Call for Foundation-supported Project Ideas

astyle

Daemon

Reaction score: 764
Messages: 1,645

In general, it's not the foundation's main task to get ports under control. It develops FreeBSD, not Chromium, Python or KDE. That division of labor between base and ports is one of the things that keeps the FreeBSD ecosystem healthy.
This point, I would like to voice my disagreement about. The ports (as in software packages, not TCP/IP variety) are something that is part of the FreeBSD project. I would think that the Foundation has quite a bit of say in how that part of the project is managed. I know that the ports are a rather open and unwieldy part of the FreeBSD project, so I hope the Foundation will put in some effort into implementing a bit more discipline into ports.
 

ralphbsz

Son of Beastie

Reaction score: 2,517
Messages: 3,377

The ports (as in software packages, not TCP/IP variety) are something that is part of the FreeBSD project. ... I hope the Foundation will put in some effort into implementing a bit more discipline into ports.
While this is a desirable goal, it is also a huge amount of work. Perhaps it would make sense to have a two-tier system, with an upper tier of audited and supported ports/packages, and a lower tier of volunteer-prepared and maintained ones?

The problem is: What to put into the upper tier of foundation-supported packages? I'm sure a lot of people would immediately say "At least one DE", including the whole hairball of what a desktop needs (printing, web browsing, productivity tools). And that right there is a huge pile of stuff already. Others might say "a small set of commonly used servers (NTPD, Apache, some database such as MariaDB or such). I have opinions, but I'm sure few people share my opinions.
 

roccobaroccoSC

Aspiring Daemon

Reaction score: 232
Messages: 774

As you got me reading jail(8) a bit more thoroughly, it should be noted that there is dependency support baked into jail.conf with the pseudo-parameter depend:
Indeed, there is, and our code relies on it.
depend does not work across different files in jail.d though, that was my point.
 

astyle

Daemon

Reaction score: 764
Messages: 1,645

While this is a desirable goal, it is also a huge amount of work. Perhaps it would make sense to have a two-tier system, with an upper tier of audited and supported ports/packages, and a lower tier of volunteer-prepared and maintained ones?

The problem is: What to put into the upper tier of foundation-supported packages? I'm sure a lot of people would immediately say "At least one DE", including the whole hairball of what a desktop needs (printing, web browsing, productivity tools). And that right there is a huge pile of stuff already. Others might say "a small set of commonly used servers (NTPD, Apache, some database such as MariaDB or such). I have opinions, but I'm sure few people share my opinions.
OpenBSD does audit their ports collection, IIRC. Limiting what goes into the ports tree is a good idea. OpenBSD's version of KDE seems to lag behind FreeBSD's, though.

Edit: just remembered that Debian does have 'community repos' that are used by derivative projects like Ubuntu... and so does Arch. Haven't checked Slackware, but I imagine it's pretty 'hands off' on its package management. There are some ideas from the Linux world that FreeBSD can look at and see what might work in their case.
 

Jose

Daemon

Reaction score: 1,164
Messages: 1,374

For the past week or two, I've been thinking about that: What would I personally want to be improved in FreeBSD? Unfortunately, this thread here was terribly disrupted by two topics (KDE and XFS). And the bigger problem is that in my use of FreeBSD (as a small home server, with PF, firewall/routing, DNS/DHCP/NTP server for internal uses, web server, and some monitoring/control software), it is nearly perfect. I really can't point at "make this one thing much better".
The only thing I can think of is traffic shaping support using dummynet for pf. I believe that's already in progress:
First, the 802.11 access point support (not client!). Today, you can configure Linux machines to be APs, and they supposedly work nearly perfectly for a single-node access point. I've tried this with both Open- and FreeBSD several years ago, and the AP driver stack just has too many bugs to be reliable, and I've switched back to using dedicated APs (first Apple, now TP). Since I have not tried again, it might be that FreeBSD today can do it perfectly, in which case please ignore this request.
I second this. I also tried both with Openbsd and Freebsd and failed. It would be so nice to be able to replace the Chinese plastic boxes running some dodgy version of Linux with a nice PC Engines box running vanilla BSD.
 
OP
jrm@

jrm@

Daemon
Developer

Reaction score: 533
Messages: 1,223

Wrong place to ask. This is a user's forum, not a developer's.
In this particular thread it's fine to make such requests. We'll be collecting all the requests in one place and some will be candidates for supported work.

There is always more work than resources for funded work, but we'll still collect all requests so anyone interested in tackling a problem can find the requests in one place. I'll update everyone here when this is ready.
 

Phishfry

Beastie's Twin

Reaction score: 2,855
Messages: 5,854

I would like to elaborate on the NanoBSD request I made.
I don't like to ask for things blindly so I have looked at the problem.

So stock NanoBSD scripts generate MBR images. There is a slow move to EFI on the embedded box front.
What we have in source tree is /tools/tools/nanobsd/ with PHK's work slightly extended.
Then under /tools/tools/nanobsd/embedded/ we have some improvements by Warner.
In general I am speaking about the UEFI formating section of /nanobsd/embedded/custom
As you might know it is possible to enhance nanobsd builds with custom functions.
Warners work under ./embedded extended that to include filesystem builds.

So what I would like is to see that work folded into NanoBSD to create optional UEFI image formats.
Some of the stuff in the ./nanobsd/embedded folder may need updating due to Arm changes.
From /tools/tools/nanobsd/embedded/custom (2015)
o need to promote much of this to nanobsd.sh and friends

I am a willing volunteer to help in any way although I am a script novice.
In the past I have modified Warners ./embedded/custom script to make it work for me.

There is a viable alternative in 'poudriere image' but the images just are not the same as NanoBSD.
It can create just about any image format you desire.
 

mark_j

Daemon

Reaction score: 797
Messages: 1,374

I know this is a pigs might fly request, but, I would like to see Solaris RBAC introduced and be an alternative to MAC which is tedious to set up and maintain and just plain counter intuitive especially for new admins to learn (compared to RBAC).

Yes, I know there's arguments for and against. It's my 2c worth.
 

richardtoohey2

Aspiring Daemon

Reaction score: 347
Messages: 674

And the bigger problem is that in my use of FreeBSD (as a small home server, with PF, firewall/routing, DNS/DHCP/NTP server for internal uses, web server, and some monitoring/control software), it is nearly perfect. I really can't point at "make this one thing much better".
Since feedback/requests are being requested, my situation is similar to the above except it's for more than home servers. FreeBSD has been rock solid and performs well for my production uses - Apache, PHP, MySQL, pf, UFS, wired networking, etc. So please don't lose sight of the server-orientated core/use cases, including the work required to support modern server hardware (e.g. 10G NICs, lots-of-cores CPUs, ARM, NVMe, EFI, etc.)

This isn't a "don't change anything it's perfect" request - there's obviously plenty that people would like added/changed for their use cases (and might prove useful for me, too, one day, obviously).
 

ralphbsz

Son of Beastie

Reaction score: 2,517
Messages: 3,377

FreeBSD has been rock solid and performs well for my production uses - Apache, PHP, MySQL, pf, UFS, wired networking, etc. So please don't lose sight of the server-orientated core/use cases, including the work required to support modern server hardware (e.g. 10G NICs, lots-of-cores CPUs, ARM, NVMe, EFI, etc.)
Completely agree. As a server, FreeBSD is not only rock solid, it is also well built and engineered. One of the reasons I like to use it is: it doesn't make me mad with crappy decisions, half-baked ideas, and sloppy implementations. It's just more pleasant to work with something that was put together with good craftsmanship. OpenBSD is the same way, only things are even neater and tighter. Linux is not at all that way, it's a good-awful mess and cluttered (even though it works well).

The other part I agree with: Make sure normal boring server workloads (the typical FAMP stack) runs well, and runs efficiently on server hardware. Not necessarily on tiny servers (people who use their Raspberry Pi 0 as a database server are like people who build a model of a cathedral out of matchsticks: fun hobby, but not very practical), nor necessarily on gigantic servers (folks who own $2M mainframes or $100M supercomputer have plenty of operating systems available already, and are probably not a good target for FreeBSD efforts), but anything from a $100 motherboard with 8gig and a handful of cores, to the SMP with $5000 worth of CPUs, 1/4TB of RAM and a few hundred cores, the typical server gamut.
 

Alain De Vos

Son of Beastie

Reaction score: 870
Messages: 2,826

Supercomputers run linux. So a logical question is where did bsd took a "wrong turn" ?
 

drhowarddrfine

Son of Beastie

Reaction score: 2,503
Messages: 4,451

Supercomputers run linux. So a logical question is where did bsd took a "wrong turn" ?

Supercomputer companies are the ones who took the wrong turn, not BSD. Probably for the same reason Google started with Linux. It's cause it's what they were comfortable with and used in school and no other reason.
 

Alain De Vos

Son of Beastie

Reaction score: 870
Messages: 2,826

In school , that is 30 years ago, we mainly used solaris. At the job it was sco.
 

sidetone

Daemon

Reaction score: 942
Messages: 1,895

Supercomputers run linux. So a logical question is where did bsd took a "wrong turn" ?
It doesn't mean FreeBSD took a wrong turn. Maybe didn't have the right software or lacked for graphical processing, or didn't move fast enough in the market place. Popularity is a big reason, and also FreeBSD is a product that markets itself by showing what it can do, than have the exposure that Linux has. One reason for the lack of popularity is the amount of setup and customization required for FreeBSD, especially for a desktop.
 

Beastie7

Aspiring Daemon

Reaction score: 652
Messages: 739

I'm sure much of these markets were heavily saturated by Red Hat since it was the first major open source company. SUN was too complacent with SPARC and didn't follow the intel gravy train. Then CentOS/Scientific Linux came along and gobbled up all HPC type customers. I remember when Dell became the first OEM to ship RHEL with their servers. That was when Linux really blew up in the server space.
 

mark_j

Daemon

Reaction score: 797
Messages: 1,374

In school , that is 30 years ago, we mainly used solaris. At the job it was sco.
To answer that, I think I would answer more broadly.
In my experience, high school & college/university kids get exposed to Windows, Mac and Linux. Why Linux? Well the back history goes into the BSD law suit and everything that surrounds it but that's sort of irrelevant. Regardless those 3 are what are in their consciousness. Then later they become CIO, CEO or whatever acronym and when pushed to save a buck go to the first free OS they know; Gnu/linux. Universities and schools adopt gnu/linux because it's free.

It's then self perpetuating. Students become managers, lecturers, high school teachers and they fall back on what they know. Then corporations see money in supporting it and the cycle continues.

Look at gnu/systemd/linux now especially with SOCs where the likes of the Raspberry Pi now dominate (in my experience) high school technology coursework, where gaggles of kids learn linux and, eek, python.

I think the foundation should push hard to align itself with a SOC like Odroid or RPI to get it out into schools and spread the love and critical mass for freebsd and perhaps add another free OS to the psyche of future CEOs, CIOs, lecturers, teachers, managers etc.
 

drhowarddrfine

Son of Beastie

Reaction score: 2,503
Messages: 4,451

I think the foundation should push hard to align itself with a SOC like Odroid or RPI to get it out into schools and spread the love and critical mass for freebsd and perhaps add another free OS to the psyche of future CEOs, CIOs, lecturers, teachers, managers etc.

That's an actual grassroots call for action. Something we could all do on our own locally and, eventually, form a national organization aligned with and supported by the foundation.

For that matter, it's something the foundation can start and get interest generated nationally.

Someone should grab the ball and run with it.
 

mark_j

Daemon

Reaction score: 797
Messages: 1,374

Indeed, it is somewhat grass roots. You are correct.

I have had two approaches, both unsuccessful.

1. I am friends with a high school technology teacher. They use RPIs for teaching robotics, programming etc. It's more fun than super serious, but nevertheless. (Edit: The kids are 15+ years of age, they can probably root/jailbreak a phone for crying out loud!)

I asked him to try using FreeBSD (what's FreeBSD he says! :rolleyes:). He was willing but the school would only accept those on the website because the kids had to install and run their own systems and they deemed it too hard. (You'd think LEARNING was a problem given that attitude!)

I even wrote a step-by-step guide. That didn't help, the school administration still didn't entertain the idea.

Now, if I had got the Raspberry Pi Foundation to make FreeBSD an official OS on their website, I think things would be different.

2. I have used (professionally and privately) a lot of SolidRun SBCs and SOCs and have tried to convince them to put work into providing official support for FreeBSD. I might as well hit my head on a brick wall. They have somewhat supported them on non-networking appliances like the Cubox et al, but it's a desert elsewhere.
 

astyle

Daemon

Reaction score: 764
Messages: 1,645

I asked him to try using FreeBSD (what's FreeBSD he says! :rolleyes:). He was willing but the school would only accept those on the website because the kids had to install and run their own systems and they deemed it too hard. (You'd think LEARNING was a problem given that attitude!)
gotta sell the idea of FreeBSD a bit better than that. Otherwise, you just come across as a FreeBSD fanboi who doesn't care what others think, and just chants. 'FreeBSD is the greatest, FreeBSD is the greatest, FreeBSD is the greatest! Try it!'. Figure out what their difficulty is with getting stuff to run on Pi, and show them how FreeBSD does the same thing so much better. If you look at comments people make about what prompts them to move to FreeBSD from Linux - the reasons vary by the person.
--
The school board would rather stick with officially listed stuff, because it's supported and easier to support and troubleshoot if something goes wrong. It's actually the exact same logic employed by the Forums and Foundation, if you think about it.
--
One idea for the tech class could be an 'extra credit' challenge: Make a backup image that you can burn right back onto the board later (if need be), install FreeBSD, share results with class. And if you can stick with FreeBSD for the rest of the class, that's guaranteed 10% extra credit. ;)\
--
Man, I just went off topic and hijacked the thread... 😩
 

bakul

Active Member

Reaction score: 92
Messages: 164

Now, if I had got the Raspberry Pi Foundation to make FreeBSD an official OS on their website, I think things would be different.
This is simply not going to happen. 'Pi founders all use Linux and originally did substantial work to get Linux working on it. A huge user community has coalesced around the 'Pi running Linux and for them FreeBSD doesn't buy anything -- its API is just like Linux but far fewer devices work with it.

What would help users of FreeBSD on 'Pi is a framework or a stub driver that allows you run drivers in user mode. Sort of like fusefs but for drivers. The same driver should work as a kernel mode driver with a #define change. Such a thing would speed up driver development for simple IO devices, the kind one attaches to a 'Pi.
 
Top