1 |
On Sun, 2013-04-28 at 11:59 +0200, Alan McKinnon wrote: |
2 |
> You don't say what your project is, but reading between the lines I |
3 |
> think it's safe to assume it's a somewhat niche project with specific |
4 |
> goals that solves a specific problem, right? |
5 |
|
6 |
This is true (I almost typed True. That's what happens when you code 8 |
7 |
days a week.) |
8 |
|
9 |
> Such projects come with their dep list as you pointed out and this only |
10 |
> affects the machines that project runs on. In eight years hanging out on |
11 |
> this list I don't recall any cases of users complaining about deps of |
12 |
> projects in such a class. |
13 |
> |
14 |
> What we complain about here is basic low-level software changes that |
15 |
> affect much more than just their own little universe, and will do it ON |
16 |
> ALL LINUX MACHINES NOW AND IN THE FUTURE. |
17 |
|
18 |
Well, this will be the case if nobody forks these projects, or writes |
19 |
competing projects. As Dale has pointed out, there already is eudev. For |
20 |
systemd, you have OpenRC as an alternative. For Pulse, you can just use |
21 |
a different DE. I understand that you don't like the direction that |
22 |
these projects are going, and I'm not attempting to convince you to like |
23 |
or use them. I'm just trying to point out that there are viable |
24 |
alternatives. |
25 |
|
26 |
Perhaps you are concerned that Gentoo will require these technologies? |
27 |
That might happen, I don't know one way or the other. If it does, |
28 |
there's always the possibility of forking Gentoo itself (there are |
29 |
already a handful of Gentoo derivatives.) |
30 |
|
31 |
> That is a whole different kettle of fish entirely and is interpreted |
32 |
> very differently from what your project does, this is the point where |
33 |
> the analogies break down. Regardless of how similar two things may |
34 |
> appear on technical merit, the reaction of users is always the deciding |
35 |
> factor. |
36 |
|
37 |
Fair enough. |
38 |
|
39 |
> udev rules changed network names for all recently updated Linux machines |
40 |
> everywhere. |
41 |
> Separate /usr caused changes to many machines not using an initrd, and |
42 |
> will continue to do that for all time. |
43 |
> systemd changes how sysadmins start and shutdown their machines, and how |
44 |
> that works for every service on the host whether the sysadmin likes it |
45 |
> or not. |
46 |
> PA makes deep changes to how the machines handles sound, and the user |
47 |
> for the most part never agreed to have those changes. The user agreed to |
48 |
> use Gnome and the change came in from left field unexpected. |
49 |
|
50 |
Yeah, I fully understand why you don't like these. I don't fully like |
51 |
them either, to be honest. I too experienced some inconveniences during |
52 |
my upgrades, particularly around udev. However, in my case, I'm willing |
53 |
to accept it. For you, there are alternatives. |
54 |
|
55 |
> With your project, the user knows upfront they will need MongoDB, they |
56 |
> make an informed decision about this before ever emerging your code at |
57 |
> all. So your analogy doesn't really hold true. A much better analogy |
58 |
> would be if your project used MySQL and one day you required them to |
59 |
> upgrade to Oracle (and not the free one either...). Plus, you don't |
60 |
> really give them a choice - you also say that all support for all |
61 |
> currently released versions will end in 6-12 months. You are giving the |
62 |
> *apparency* of choice, whilst creating the *reality* of no (or very |
63 |
> little) choice. Does this not look to you a lot like lock-in? |
64 |
|
65 |
Yeah, I think I understand more where you are coming from. And I do see |
66 |
how my project is different. My project is unlikely to be installed by |
67 |
default on all the major Linux distributions, so there's no worry that |
68 |
we will be bringing MongoDB to be on each distribution. |
69 |
|
70 |
I just want people to focus on the fact that there is still choice. You |
71 |
are not alone in your dislike for these technologies, and for you there |
72 |
are options. Yeah, maybe you will be using a technology that is only |
73 |
used by a minority, but we're all used to that on this list, right? :) |
74 |
|
75 |
-- |
76 |
Randy Barlow |