near, mid, and long term prospects of dragonfly

[Michael Neumann]

On 1/13/23 01:15, Will Senn wrote:
> Hi,
> 
> I'm definitely having fun learning Dragonfly BSD, but I just read the 
> linux-magazine interview with Matt Dillon from last May:
> 
> https://www.linux-magazine.com/Issues/2022/258
> 
> where he said "As projects go, I think all the BSDs are aging out, 
> including DragonFly, but can remain relevant in this world of Linux as 
> long as we are able to offer installation on modern systems".
> 
> I know that the BSD's are less popular than Linux, but I don't think 
> FreeBSD is gonna go belly up anytime soon, or OpenBSD, or NetBSD for 
> that matter. As for DragonFly, another article I just read, from 
> yesterday, said that an effort is underway to port Hammer2 to NetBSD:
> 
> https://www.phoronix.com/news/NetBSD-HAMMER2-Port
> 
> I admit to being a little confused...  so, before I jump down the rabbit 
> hole and go further trying to learn about DragonFly BSD and how it does 
> things, I've got a few questions for y'all and would appreciate hearing 
> your opinions:
> 
> * Do you see dragonfly bsd as a going concern (or did I miss a defunct 
> notice somewhere)?
> * Do you see it continuing to be actively developed and maintained in 
> the mid (3-5 year) term future?
> * Do you see it continuing to be actively developed and maintained in 
> the long (5-10 year) term future?
> * Or what?

Hi Will and welcome!

First of all, it's really impressive how DragonFly evolved during the 
last ~20 years with such a low manpower, compared to the much much 
bigger projects like FreeBSD or even Linux (or any other BSD out there).

I think the DragonFly project mainly depends on two things:

* Matthew Dillon - without him, I guess the project would die out sooner 
or later.

* People, willing to dedicate their spare time. Small things matter.

And to some degree, payed developers. Of course it would be nice to 
attract more new developers; it seems other projects like OpenBSD or 
NetBSD are doing a much much better job here.

It will always stay niche, and users should be willing to put a little 
bit more effort into it. On the other side, when things go wrong, it is 
much easier to find out why... and get support from kernel developers - 
quite often directly from Matt in person :)

It's lacking support for some newer (laptop) hardware, mainly graphics 
and wifi, but works perfectly fine on a bit older models. Keeping driver 
support up to date is tedious. It looks much better on the server side, 
where new hardware is not release all that often. One thing that is 
missing here is VirtIO 2.0, which is required by some KVM cloud providers.

I am using DragonFly BSD almost since it's beginning, on all sorts of 
hardware. Even once giving a live presentation at my university about 
the HAMMER1 file system, where I was running a demo on an ASUS EEE PC 
(you know it's not very powerful), where I was mirroring the HAMMER 
filesystem in real time with my server, and students could upload files 
onto my server and that file would be mirrored to my EEE PC in real time 
and an on-screen message would pop up during my presentation :). 
Students soon realized that they can chat by naming the uploaded files 
accordingly :D :D. All with a simple `hammer mirror-stream` running in 
the background over poor WIFI.

I continued running DragonFly on smaller servers and for some time on my 
laptop, worked on Rust and Crystal support.

I really wish DragonFly a bright future, the community is awesome (even 
so I am not very active), and most of the time it just does it's job.

Regards,

   Michael

> 
> Thanks,
> 
> Will
> 
>