Is there a real interest in pushing FreeBSD on the desktop space?

My goal when i said i do not post here is that the qwerty in a phone is small. I am sorry
My apologies, I was trying to make a joke, a deviation from "goose feathers". Lesson learned, silent = smart (works both ways)
Never intended to mock you.
 
Is it important? …

Clearly, yes (me quoting the Executive Director of The FreeBSD Foundation):

❝… focus will be to improve the desktop/user experience to make it easier to start using FreeBSD, …. More specifically, …❞

… no one will agree on one desktop …

True, in that I don't foresee any future -RELEASE of FreeBSD including any one desktop environment.

Inclusion of other things must (and do) take priority.

… FreeBSD 14.x-DESKTOP or something. …

To my eye, that's amongst the most thought-provoking suggestions in this topic.

If FreeBSD-DESKTOP will be distinct from FreeBSD-RELEASE (where RELEASE is preceded by CURRENT (main), STABLE, alpha, beta, and release candidates), then:
  • what might be the processes that precede release engineering of DESKTOP?
FreeBSD-RELEASE is very carefully engineered. The five-year support policy involving well-defined security advisories, well-defined errata, and so on. So:
  • what might a very diverse user base expect (or demand) from a release of DESKTOP?
Focusing on processes (not on possible naming conventions): there's commonplace underestimation of complexity.

This is not to dampen enthusiasm. It's to help readers focus on the realities of things such as:
  • what will be achievable
  • the contexts in which those things will be achievable
  • timescales.

… Make it happen.



My solution for all of you who want that is:
  • Make a forum post to discuss how it should look.
  • Start to work on it.
  • When a final product is done, introduce it to the community.


A dedicated space would be nice. First thought:


The desktop usage area could gain an experimental sub-forum.

Second thought: a significant proportion of developers simply do not gravitate towards The FreeBSD Forums. Again: that's not to dampen enthusiasm; we must be realistic.

It is human nature to like what others like.

+1

Also, natural to question and, sometimes, disagree.

Sadly, some people will find selfish pleasure in things such as negative provocation; disruption; and derailment. Derailment. Train crashes. Disagreements can become ugly.

Behaviour is learned.

Naturally, I dislike some of the behaviours that people learn from each other. I admire people who truly lead by setting a positive example.



Back on topic, positivity, and people who are known for making things happen. The absence of christhegeek from this topic is notable.

He's present in places such as Discord, and in another multi-page topic:

 
My solution for all of you who want that is:
  • Make a forum post to discuss how it should look.
  • Start to work on it.
  • When a final product is done, introduce it to the community.
This is how it should be.
Agreed. This is the only reasonable solution.
It won't help if the goal doesn't make much sense. Talk was about a "desktop edition" of FreeBSD. Reasons why this idea doesn't make sense were given. I'll just repeat the (IMHO) most important ones:
  • FreeBSD is a general-purpose OS, it doesn't have "editions" (and, shouldn't have them). The base system offers everything technically needed to use FreeBSD for any purpose (server, embedded, desktop, ...?), and no, this doesn't (have to) include userspace GUI software like a windowing system, "desktop environments", etc, as all these can easily be installed from ports, there's zero technical reason to pull them into base, but many technical reasons not to do that (lots of complexity and churn, just bloat for anyone using FreeBSD for something other than desktop, ...)
  • The "leave it as is and just automate installing a desktop in some graphical installer" approach makes no sense either, it would introduce a dependency from base to ports. Base uses a classical release schema with "long-term support" stable branches and point releases created from them for a whole of 5 years for each stable branch. Ports are "rolling release" and just offer quarterly snapshots. This is in line with the idea of base being the platform for 3rd-party software from ports. A dependency in the opposite direction would break the whole concept.
  • So, the next idea would be, just do it entirely in ports and offer an automated desktop installation there. Surprise, this already exists: sysutils/desktop-installer. Sure, if you want more "bells and whistles" there, get wild, why not make this a "graphical installer" with some wizard-style UI, why not add some option to the base installer offering to directly install and execute the latest version, whatever you like, but this certainly won't be a "desktop edition". It still might be something some people would like.
  • There's of course nothing wrong with doing whatever you like outside the FreeBSD project, like GhostBSD is doing. It must be outside because it will need its own "release engineering", and the result will be suitable only for desktops. Still, these projects are certainly useful to some users.
Of course, the issue mentioned here a few times that "nobody would agree upon which software and DE to use for that" is valid as well, it's not on my list because it can be "solved" by offering many choices here. This will greatly increase complexity and maintenance workload, but at least it's not a technical showstopper.

Written from a desktop running "vanilla" FreeBSD for 7 years now...
 
Sorry to need to necro-post this thread, dead since Xmas eve, but I don't want to waste considerable digging over the last year or so about problems with the -RELEASE DVD1 contents that seem pertinent.
Feel free to ignore <&^}=

Quoting post #188:

For two desktop environments: things are simplified, but not automatically installed.

Re: <https://www.freebsd.org/where/#choose-image>, FreeBSD-provided ⋯dvd1⋯ files include packages for x11/gnome and x11/kde5. Context:


That's regarding FreeBSD 14, where pkg-stage.sh properly covers all DVD1 packages, however it had been broken ever since at least 12.2 and 13.0 in several respects.

Ignoring github, and because for me it's easier to compare different releases, using:
https://cgit.freebsd.org/src/tree/release/scripts/pkg-stage.sh

where main (= head?) and 14.0 do not currently differ.

_DVD_PACKAGES="archivers/unzip
devel/git
emulators/linux_base-c7
graphics/drm-kmod
graphics/drm-510-kmod
graphics/drm-515-kmod
misc/freebsd-doc-all
net/mpd5
net/rsync
ports-mgmt/pkg
ports-mgmt/portmaster
shells/bash
shells/zsh
security/sudo
sysutils/screen
sysutils/tmux
www/firefox
www/links
x11-drivers/xf86-video-vmware
x11/gnome
x11/kde5
x11/xorg"

All of 12.3, 12.4, 13.1 and 13.2 wanted the no longer existing
emulators/linux_base-c6; it had been -c7 for ages.

12.4, 13.1 and 13.2 wanted
x11/gnome3
after gnome3 became gnome
so no linux or gnome on those.

12.3 and 12.4 lacked the then drm-fbsd12.0-kmod pkg, instead it had the (then not existing for 12.x) drm-{legacy,stable}-kmod, so it took me weeks to get S3 suspend / resume working.

Oh, and 13.1 still had
devel/subversion
devel/subversion-static
after the change to
devel/git

Is this a complaint? No, it's a description of a breakdown in communication and/or a lack of knowing who's responsible for maintaining that list in /src/release/scripts/pkg-stage.sh

I wrote to Glen Barber a few times during the 12.4 release cycle, pointing this out, but at the time the emphasis was on getting a broken link fixed, that hadn't been merged from 13.x, re which Graham thankfully intervened on my behalf.

Glen figured, I gather, that it wasn't re@'s job to make sure that list was up to date with present ports reality, citing possible build failures. re@ in general appeared to have little interest in the DVD, maybe due to having great bandwidth, where it's hard to imagine "third world" conditions.

My interest was then due to negligible bandwidth, broken sessions and efforts over months to get 'bsdconfig packages' working again, having been broken for years due to asynchronous changes to pkg(8) and some personal differences.

I had achieved fixing that, but my patches got left behind when the relevant PRs were closed, by which time I'd run out of puff, literally, and noone else indicated any interest.

My point is that if release CDs or DVDs will include packages, whether official -RELEASE or any other collections or whole alternative package sets any people care to assemble using the /src/release code, someone(s) will have to maintain accuracy and facilitate distribution.

smithi out.
 
smithi I can't recall the details, my strongest advice going forward is to not go behind the scenes with email. It runs contrary to having things trackable, and so on.

Do please make best use of Bugzilla.

My lasting impression of Glen, not only from the months during which he was my primary mentor, is that he is impeccably professional. In all things FreeBSD: never (to my nose) the faintest whiff of allowing personal feelings to interfere with technical decision-making and prioritisation.

There's some discussion of images in Discord.

HTH

 
It won't help if the goal doesn't make much sense. Talk was about a "desktop edition" of FreeBSD. Reasons why this idea doesn't make sense were given. I'll just repeat the (IMHO) most important ones:
  • FreeBSD is a general-purpose OS, it doesn't have "editions" (and, shouldn't have them). The base system offers everything technically needed to use FreeBSD for any purpose (server, embedded, desktop, ...?), and no, this doesn't (have to) include userspace GUI software like a windowing system, "desktop environments", etc, as all these can easily be installed from ports, there's zero technical reason to pull them into base, but many technical reasons not to do that (lots of complexity and churn, just bloat for anyone using FreeBSD for something other than desktop, ...)
  • The "leave it as is and just automate installing a desktop in some graphical installer" approach makes no sense either, it would introduce a dependency from base to ports. Base uses a classical release schema with "long-term support" stable branches and point releases created from them for a whole of 5 years for each stable branch. Ports are "rolling release" and just offer quarterly snapshots. This is in line with the idea of base being the platform for 3rd-party software from ports. A dependency in the opposite direction would break the whole concept.
  • So, the next idea would be, just do it entirely in ports and offer an automated desktop installation there. Surprise, this already exists: sysutils/desktop-installer. Sure, if you want more "bells and whistles" there, get wild, why not make this a "graphical installer" with some wizard-style UI, why not add some option to the base installer offering to directly install and execute the latest version, whatever you like, but this certainly won't be a "desktop edition". It still might be something some people would like.
  • There's of course nothing wrong with doing whatever you like outside the FreeBSD project, like GhostBSD is doing. It must be outside because it will need its own "release engineering", and the result will be suitable only for desktops. Still, these projects are certainly useful to some users.
Of course, the issue mentioned here a few times that "nobody would agree upon which software and DE to use for that" is valid as well, it's not on my list because it can be "solved" by offering many choices here. This will greatly increase complexity and maintenance workload, but at least it's not a technical showstopper.

Written from a desktop running "vanilla" FreeBSD for 7 years now...

Sometimes, you must let it go and let people discover why FreeBSD is as it is.

Also, you never know if someone could add an option in bsdconfig. There would be no need for separate ISO.

I think FreeBSD that boots to a desktop would be a waste of time.
 
Is human nature to do as other do, i.e. use Windows.
Well, there are frustrations with how Windows functions. Some technical, some political, some arise out of lack of useful info (real or perceived). Using something other than Windows is seen as a remedy for those frustrations.
 
There are people who always follows other (eg. using Windows), and there are people who decides to try something different (be it FreeBSD or Linux). At any moment in time a "follower" can wake up and decide to try something different. This is a good thing.
I see no upsides, no - none at all - to "lower the bar" so to speak on FreeBSD just to make it easier for "followers" to use.
In my view, we want quality users (those who manage to learn how to use it despite the flaws it has) over "followers". All the time.
 
Well, there are frustrations with how Windows functions. Some technical, some political, some arise out of lack of useful info (real or perceived). Using something other than Windows is seen as a remedy for those frustrations.
The point is not in how Windows works. The code is secret. No person can see it. The political factor is more a who can you buy factor. Is, by the way, Windows politically correct?
 
The point is not in how Windows works. The code is secret. No person can see it. The political factor is more a who can you buy factor. Is, by the way, Windows politically correct?
There's plenty of workplaces that use Apple hardware. In the workplace, you just use whatever's provided, and you're expected to be productive with that. Liking or not liking a particular OS is irrelevant.

I actually saw a few stories where a workplace actually accommodated a new hire's wish to use Linux on the job while everyone else was using Windows... That new hire ended up wasting so much time on resolving compatibility issues between their machine and the rest of the company's infrastructure that the company ended up terminating that new worker for lack of productivity. Point of sharing this story is that I just don't see Windows (or any other OS) as 'politically correct'. My assessment is that the very question of whether an OS is 'politically correct' is pointless. Just pick the tool that gets the job done in a reasonable amount of time, and don't get stubborn about it.
 
smithi I can't recall the details, my strongest advice going forward is to not go behind the scenes with email. It runs contrary to having things trackable, and so on.

I cannot agree. Not every discussion benefits from being undertaken in public. Bugzilla is public, that's good, but it wasn't getting traction when needed. I've done better with email since '98 (FreeBSD) after Netmail since '89 (Fidonet).

Besides, apart from the first message you, re@ and emaste@ were cc'd for all of that exchange - and with a successful outcome, at least toward the 12.4 DVD1 having that essential symlink fixed - again, thanks to you.

Do please make best use of Bugzilla.

I did try, then ran out of puff. If my health improves I may try again. See below.

My lasting impression of Glen, not only from the months during which he was my primary mentor, is that he is impeccably professional. In all things FreeBSD: never (to my nose) the faintest whiff of allowing personal feelings to interfere with technical decision-making and prioritisation.

Agreed 100%. I made a point of being clear about that:
Is this a complaint? No, it's a description of a breakdown in communication and/or a lack of knowing who's responsible for maintaining that list in /src/release/scripts/pkg-stage.sh

Certainly not Glen's fault, nor his job - but neither was it anyone else's. That's my point; it's a procedural shortcoming, that can't be reported as a bug.

There's some discussion of images in Discord.

HTH

Thanks. I may find more time, sometime.


The third link there - bugs
In Progress - shows 2 bsdconfig bugs assigned to Devin Teske (bsdconfig and sysrc author) since ~2018, not resolved. It was differences between Glen and Devin's views on bsdconfig's utility that I was referring to, but I've lost track of those PRs.

Except this one, which contains my incomplete patches and rationale, and success re fixing 12.4-RELEASE:

https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=238314

Are we clear that I haven't blamed Glen, nor re@, nor dteske@ for any of this?
 

There's plenty of workplaces that use Apple hardware. In the workplace, you just use whatever's provided, and you're expected to be productive with that. Liking or not liking a particular OS is irrelevant.
Yes. Apple is not gone. But apple is not like it was in the 80s. Why their ads are mostly white ads. Same in the Windows ads. But that was the 80s. Good story. In an Apple workplace that user can be fired.
 
Yes. Apple is not gone. But apple is not like it was in the 80s. Why their ads are mostly white ads. Same in the Windows ads. But that was the 80s. Good story. In an Apple workplace that user can be fired.
Any workplace, frankly... you did miss my point pretty completely here, Rock ... 😩
 
Just pick the tool that gets the job done in a reasonable amount of time, and don't get stubborn about it.
Agree. Whenever I have to work on Windows as part of my job, I think: “How easy it can be to do this on Linux!”. When I work on Linux, I think 'how easy is it to do this on FreeBSD!!". When I work on FreeBSD, I sometimes think: “why is there no problem with this in Linux?”. It all depends on the goal. If the goal is “to work on FreeBSD every day”, that’s one thing. When the goal is “to get the job (for which you are paid) done efficiently and on time,” this is completely different.
In other words, if your job requires you to work on a given project on Windows, then it's better to find the right tools to get the job done, rather than lamenting how easy it could be done on Linux or FreeBSD. And don’t waste time and energy trying to convince the customer to choose another OS. It is expected that in this case the customer will choose another contractor.
 
I am late to this thread, FreeBSD default for desktop environment, since years and years ago users or testers have opened topics why FreeBSD is not default for desktop environment, sterile answers that the server world ate it to the system, in conclusion FreeBSD is a default base system for professionals, and is not for the end user as the vast majority of the world uses default systems with a desktop environment.
 
Users, in case they only had a CL, would be lost. I mean, using a keyboard to create a DVD or to see a thing like
Code:
#
on the upper left of the screen, is their worst nightmare. Thank Jobs for the graphical interface.
 
Last edited:
Users, in case they only had a CL, would be lost. I mean, using a keyboard to create a DVD or to see a thing like
Code:
#
on the upper left of the screen, is their worst nightmare. Thank Jobs for the graphical interface.
Oh not that old lie.
Creating a DVD on the command line? No problem. Seeing # instead of >? Ok, that should ring some alarm bells in the unsuspecting users. But that Jobs created the graphical interface? Nope. Far from it.
 
Back
Top