Re: [9fans] Software philosophy

[Lucio De Re <lucio.dere@gmail.com>]

On 8/19/21, hiro <23hiro@gmail.com> wrote:
>> I never said it was *the* development branch.
>
> lol
>
> but it is.
>
Correct. On two fronts, to coin a phrase. It is active and it is well supported.

It is the pragmatic end of the Plan 9 spectrum, courtesy of Cinap who
clearly would be a Torvalds if Plan 9 had gained the traction of
Linux.

We are all lucky, in a somewhat masochistic sense, that Cinap is not
Torvalds and Plan 9 came a little late on the scene and was "licenced
to kill" itself. Wrong generation, that was, but for those who want
everything and the kitchen sink, Linux is by far more convenient: I am
typing this into Gmail's "basic HTML" in an obsolete and unsupported
version of 32-bit Chromium under Linux Mint. I guess it IS my
preference, even though I greatly regret that there aren't enough
seconds left in my universe to migrate everything I can to my Plan 9
network.

Which, given that this is a "philosophy" thread and both Rob and Russ
are on board at least occasionally, may entitle me to ask: my
understanding is that both R&R use P9P under a version of MacOS "du
jour". I rely on P9P to relieve some of my frustrations with Linux
(and there are hundreds, some go back to my NetBSD days) and in some
crazy way I would tolerate logging in to Rio and I am totally sold on
acme as my editor from my remote workstation to a suitably tweaked
development server 300km away (ssh -fX...). Hell, this saved my bacon
recently after a serious outage: "ssh -fX devbox acme -l acme.dump".
How do I reward R&R for their efforts? And what have they neglected to
contribute to the mental health of suffering people like me in the
last few years :-) ?

In any case, the point is not where I am, but where I come from and
where I wish to be.

I don't run Windows on my premises in any size or shape. I'll probably
regret dropping it from my (cheap, non-plan 9-compatible) HP laptop on
my most recent Linux Mint upgrade, but I have learnt what I can do
with GPT and it doesn't bug me in the least to run Windows on a USB-3
connection to my drive (haven't needed to do that, yet, I may have to
eat these words).

I don't want to run Linux, either, in principle. Now that I have to
think about it, I might not mind using Linux as a hypervisor and
everything else under KVM, but for now the only hardware I own that
supports Qemu-KVM is what I use for the Fossil/CPU server. If I can't
see Linux, I can deal with it. One way or another, though, Linux has
the ability to remind one frequently that it is in charge, in a manner
neither NetBSD nor Plan 9 do.

I don't see how I can create the seamless environment I seek by
glueing together divergent systems such as 9front, 9legacy (my kernel
I label 9miller, a version configured for my server and workstation,
32-bit X86 even though the server is the most advanced platform in my
network, short of my Samsung phone), P9P under NetBSD
(acme-over-remote-X works fine there, too) if these extremely
preferable platforms (and I excluded P9P under Linux, but in fact that
remains the main option, like right now, it's just the least
preferable) continue to diverge, nay, are encouraged to diverge.

And of course, the scarier possibility is that one or more of my
essential ingredients will slip beyond my equipment's ability to run
it. Already, Linux Mint 20.2 with Skype and Chrome is too much for
2GiB of memory in my laptop and I don't have the income that allows me
to keep up with hardware advances.

In summary, I am entirely contrary in attitude to Keith, because my
interest lies in smaller, not bigger (I keep hoping I can afford a
recent rPi model, but I can't entirely justify that, yet).

Another way to get my point across may be to point out that I have no
issue with improvements to Plan 9. Its philosophy is sound and
palatable, much more so than the monstrosities of Windows and Linux
that no sane individual should willingly enslave oneself to. Windows
is still extremely insecure and not even slightly open to a security
audit, although I bet the NSA has no qualms exploiting what the see as
security "features" from the comfort of a source licence paid for by
the U.S.'s subjects; Linux is bloated beyond comprehension and I'm
also not impressed with the OSS's approach to software quality.

But divergence is in no manner "improvement", it needs to be
negotiated back into the "core values". I appreciate that there are
various costs associated with such "upstreaming" and that is why I'm
suggesting that the P9F should take it on, identify the costs and also
arbitrate, from a position of common wisdom, what is "core" and what
is tangential. Note that "core" then becomes a future entity, not a
past one, in this case.

Note: 9fronters may well believe that outsiders refuse to grant them
an identity they feel they have worked hard at earning. What they seem
to miss that even though there may actually be an "inside", the
outside is not the homogenous enemy they paint us as. Interestingly,
what made 9front the success story it is - and Plan 9 the orphan OS -
is precisely that "inside". I see the same happening here in South
Africa with factionalism and in the U.S. with the various nationalist
movements. It is clearly a very powerful human emotion.

A transparent history of decisions in this matter would prevent losing
any interesting proposals - yes, we need better than Git, but Git is
painfully "enough" to start with, even if as I get more familiar with
Git I'm starting to believe, hopefully wrongly, that Plan 9 may have
to bend towards supporting symbolic links to deal with it if it is
going to be a long run - and will raise a chuckle or two when future
archeologists come across it. I doubt they'll be able to do any more
than raise eyebrows when they try that with Linux.

One last, not quite related matter: Plan 9 seems limited never to
provide a conventional browsing experience for its audience. What does
that actually say about Plan 9's future?

That incidentally, must be qualified by the reality that NetBSD
doesn't fare much better: I have certainly not had the stomach to
migrate my workstation  from Linux to NetBSD - even less so my laptop
- precisely because even Chromium isn't "du jour" enough for the real
world.

Lucio.

------------------------------------------
9fans: 9fans
Permalink: https://9fans.topicbox.com/groups/9fans/T9ef6430f3025e731-Mb4d30b419265a251efc01435
Delivery options: https://9fans.topicbox.com/groups/9fans/subscription