On Fri, May 16, 2025 at 03:05:31PM +0200, qwx via 9fans wrote:
> On Wed May 14 23:47:26 +0200 2025, [email protected] wrote:
> > Hi all, recently i've decided to make my compsci bachelor capstone
> > project about a plan9 tool and I?m curious to know and it would be a
> > bit relevant for me: what?s the current state of Plan 9/9front in
> > actual practice? Are there any companies, consultancies, or platforms
> > still using it in production or semi-production contexts?either
> > internally or commercially? I understand it became a niche system, but
> > I wonder if there are efforts where it sees real-world use beyond
> > experimentation and the usual stuff we see around here, especially
> > since it seems there was/is proprietary versions around. Also, are
> > there specific roadblocks that you feel stand in the way of some niche
> > use case that it's nearly viable? I would love to take those into
> > consideration while writing my document.
> >
> > Thanks,
> > Lucas
>
> I don't know of any commercial products based on Plan 9 other than
> those already cited, but it is being used in actual practice by a
> number of people, some using real hardware, others through virtual
> machines and drawterm, to do work beyond hobbyist projects. In my
> case, my PhD project is about developing a tool for genome graph
> visualization relevant to biologists and bioinformaticians and was
> initially implemented and showcased on 9front before being ported to
> Linux and others. I've heard of other researchers in Europe using it
> right now in some capacity for their own projects, if only for
> testing. In my branch of research, many existing popular tools
> already work fine on Plan 9 through APE or NPE. The only thing that
> makes Plan 9 impractical for applications is the memory allocator's
> implementation (a lot of software just assumes it can allocate huge
> buffers, eg. on a cluster) and the lack of a C++ compiler and various
> dependencies. The leap from academia to industry is quite small
> especially for non-copyleft projects. Both startups and the largest
> bioinformatics firms use other opensource projects in their products,
> often unmodified and as part of large data processing pipelines. More
> broadly, you could draw comparisons between 9front and OpenBSD, which
> is a relatively larger and much more well-known project, and you might
> quickly realize that outside of ports, the differences in terms of
> capabilities aren't particularly large. For my specific case, OpenBSD
> is just as impractical because of performance and compatibility issues
> across the board. Note that as far as I'm concerned, 9front is Plan 9
> and vice-versa.
>
The question seems for me to not be just "is it used commercialy by
now" but: is it relevant, as is or modified?
Because my own current experience is that some Unix kernels are about
to die because stupidity or lack of thinking about the whole design is
a luxury that can only be afforded by few, wealthy (in cash and in
resources) organizations: the "blob" (I mean this basic "thing" that
is said to "learn" because it grows in every direction and simply shifts
focus in the directions where it finds things to absorb), the "blob"
development is
affordable only to organizations that can waste time and resources to
test every direction---and it shall be noted that one can kill a
"blob" by attracting it with food in one direction leading to a desert
so that it is amidst a sterile zone with a sufficient gap in every direction
that it will waste its whole energy in every direction without
reaching out something in time to survive.
The Linux kernel is a monster, a pile of ad hoc code, and yes: it
works due to the people and organizations investing in it. But for
kernels (I speak of the BSDs) that try to keep pace by incorporating
more and more code taken from Linux without having the hundredth of
the resources to manage it and make it blend correctly with the
kernel, this is a sure way of disaster---the NetBSD DRM integration and
fiasco is for me the exact example. This is the development equivalent
of "thrashing" for an O.S.
So it's time to take a break and think twice. And to keep things "maintenable"
(holding in one hand), Plan9 is still relevant and still a starting
point. If only for inspiration.
--
Thierry Laronde <tlaronde +AT+ kergis +dot+ com>
http://www.kergis.com/
http://kertex.kergis.com/
Key fingerprint = 0FF7 E906 FBAF FE95 FD89 250D 52B1 AE95 6006 F40C
------------------------------------------
9fans: 9fans
Permalink:
https://9fans.topicbox.com/groups/9fans/Td32c2e656057a991-M4c8a145b6dc059fa34c57f68
Delivery options: https://9fans.topicbox.com/groups/9fans/subscription
