1 |
Am Mittwoch 28 September 2011, 14:44:06 schrieb James: |
2 |
> Volker Armin Hemmann <volkerarmin <at> googlemail.com> writes: |
3 |
> > > > Breaking the user experience in order to ???fix??? something |
4 |
> > > > is a totally broken concept; you cannot do it. |
5 |
> > > |
6 |
> > > That's hilarious. |
7 |
> > > |
8 |
> > > The Linux developers are _constantly_ changing APIs in ways that |
9 |
> > > break |
10 |
> > > existing device driver code. There are repeatedly wholesale |
11 |
> > > re-designs of some APIs that happen between minor versions of a |
12 |
> > > supposedly "stable" kernel. |
13 |
> > |
14 |
> > which is seriously not a problem and does not matter in the slightest. |
15 |
> |
16 |
> Some perspective may ease the pain here. Folks on this list are focused |
17 |
> on *their personal pain*. Welcome to unix/bsd/linux. (too many decades now) |
18 |
> No pain, no gain. Gui experiences are what consumers see, feel and purchase; |
19 |
> so Volker is very right here. |
20 |
> |
21 |
> The kernel gyrations are all really about something much more important. |
22 |
> *MONEY* |
23 |
|
24 |
well, if you make money with linux, their are many choices for you. Nobody |
25 |
forces you to target the latest kernel. You can always go with one of the many |
26 |
stable releases out there. Look them up. |
27 |
|
28 |
> |
29 |
> Just think about it, on this list in the last few months, we have discussed |
30 |
> how the stock market runs on linux, Some folks use GPU + CPU for very |
31 |
> advanced things, Commercial distros like Apple's offering are making |
32 |
> billions. Android. (on and on). The point is the Linux Kernel is |
33 |
> the battle ground for software deployment, particularly firmware. |
34 |
> An infinite number of "user experiences" |
35 |
> can be packaged and sold on top of the Linux kernel. |
36 |
|
37 |
so what? what does this have to do with linux changing internal apis that are |
38 |
not supposed to be public? (hint: nothing) |
39 |
|
40 |
> |
41 |
> Here's another one: Carrier Grade Linux (runs most of the worlds |
42 |
> communications systems, including most carrier grade cisco gear. Most |
43 |
> legacy comm system at some point now, get boosted on top of private IP |
44 |
> networks run by the carriers (or military). Cisco recommends embedded linux |
45 |
> on their carrier switches and IOS is an unmanaged *hacked* pig, with little |
46 |
> future. |
47 |
|
48 |
see above |
49 |
|
50 |
> |
51 |
> |
52 |
> The "gymnastics" about the kernel and drivers are the public manifestation |
53 |
> of a much deeper battle for embedded systems supremacy using linux. Wind |
54 |
> River, unquestionable the largest commercial offering of embedded solutions |
55 |
> has products based on both bsd and linux kernels. In "ka-hoots" with chip |
56 |
> vendors they routinely offer "enhanced" drivers to companies that build |
57 |
> products, with features never to found in the linux published sources. |
58 |
> Binaries are available and yet clearly violate the spirit of the whole |
59 |
> (whore) open source movement. WHY? *MONEY*. Governments and miltaries also |
60 |
> feed at this trough. Linus would have his tits |
61 |
> slapped together, if he every interfered with these industries. |
62 |
> He in only in charge of the gyrations.... |
63 |
> |
64 |
|
65 |
tell that yourself to make you happy. |
66 |
|
67 |
> Tons of products still use embedded linux for the 2.4 kernel series. |
68 |
|
69 |
and there are even products with 2.2 kernels. What does that prove? Nothing? |
70 |
|
71 |
|
72 |
> Companies build very large data base systems, using the latest technologies |
73 |
> that work with the linux kernel. Often these technologies only appear |
74 |
> for the masses, years after companies use a "in house" version as |
75 |
> the key pillar for commercial success (MONEY). |
76 |
|
77 |
and again, what does that have to do with internal api changes? |
78 |
|
79 |
> |
80 |
> Take for example the company that does backups for one of the worlds largest |
81 |
> and most complicated database needs. The good old US ARMY. |
82 |
> They use linux, the latest open source databases and the newest |
83 |
> file systems like CEPHS, yet they are years away from public consumption. |
84 |
> Well financed companies are buying up the young (phd) experts whom |
85 |
> have hack out versions and code that makes CEPH usable. Billions of dollars |
86 |
> are being made and it's a real threat to Oracle. Customizations |
87 |
> of low level drivers in the latest linux kernel are the key, and |
88 |
> much of that work will not even be introduced to the linux kernel |
89 |
> community..........TOO MUCH MONEY AT STAKE! |
90 |
|
91 |
see above. |
92 |
|
93 |
> |
94 |
> (and you wonder why Oracle hates linux?) |
95 |
|
96 |
yeah, they really must hate linux. One of the first databases running on it, |
97 |
sponsoring btrfs etc pp. That is hate. |
98 |
|
99 |
> |
100 |
> What amazes me if that we get any real progress on the kernel at all. |
101 |
|
102 |
not me. Because keeping internal apis backwards compatible for some out-of- |
103 |
tree code is a sure way to go down the drain. |
104 |
|
105 |
|
106 |
> > They NEVER change user-space APIs and ABIs in incompatible ways. THAT is |
107 |
> > important. |
108 |
> > |
109 |
> > > We have to touch our NetBSD and FreeBSD drivers maybe once every 3-4 |
110 |
> > > years. |
111 |
> > |
112 |
> > and look how much devices they drive - because nobody has to send their |
113 |
> > drivers upstream, nobody does. |
114 |
> |
115 |
> Because embedded BSD, although still viable, does not have mindshare |
116 |
> any more. Most do not care. The battle it to spin your version |
117 |
> of embedded linux, and sell it to the product manufacturers. |
118 |
|
119 |
and thanks to that mindset BSDs are pretty much stagnant. Think about it.. |
120 |
|
121 |
> |
122 |
> > > Often our Linux drivers have to be updated every 3-4 _months_ |
123 |
> > > to keep up with changes in the kernel that break things. |
124 |
> > |
125 |
> > which is your own fucking fault. |
126 |
> > Get your drivers into the kernel. Problem solved. |
127 |
> |
128 |
> Volker is right, again. However, this is where the true |
129 |
> fun begins, particularly when an innovative startup |
130 |
> looks to gain market share in an area where other |
131 |
> have made lots of money. Many drivers, not thought |
132 |
> to be strategic, have little issue. Some vendors, |
133 |
> Motorola comes to mind, put one driver into the kernel |
134 |
> and offer another quietly through vendors or |
135 |
> directly. Many Chipsets have always had "secret hardware features" |
136 |
> and the ability to use those features is still a well guarded |
137 |
> secret and costs tons of money and is often limited to |
138 |
> who can use those chipsets. There are some NDA, if you |
139 |
> violate, your ass is dead. |
140 |
> |
141 |
> Linus a "showboat" and making some serious cash, keeping |
142 |
> the public focused on linux (mindshare) and playing |
143 |
> as puppet as the big boys joust behind the scenes. |
144 |
|
145 |
question: do you think the moon landings were fake too? |
146 |
|
147 |
> From a |
148 |
> modeling point of view, the gyrations of the linux kernel, |
149 |
> chipset's hidden features and the device driver delusions |
150 |
> are very much akin to what is going on in the hacker |
151 |
> (interloper) world. The hilarious twist is the kernel |
152 |
> game is controlled by globalist. Hacking is everybody's |
153 |
> economic playground. |
154 |
> |
155 |
> Why Greg even offers to develop drivers free for folks, |
156 |
> yet hardly any corporations take him up on this generous |
157 |
> offer? |
158 |
|
159 |
because they are scared for their precious 'ip' not realizing that most of it |
160 |
is well known by their competition anyway. |
161 |
|
162 |
> |
163 |
> http://en.wikipedia.org/wiki/Greg_Kroah-Hartman |
164 |
> http://kerneltrap.org/node/7636 |
165 |
> |
166 |
> |
167 |
> Common it's all about *MONEY* The rest is just smoke, mirrors |
168 |
> and BULLSHIT.... |
169 |
|
170 |
you must know it... |
171 |
|
172 |
> |
173 |
> hth, |
174 |
> James |
175 |
-- |
176 |
#163933 |