all the different media types and security add, but to me browsers are trying to become the Emacs of the Web (practically an OS).
Using external lib: perfect, why reinvent the functionality? Some people "Not Invented Here" syndrome.

Circling back to the OP, I think his first sentence is not quite accurate. The reason for the "Linux" stuff is because of applications. The solution there is to create BSD native applications that do the same functionality. But then we wind up with the NIH again.

I agree with kpedersen up in #7: If the Linux world really moves everything to Wayland, it opens up the possibility for cleaning X up on the BSD side
 
> Linux -> Wayland = X -> BSD
yeah, I can see that. Great perspective kpersersen.

mer, I would absolutely love to pick your brain more (that's a less generic "you" than your "you" [not sure if that joke will land or not but it made me chuckle while I was typing]) -i.e. that's a great point and I'm fighting for a good (real world, small) example to use for a talking point. I mean; I guess I do have a recent need/project I'm playing with that I could use but that may seem like I'm actually fishing (I'm not). However, the "cat | awk | sed" example is a great point in itself. Where is the line where we accepted tool X (with all these added functions) when we just needed 'cat|awk'. The reason why that's a great point is because it takes a more experienced user to point out major flaws in, or missing, test cases. ...It would be wonderful to have discussions about those types of things with people because the time my abilities have been high enough to actually do some of the things I want to do is far less than others' time (I've used "cat | awk | sed" ten times so my tool is better than yours). ...is there such a thing as an "experienced BSD user pen-pal program"?
 
JohnK :) I have no problems answering DMs (but I do have to keep in mind I still work to keep the dogs fed).

I think some of the disconnect is "how we were brought up". I've been a CLI guy for a long time; Started with dads TI thermal TTY over 300baud back to Bell Labs, various jobs with VT100s up to VT220s (huge improvement). Others coming from Windows, want a GUI for everything. GUI for everything leads to complexity and bloat, CLI tends toward lean and simple with chaining.
 
Oh, honestly I really wouldn't want to burden anyone (and I certainly would only be looking for "opinion"--not "give me code"--type talk). But I too am a CLI guy so it would be really cool to have discussions! ...but, maybe I'll create a post about a project and try to spur some discussion about what is and what isn't a "need".
 
You must log in or register to reply here.
Back
Top