1 |
On 11/21/14 07:00, Rich Freeman wrote: |
2 |
> On Fri, Nov 21, 2014 at 2:32 AM, Canek Peláez Valdés <caneko@×××××.com> wrote: |
3 |
>> It's actually a great thing for a lot of use cases. But it doesn't |
4 |
>> seem that Gentoo will change defaults soon, although systemd works |
5 |
>> great with it. |
6 |
>> |
7 |
> |
8 |
> My (personal) sense is that in the medium-term we may end up moving to |
9 |
> not having any default at all, just as with bootloaders, kernels, |
10 |
> syslog, crontab, mail, etc. That is pretty-much the Gentoo way |
11 |
> everywhere else when there are options. |
12 |
> |
13 |
> As you already pointed out, as long as somebody cares to maintain |
14 |
> openrc and write init scripts for it, there will be support for it. |
15 |
> Many init scripts and systemd units are contributed by outside users |
16 |
> already, and policy is that maintainers cannot block them from being |
17 |
> added to packages (though they do not have to write/maintain them |
18 |
> personally). |
19 |
> |
20 |
> Gentoo doesn't really tend to exclude anything, and inclusion is a |
21 |
> matter of whether somebody wants to put in the work. |
22 |
|
23 |
> Rich |
24 |
|
25 |
Very wall expressed and neutral. This is an "800 lb gorilla" that nobody |
26 |
seems to be talking about; that is embedded linux. Embedded linux now |
27 |
accounts for at least 20 times the number of deployments of linux than |
28 |
workstations and servers combined; some argue it is far more, others a |
29 |
bit less. Regardless, embedded linux is a force that is driving the |
30 |
semiconductor markets. There's not much margin on 32bit or less. etc etc. |
31 |
|
32 |
The main point of embedded linux is take what Rich has articulated above |
33 |
and multiply it by a billion. There is no such thing as standard |
34 |
embedded linux. If Systemd is successful, with very large embedded |
35 |
systems (dozens to hundreds of cores) then it has a future. If it |
36 |
fails in that space, it may survive, but not likely. It will old serve |
37 |
to isolate those distros that go down that path, exclusively, *imho*. |
38 |
|
39 |
Regardless, the smaller, cheaper embedded linux crowd is very unlikely |
40 |
to ever embrace systemd. Why? Glad you asks. Thousands of reasons, but, |
41 |
here are a few: It is very common in embedded (anything) to run |
42 |
multiple and often different rtos (real time operating system) on |
43 |
different embedded systems products, often to circumvent licenses, |
44 |
royalties, duplication, security and a plethora of other reasons. |
45 |
Furthermore, many embedded systems run simultaneous codes on a single |
46 |
core and systemd does not fit into that scheme of things, at all. |
47 |
|
48 |
|
49 |
So, even if gentoo becomes stupid and decides to abandon openrc. Many |
50 |
folks will just move to embedded (gentoo) linux and play "follow the |
51 |
leaders" with bootstrapping there cores. |
52 |
|
53 |
Rest easy as the devs fight this one out. I hope systemd survives |
54 |
and prospers. I can tell you one area of massive failure and that is |
55 |
clustering/cloud computing. Sure the "big dogs" with big buck are |
56 |
claiming to use systemd, but they only roll out binary offerings. When |
57 |
others try to use one of the commercial brands of linux and build a |
58 |
cluster/cloud from the source-codes up, there are all sorts of problems. |
59 |
|
60 |
Hmmmmmmmmmmmmmmmmmmmmmm. Very strange, batman. Very strange. |
61 |
|
62 |
|
63 |
What is going on, is wildly variant. YMMV. But, should I be a sporting |
64 |
man, my money is on the embedded folks deciding if something other than |
65 |
systemd survives. Why do I bet on embedded folks? Easy. I personally |
66 |
know of dozens of folks that code in machine, assembler all the way |
67 |
through any language they choose. They routinely build entire systems, |
68 |
custom on a wide bit of processors. Only a few of those folks are |
69 |
necessary to keep alternatives to systemd alive, prosperous and clearly |
70 |
documented. There are most likely tens of thousands of the folks around |
71 |
the world. Do the math. Each time one of these experts build an embedded |
72 |
(linux) system, it is usually optimized and so wonderful, that companies |
73 |
clone them in counts of thousands to millions of deployed |
74 |
linux systems. The fact that the majority rare require human tinkering, |
75 |
is both a testament to how well they run and the wisdon of these |
76 |
brilliant developers to keep the rank and file humanoids using winblows |
77 |
and OXlooser operating systems. |
78 |
|
79 |
A forking of the linux kernel would be the best thing to happen to |
80 |
opensource, in a very, very long time. The kernel development has become |
81 |
a "good ole boys club" imho. |
82 |
|
83 |
Embedded linux runs everywhere; so rest easy! |
84 |
|
85 |
|
86 |
peace, |
87 |
James |