GhostBSD jumps ship; Drops FreeBSD to use TrueOS (PC-BSD)

Chris_H

Chris_H

According to their latest announcement. GhostBSD is dropping FreeBSD as their base, and taking up with TrueOS.
Their claim is people complaining about increasing problems with setup, and maintenance. Some claiming the problem is not using OpenRC. I remember someone in the Forums here, bringing that up sometime in the last few days, and them being slammed. Anyway, if you're interested in the details: GhostBSD is switching it's system base

--Chris
 
Soon we will have people complaining about FreeBSD not using SystemD. Oh, well, that actually already happened. :D

I don't know if you saw it on that other thread but there is at very least one BLOCKER for OpenRC come to FreeBSD, but those people don't care for stability just coolness.

Also, the amount of the work eventually spent to switch to OpenRC (what would include re-write all rc scripts over the time) would do better importing/merging/finishing the NextBSD work ( IMO ).
 
drhowarddrfine said:
What is the system base for TrueOS?
Click to expand...
Essentially FreeBSD. That is, Kris cherry picked what he wanted (some of which he wrote while with FreeBSD), and bolted on, or "enhanced" the rest.
It's kind of a slick system. But not my cup of tea. Nor do I really have any use for it.

--Chris
 
What disappoints me about these "other" FreeBSD's. Is that FreeBSD became so fragmented. Which means you have several small groups struggling to keep their version(s) useful, and interesting. When what is truly needed. Is to rejoin the group. Pooling their resources, and skills. Making one truly great product.

--Chris
 
It seems to me that both TrueOS and GhostBSD have similar goals - i.e. the development of a consumer oriented OS. So the "rebasing" does seem to make sense. Both are not really so different to anything else FreeBSD based, whether closed or open source.

I don't really see the issue, nor see it as "fragmentation". As with other *BSDs, FreeBSD can choose to use whatever comes from these projects that might be useful. But "pooling their resources" could also mean getting all the nonsense you don't want along with the good stuff - a "too many cooks" situation.
 
Chris_H said:
When what is truly needed. Is to rejoin the group. Pooling their resources, and skills. Making one truly great product.
Click to expand...
The problem with that idea is that there isn't a single way to define "great". Different users, supporters and developers of FreeBSD have different goals and aspirations. And this is not only true for FreeBSD, but also for other operating systems it competes with (*BSD, Linux, other Unix variants, Windows, ...). For example, I personally don't care at all about GUIs and windowing systems, but I want a server OS only, with certain characteristics. As far as I'm concerned, a very simple console, enough to get the server onto the network is all I need, and then I never install anything that needs or wants Xwindows. But I know that a lot of other folks here on the Forum are very adamant about wanting to use FreeBSD as a desktop system with a GUI. To each his own.

The problem with "to each his own" is: optimizing the whole OS stack (not just the kernel, but the lower operational layers like init(8) and configuration (text files in /etc versus databases), and the philosophy of administration, has to be done differently for different use patterns. An extreme example is sytemd, which was written from the vantage point of a laptop with a GUI (and has certain advantages in that setting), but is foot-shaped gun when deployed on headless servers.

From this viewpoint, it might perhaps be a good thing that the TrueOS and GhostBSD people are working on a different code base: less opportunity for them to break FreeBSD. Their goals are clearly different from mine:

"TrueOS is a cutting-edge FreeBSD graphical desktop operating system designed with ease-of-use in mind."
I don't want cutting edge. I want really well tested systems. I don't want graphical desktop; tuning or organizing the system for graphical desktop will just get in my hair. And I don't care about ease-of-use; I understand that complexity exists in the real world, and trying to hide it behind easy layers just breaks things later on.

"Built on top of FreeBSD, GhostBSD provides a simple desktop-oriented operating system pre-configured with the carefully selected minimal commonly used set software required to start using it to its full potential."
I don't want a desktop. I can configure things perfectly well myself, and I'd rather not have anything pre-configured, other than a minimal functioning base system, roughly at the level of V7 Unix with networking. I don't use the system to its full potential, rather for a very narrow and specific purpose.

As you can see, I'm not in favor of them rejoining the group. But other people will likely have very different opinions, and as long as FreeBSD works reasonable well for me, I'll continue to tolerate these other opinions. If GUI and ease-of-use people break FreeBSD for me, there are many other OSes I can use.
 
ralphbsz ,
You make some nice counter points. I generally follow your path. With the exception(s) that my "build"/development box has a DE. I chose that approach because I find I have a better/wider "macro" view working in that environment. I can have 150 tabs open in my editor/IDE, and at least as many in my (web)browser (for any research). So I build all my "deployment" media on this server/DE, and go from there. Where all my servers are console-only. An environment I find equally comfortable in that context.
This is what has always kept me riding the (Free)BSD train. It's everything-to-everybody. Who doesn't like choices?
But, and this probably speaks to your assertions; I do have a problem with the way pkg(8) was implemented. In that it is a requirement. While I fully appreciate the convenience it provides. I am ill served when I am building/using for my chosen environment. It just gets in my way. Making it harder; if-not, impossible to accommodate my end goal. So, again to your point(s), it can be difficult-to-impossible to have/keep everyone (all developers) on the same path, with equal goals.
My only point above; is that it would be advantageous to have more (skilled) develop(ers|ment) on the same team. :)

Thanks for the "counterpoints", ralphbsz !

--Chris
 
ralphbsz said:
An extreme example is sytemd, which was written from the vantage point of a laptop with a GUI (and has certain advantages in that setting)
Click to expand...

It has certain advantages in the server world too (launching thousands of VMs/day). What doesn't seem to have advantages in servers but seems to be getting worked on is drm, video and wifi drivers.
 
The big issue with this kind of discussion/practice is no one ( talking about the people who are leading either FreeBSD and GhostBSD/TrueOS ) tell/set who is the actual target audience clearly. I mean, who is the FreeBSD target audience?

We ( users ) are the most ones telling it to everyone that came in here or IRC or mailing lists who is the FreeBSD target audience and what kind of behavior is expected from them when the things get too out of normal, but there is no official direction ( as far I am aware ) from the project nor the foundation about this subject. We can just assume we are 'right' because no one come and say NO, you are wrong.

The point is, what is a 'Desktop OS' and what is 'Server OS'?

For most people a 'Desktop OS' must be something like MacOS or Windows ( both have clearly defined audiences ). Same can be said about a 'Server OS', for some MacOS server is perfect and they don't need anything more than that, there are plenty of business running Windows Servers, others need carefully tailored setups or mainframes... Still, we can discuss, what is easy to use? For me i3 is easy to use, to my father it is black magic.

Why not do a research about the current users and of a defined target audience, preferences, suggestions etc. and follow the path brought by the result of it? It is easier to get more users/developers joining a community/project when it is clear for people they are the target ( or not ).

About desktops in particular, I guess the vastly majority of FreeBSD desktop installations are using: ( mostly ) WMs, followed by XFCE and then KDE. At same time there are a bazillion of DE/WMs available on ports used by 2 people each. This looks cool but it leads to a lot maintainership overhead.

IMO the right way to deal with it would be either: adopt the code most people use or write its own home grew desktop/wm, and drop everything else. But if something like that would be done, do it right. Lumina is a nice initiative aiming FreeBSD 'end-users', but I bet 50%+ of FreeBSD desktop users are using WMs and not a full feature DEs...

[EDIT]

Linux for instance has 1000+ distribution but if you look at the most successful ones they always have a clear target audience, for instance:

Arch Linux: programmers, developers.
Ubuntu: people aiming user friendly, magically configured stuff - and currently include enterprise servers.
RedHat/SuSe: enterprise, servers and desktops.
Debian: servers and desktops.

Now if you look at the ones that actually support desktops they almost always do actively support just one or two, three maximum, and sometimes don't ever have anything more in the repositories, and if people want to use something different they have to create and maintain it or use third party repositories.

So, the FreeBSD Server-side audience is more or less well defined by Juniper/Netflix etc., but not the desktop one. IMO, FreeBSD should market/work/develop the Desktop-side in similar way of Arch Linux/Gentoo -> programmers/developers - people who can and often prefer to use/maintain a desktop using the same or similar tools used on servers, or more towards to server usage ( I m not implying both are the same but it is easier to compromise ).
 
I don't think FreeBSD should market itself to desktop users at all IMHO. MacOS / iOS users look forward to auto-magically configured / pre-installed systems (it just works, etc), Windows users are more or less the same (yes -> next -> ok -> i agree, etc) but probably think FreeBSD is a Linux distro, while the rest are using Ubuntu (magically configured stuff). I know those are fairly broad strokes but still.

The ones that will realistically consider FreeBSD (as a desktop) are mostly Arch, Gentoo and/or <insert_niche_linux_distro> users. From those users, some will install FreeBSD and a subset will continue to use it. For those that eventually switched, the bottom line is whether or not FreeBSD provides a comparatively superior desktop experience compared to Linux. If it does, they'll keep using it. If it doesn't, they'll go back.
 
xtremae

If some project ( commercial or not ) does feel that part of it should not be marketed for some reason, it is because that part is worthless and then it should be completely dropped to freely resources to deal with what is considered important and worthy - specially when they have very limited resources.

Note that marketing something is not just about advertising ( marketing != advertising ) and may even not include it, but have a properly designed project/plan of how the product is positioned at the market, who are the customers, what they want, and what/how you should delivery it for them.

The current situation, I feel while doing some minor ports works, is more like throw everything each one want in the ports tree to try get every single user pleasured, and then spend a lot of time ( read resources ) to maintain a ton of ports that 2 people use, none of them maintain or do anything, and still want everything updated ASAP ( and they are not essentially wrong because they are not breaking any rule ) - what means bug reports to be handled, patches to be written, etc.

That said, if Arch/Gentoo-like users are supposedly the target or the ones that worth, instead of maintaining a ton DE/WM with 2 users each, research what those "worthy" users want need and design the thing ( what may be something new or something that already exist ) for them, and drop the rest. I give some ideas in HERE.

Are you aware we have +32K ports on the tree and currently should have about 400 porters to maintain them, almost all of them working at they spare time? Yes, there are a lot of users who maintain ports and many of them are nothing more than 'the maintainer'.

About the TrueOS desktop initiative, they are apparently looking for Ubuntu-like users, exactly those users that do not live without Netflix, Steam, Spotify, and similar that do not work on FreeBSD, or work poorly, or need some hack to work... exactly what those users don't want.

[EDIT]

What I want to say is the way the thing is being managed is exactly the way you should follow when you want to be left behind.
 
lebarondemerde said:
If some project ( commercial or not ) does feel that part of it should not be marketed for some reason, it is because that part is worthless and then it should be completely dropped to freely resources to deal with what is considered important and worthy - specially when they have very limited resources.
Click to expand...
Of course, we wouldn't be having this discussion if resources were unlimited.

lebarondemerde said:
Note that marketing something is not just about advertising ( marketing != advertising ) and may even not include it, but have a properly designed project/plan of how the product is positioned at the market, who are the customers, what they want, and what/how you should delivery it for them.
Click to expand...
I don't think this is an issue for FreeBSD since the project has a well defined and appealing feature set as a server OS. As a desktop OS, it's just not incredibly compelling to satisfy the particular market, at least not for most people. Trying to double down on desktops just doesn't make sense to me and i wish i could be wrong.

lebarondemerde said:
The current situation, I feel while doing some minor ports works, is more like throw everything each one want in the ports tree to try get every single user pleasured, and then spend a lot of time ( read resources ) to maintain a ton of ports that 2 people use, none of them maintain or do anything, and still want everything updated ASAP ( and they are not essentially wrong because they are not breaking any rule ) - what means bug reports to be handled, patches to be written, etc.
Click to expand...
This is a result of bike shedding and poor management but thankfully it's never late to start dropping the fat. Countless ports were created because certain users needed them. With those users gone, the burden has fallen on the shoulders and the good will of the very few. This is a very big problem.
 
xtremae said:
This is a result of bike shedding and poor management but thankfully it's never late to start dropping the fat. Countless ports were created because certain users needed them. With those users gone, the burden has fallen on the shoulders and the good will of the very few. This is a very big problem.
Click to expand...

You are getting the point. For instance, based on what I see in here, IRC and mail lists, the majority of FreeBSD users that use full featured DE use XFCE and it is very well maintained, and also KDE (now being maintained with help of Qt Company, it seems).

Then there are Mate, LXDE, LXQt, Cinnammon ( was broken from long time and at least until some time ago, if it still isn't working ), and probably something I don't remember now ( and I am not ever talking about Gnome with its hard SystemD dependencies ).

IDK the state of all of it but I guess most are poorly maintained because of very few users of it, and so why keep these things? What are the real advantages or unique features of these bunch of DEs ( those are many ports that often should be updated at once ) in relation to XFCE? I don't see any but some minor different details.

What I was pointing is, for instance, we "know" most people use XFCE ( and also KDE ), it is solid, well maintained and the alternatives offer no advantage ( with the exception of pcmanfm what is useful for WM users ). So, we know this part of the FreeBSD desktop market, and then why not drop everything else, and adopt XFCE as the only supported full featured DE ( + KDE )? Not to say poorly maintained ports are ugly and looks lazy.

That would freely a lot of resources that eventually could be interested in improving the XFCE and/or KDE support/integration on the system. Same work for WMs ( but WMs alone are often easy to maintain ), and everything else related with desktop. But to archive that, beyond the project leading, there is a need to do some marketing to properly know what should be dropped and be kept.
 
... talking about systemd:
xtremae said:
It has certain advantages in the server world too (launching thousands of VMs/day).
Click to expand...
And nasty disadvantages. If Lennart had had server computing experience, and respect for it (supposedly, he has neither), he would have designed something completely different. Matter-of-fact, if a group of competent and responsible engineers from the large commercial users of Linux (there are many, and I've been an employee of several of them) had designed a replacement for init, it would probably have been very different too.

Alas, in the Linux world, it is what we have now. Like it or not, we'll live with it, and we'll have to be productive.

lebarondemerde said:
The big issue with this kind of discussion/practice is no one ( talking about the people who are leading either FreeBSD and GhostBSD/TrueOS ) tell/set who is the actual target audience clearly. I mean, who is the FreeBSD target audience?
Click to expand...
That is a very good question. Another question discussed here is "marketing" (which is not a bad word, but a very necessary part of a product.

I would go even one level deeper and ask a meta-question: Given that FreeBSD is mostly run by volunteers (with very few people paid to work on it, and very few companies using in-house developers for stuff that gets fed back into the public release), who can even make this decision, and can such a decision even be made coherently? Is there even a marketing group, is it coherent, and does it know what its direction needs to be?

Maybe the answer is: every volunteer does what they feel like, and we hope they're productive this way. So far, the result has been pretty darn good.
 
Given the current FreeBSD state, the good and the bad, IMO the direction to be followed is to become the "BentleyOS". Yeah, it is niche, old, heavy, expensive, but it is the best. Linux is Ford. MacOS is Rolls-Royce ( too fluffy ).

Of course to marketer something like that there should have some proper preparation ( this is a long time investment ), the work to come from Intel ( Thread 66293 ) already help a lot, and niche products are not intended to be full featured but specialized. The clearly bad things they cannot get rid of are worked to seem good ones, and so the first thing to do would be dropping stuff.

If desktop is important choose very few alternatives as I already pointed before ( this is not MacOS who can afford imposing just one ), invest on them and drop everything else ( desktop would need to have its separated policy ). If it is not important drop it completely, or cut to the very minimal just to not lose the functionality.

Most FreeBSD people apparently use WMs, I am no expert but to develop ( or fork ) a WM should not be that "expensive" since every month one guy come out with a new WM.

Look what people are using ( tilling: i3, awesomewm, bspwm, xmonad ), study them ( and they bug trackers ), discover what the users like and miss on them, what are they design flaws, design a proper one using that data, don't care for portability ( leave it for server stuff only ), and heavy integrate it with the system ( make it personal ).

The big blocker to do something like that is the obvious need of specialized marketing professionals, and that cost ( a lot of ) money. However, all partners certainly have very capable marketing departments and probably wouldn't harm to ask if they can lend them a bit. :)
 
The direction of FreeBSD is determined by those whom contribute; either by developed/maintained code, or by $$ (paying to have something done).
You want to turn FreeBSD into a Bentley lebarondemerde ? (good choice BTW). Then whip up some Bentley code, and submit it. :)
Alas. That's about as good as it's going to get, I'm afraid. But all-in-all, I wouldn't choose any other OS as my "daily driver" (pardon the pun).

SystemD (and Linux)?
Well, I know that one of the original distros, disavows SystemD. What's more; it's the closest FreeBSD (feeling) Linux I know of. Namely Slack Linux.
OK it's not the newest kernel. But if it ain't broke. Why fix it? At least one Linux distro stood up for what's right! :)

--Chris
 
Chris_H said:
The direction of FreeBSD is determined by those whom contribute; either by developed/maintained code, or by $$ (paying to have something done).
You want to turn FreeBSD into a Bentley lebarondemerde ? (good choice BTW). Then whip up some Bentley code, and submit it. :)
Alas. That's about as good as it's going to get, I'm afraid. But all-in-all, I wouldn't choose any other OS as my "daily driver" (pardon the pun).

SystemD (and Linux)?
Well, I know that one of the original distros, disavows SystemD. What's more; it's the closest FreeBSD (feeling) Linux I know of. Namely Slack Linux.
OK it's not the newest kernel. But if it ain't broke. Why fix it? At least one Linux distro stood up for what's right! :)

--Chris
Click to expand...


You bring up a curious point about direction. Would you say that FreeBSD has a leadership/management (i.e, vision, trust, morale) problem?

--Vic
 
Last edited by a moderator:
Chris_H said:
Then whip up some Bentley code, and submit it.
Click to expand...

That is not really how it works! If you follow IRC and some mail lists you will find some situations like:

There is an issue with bug reports that some dev assign it to him/her and for some reason do nothing later ( nothing dramatic and happens everywhere, people just forget some things ), but one time the bug is assigned to some one it is not more in the pool and being "watched", for obvious reasons. There were some discussions and people found out the best solution would be to automatic un-assign it if there were no movement from the assignee between X (like 2 months) time.

Some folk worked on it, modified the bug tracker code, bring the patch, and bug-meisters said NO, blanket.

And there still has this kind of situation: Thread 59705.
 
Remember, developers and port/package maintainers are for the most part unpaid, volunteers, amateurs, hobbyists. They do something because they want to. A bug tracking system exists for their convenience, and because they enjoy it, not as a punishment or shaming mechanism for making them work harder.

This also implies that developers/maintainers will sometimes work very slowly. But that's not specific to free software projects. Even when I was a paid developer, I would sometimes have bugs assigned to me that were 1 or 2 years old. This may simply mean that fixing this bug is not considered a high priority, that it affect few people, doesn't pose high risks, or has easy workarounds; or perhaps that it is hard to fix now, but will become easy to fix when some other major work will be going on.

And to keep a crew of volunteers working together requires keeping the community spirit going. That unfortunately may require admonishing developers who interact in a fashion not compatible with the group, and if they don't change their behavior, excluding them. In a commercial setting this is much less necessary (it is rare), because there the reward of the great big paycheck causes people to behave somewhat better.

So there is no point getting upset about lack of maintenance. Do it yourself, or use a different OS.
 
You must log in or register to reply here.
Back
Top