| 1 |
On Fri, 2007-03-30 at 03:07 +0530, Anant Narayanan wrote: |
| 2 |
> Sure it's not ideal and I acknowledge that. But portage is tied very |
| 3 |
> closely to Gentoo for historical reasons, and it is not reasonable to |
| 4 |
> expect an alternate package manager to replace it (not in the near |
| 5 |
> future atleast). |
| 6 |
|
| 7 |
Historical reasons aren't necessarily the correct reasons. I'd almost |
| 8 |
say that your sentence has officially heralded the age of Debianisation. |
| 9 |
|
| 10 |
> How about implementing the features you mention in |
| 11 |
> portage? I know what your response would be though: portage is too |
| 12 |
> much "spaghetti" code to even think about it. |
| 13 |
|
| 14 |
Have you ever tried to add features to a frankenstein of a beast? What |
| 15 |
is the value to you in doing something like that? Isn't there more |
| 16 |
value in designing something based on what you've learned instead? We |
| 17 |
can all go all day about this and not convince each other, so please |
| 18 |
let's just drop this line of thinking. |
| 19 |
|
| 20 |
|
| 21 |
> But guess what, if you |
| 22 |
> do succeed in making a patch that adds a feature to portage, it'll be |
| 23 |
> accepted faster than you think. Maybe, given the current situation, |
| 24 |
> that is the best way to provide a "better experience" to the users |
| 25 |
> you are so worried about; atleast for those users who don't want to |
| 26 |
> try out package managers unsupported by Gentoo. |
| 27 |
|
| 28 |
What are you basing any of this on? Sounds like speculation that |
| 29 |
doesn't help anything. |
| 30 |
|
| 31 |
|
| 32 |
> You are comparing Gentoo with the wrong distributions. Both Ubuntu |
| 33 |
> and Fedora have people working on it 24x7, and they are being *paid* |
| 34 |
> to do so. Gentoo is a community distribution which is entirely |
| 35 |
> volunteer driven, and you can't expect it to match with the pace of |
| 36 |
> commercial distributions such as the ones you mention. Debian is a |
| 37 |
> distro you could compare with, and you'll have to accept the fact |
| 38 |
> that they develop *for* the developers, much like Gentoo. |
| 39 |
|
| 40 |
Debian was never a distro that I thought we'd emulate, or should |
| 41 |
emulate. Turns out I was wrong, I suppose. |
| 42 |
|
| 43 |
|
| 44 |
> So, really, I don't care if Ubuntu becomes more popular than Gentoo. |
| 45 |
> Isn't it already?! |
| 46 |
|
| 47 |
Here we agree. I don't think Ciaran is arguing popularity either. He's |
| 48 |
arguing that the compelling case for using Gentoo is what's fading. |
| 49 |
There's a difference. |
| 50 |
|
| 51 |
|
| 52 |
> Point is, the day when more than 50% of the devs feel we need a new |
| 53 |
> package manager, will be the day a replacement will be made. |
| 54 |
|
| 55 |
I'm not entirely sure on your reasons for this statement. If |
| 56 |
developers' don't face any API changes, why should we have to have a |
| 57 |
political vote on which package manager gets dubbed the one true |
| 58 |
official one? Why should it be a popularity contest? Why can it not be |
| 59 |
a technical superiority issue? If there is a compelling set of |
| 60 |
technical reasons to replace portage, why ignore that set? |
| 61 |
|
| 62 |
Portage is more than the package manager. Its life comes from the |
| 63 |
portage _tree_. Portage is just the tool that is used to use that tree. |
| 64 |
If that tool is outdated (and let's be honest, it kind of is), then |
| 65 |
switching it is not actually a bad thing. |
| 66 |
|
| 67 |
In sum, I'm not sure I like this direction of basing technical things on |
| 68 |
political decisions. |
| 69 |
|
| 70 |
Thanks, |
| 71 |
|
| 72 |
Seemant |
Archives