1 |
Hi! |
2 |
|
3 |
Although I am not a Gentoo developer, I saw the wiki post and thought I would |
4 |
donate my 2-cents. :) FYI, my background is science/engineering not software |
5 |
development so this might give you a little different perspective. |
6 |
|
7 |
>Questions to answer: |
8 |
>Why do you love Gentoo Linux? |
9 |
|
10 |
Over the years I have migrated from slackware --> redhat --> debian --> |
11 |
Linux from Scratch (LFS) --> gentoo (yea!). My debian period was motivated by |
12 |
(IMHO) the superior package management. However, my need for easy optimization, |
13 |
greater control and the most current software versions (esp. |
14 |
science/engineering apps)lead me to LFS. LFS provided the optimization, |
15 |
control and up-to-dateness (a good learning experience too), but the |
16 |
accumulation of cruft from lack of package management was annoying. For me, |
17 |
gentoo provides all of the above. In particlular, gentoo advantages for |
18 |
science/engineering apps are: |
19 |
|
20 |
1. optimization: even a modest 3-10% performance improvement can make a |
21 |
significant impact on runtime for number crunching apps, i.e a 72 hour |
22 |
molecular dynamics simulation. |
23 |
|
24 |
2. control: easy to play with configuration for tweaking the build (see 1 above) |
25 |
|
26 |
3. current versions: a lot of science/engineering apps are under constant |
27 |
development to take advantage of the latest compilers, libraries, algorithms to |
28 |
tweak performance (see 1 above). |
29 |
|
30 |
4. did I mention optimization? :) |
31 |
|
32 |
>What should our business strategy be? (achim and I have our own ideas, but how |
33 |
>would *you* like things to evolve?) |
34 |
|
35 |
You might have guessed, but I believe gentoo would be appealing to people |
36 |
wanting a science/engineering workstation. This might be an additional market |
37 |
segment for you. (I believe HP, IBM and others market Linux sci/eng |
38 |
workstations.) |
39 |
|
40 |
>What is Gentoo Linux going to be like in 18 months? |
41 |
Here are some features I would like to see in increasing order of pie-In-the- |
42 |
sky speculation. |
43 |
|
44 |
1. kernel ebuilds check for /etc/kernel.config for default kernel config for |
45 |
system. |
46 |
|
47 |
2. a separate (build) config command in portage, |
48 |
i.e ebuild foo.1.0.ebuild config |
49 |
to allow manual ./configure tweaking between automatic unpack, make, install, |
50 |
qmerge for cases where you want a rather esoteric build config that isn't of |
51 |
suffieciently general interest to put in the ebuild script |
52 |
|
53 |
3. app-security packages, i.e. portsentry, tripwire, etc. |
54 |
I might try to contribute some security related packages in the future |
55 |
(yes, I had a box rooted by a bot recently. That's what I get for not updating |
56 |
bind, on a dial-up connection no less.) |
57 |
|
58 |
4. System build configuration management via esr's CML2 kinda like what their |
59 |
doing with CML2+OS at the embedded debian project. Hey, why not? Python is |
60 |
already part of the gentoo system. :) Would allow fine-grained configuration |
61 |
of USE setting. For example USE bar in the build of foo1 but not in the build |
62 |
of foo2 when both foo1 and foo2 of the option of using bar. |
63 |
|
64 |
That's my two-cents. |
65 |
I think Gentoo is the greatest thing since sliced bread. Keep up the good work! |
66 |
|
67 |
Tod M. |
68 |
Neidt |
69 |
|
70 |
--------------------------------------------- |
71 |
Sent using Fidelity Networks, Inc. Mailman. |
72 |
http://www.fidnet.com/ |