Am I harsh? I don’t think I am: I am just calling it how I see it. And, I certainly mean no offense by anything I say. I thought initially to just respond to the comments in my last blog, but as you can see, I had a lot more to say on the subject: enough to warrant a new post. So here it is.
Apparently, at least a couple of people believe Gentoo to be a User focused distribution. And that is OK, to be honest. Gentoo isn’t a strict set of anything for anybody, so each person brings in their own vision and their own baggage: they join for their own reasons and leave for their own reasons.
Most of the time, I would venture, people join Gentoo development (or really, any open source project) to scratch their own set of itches. Hell, my itch was getting rid of the cruft NLS stuff and GKrellM plugins (though I do see an XMMS version bump request there too). And you know why? Because I wanted that stuff in portage. I wanted all those plugins, I wanted nls-free gnome installations on my computers, and I wanted to use yahoo messenger, God help me.
If people are expecting something other than a developer platform out of Gentoo, then I would have to wonder: what exactly do you expect? And how will you get everyone to agree with you? Because, now you’re faced with trying to define the user. Surely, Tiago is a user. But hey, so is Alex. Which of them is the right user? And there are other users with different needs: the people who want gentoo to power their firewalls, those who want gentoo on their gaming machines, those who want them to power their sparcs and alphas and other 64-bit platforms, their macs; there are those who want a great desktop system, those who want a better multimedia platform; those who want to replace their digital video recording devices; you name it. Who is the right user? Which is the right audience?
I don’t know the answer to that. Let’s not even forget the “holy grail”: the Enterprise users. If you know the right answer, pipe up.
As for this being some sort of “new-found” developer focus. It isn’t new-found. It’s very old-found. It’s, in fact, original-found. What else would you make of Daniel’s thoughts on the subject?. After all, he founded this thing in the first place. Or hey, read Aron’s observations (though gmane may be a better way for some to catch it).
I’m sorry, but if you want a user distro, there are far better choices out there: ubuntu being the primary one. You get coddled and you get sane and sensible defaults. Gentoo doesn’t do that. We give you the tools for you to build whatever you want. So you can fall into all of the above audience choices and more (thanks to the power of profiles and USE flags, primarily). Hell, go nuts with your compiler flags, we don’t care. Think that flies as a user-centric approach? It doesn’t. It’s a support nightmare, ask any Gentoo developer about it.
And finally, the pi
But please, stop being sad that Gentoo isn’t something you thought it was. We never claimed to be the user centric distribution that people imagine it to be. I’m sorry if reality hits that way, but there it is.
Someone should re-word the philosophy page then[1].. It’s very misleading with phrases like “The goal of Gentoo is to design tools and systems that allow a user to do that work as pleasantly and efficiently as possible, as they see fit.”
Not meant to be a flame, I’m just “calling it how I see it”.
[1] – http://www.gentoo.org/main/en/philosophy.xml
Corey, I appreciate the debate. On the subject of the philosophy, you do yourself and other readers a disservice by picking one sentence out of context. When you read through all three paragraphs, you’ll see that what I am saying is in no-way opposite to any of that. The *tools* that gentoo provides allow you to do what you want — up to and including: rescuing XMMS for yourself!
The tools exist to make the user’s system do what they want. The possibility is open as to “the final form of the raw materials (the source code).”
She felt that since the Gentoo project/product is out there and public, we have a responsibility.
I couldn’t disagree more. If someone scratches his itch and is magnanimous enough to share it, who am I to tell him what his responsibilities are? I think public controls of private business should be as laissez-faire as possible. There doesn’t need to be any public control of nonprofit software, except reasonable legal controls (that’s a different debate) and an individual’s own contributions.
That being said, the one and only problem I have ever had with Gentoo is how hard it is, and long it takes, to really contribute. It’s a lot of trouble to get something from bugzilla to the portage tree. If you don’t have commit access you have to hunt down and bug a dev until he does what you want. User overlays help, but don’t solve this problem because using overlays isn’t as straightforward and sensible as the main portage tree.
But I don’t think it’s Gentoo’s responsibility to make sure I can contribute my crap solution to the main portage tree. It’d just be convenient for me, (and possibly terribly inconvenient for everyone else).
The question is indeed, ‘where do you draw the line?’. I draw it pretty close to the start. I wouldn’t presume to require anyone to pay out of pocket for the web space to host his free software; I wouldn’t demand someone write documentation for their <COUGH>E16</COUGH> software, and I certainly wouldn’t claim that Gentoo has a responsibility to make sure my computer works.
Here’s my opinion: For those who think Gentoo has a responsibility to you, you’re a hypocrite if you’re not contributing software, help, or money.
I agree with kojiro, the wall between developer and user is too high, the line is too sharp. I’ve contributed patches, sometimes to let them rot in Bugzilla, sometimes to see them applied. After all the outrage, how many people applied for to be developer in order to maintain XMMS? You’d guess that in the group of outraged users (or maybe the more sane, rational annoyed XMMS users) there would be some people wanting to scratch an itch. Are there any?
The process of becoming a developer is really opaque, and certainly, you cannot become a developer anymore in order just to scratch an itch. As far as I can see you have monthly bug quota (?), and you have to be part of a team that maintains a lot of software.
So basically you just say to people wanting to maintain XMMS: sorry, make your own infrastructure, keep your own mirrors, we (Gentoo) will never support this, our infrastructure is closed to you. I know there is o.g.o., but in order to have an account there, but first of all you need to part of a project (which implies there are developers backing it) or you need to be a developer yourself. Secondly, the audience for an overlay is much smaller then the audience for a regular ebuild, certainly so for non o.g.o. overlays. Third, you need to maintain your own server/website/mirror infrastructure, which is a huge hurdle.
So summarizing: Gentoo makes it easy to maintain your ebuilds, but the process of those ebuilds (or patches) making it into the tree is a long and hard one. This scares people away from doing work such as suggested above. I know there are problems with ebuilds breaking the tree, but that is a technical problem which asks for a technical solution.
Sorry for this (semi-offtopic) rant, but I just read yours and kojiro’s post and something went just *twang*, this post expresses something that I’ve felt for Gentoo some time now. On the one hand I see developers swamped with work, and on the other hand contributing is hard.
I have been called a ranting idiot for rising the question of how to define Gentoo, and again I have to say it’s very funny to read all this from highly acclaimed longer-term devs like you 😛
This is an actual problem, although some people boldly disagree, and it has to be solved – and no, it has not been solved yet.
Looks like I won’t be around to see whether that happens, for I don’t like to be treated like an idiot for trying to talk about serious isses, but I sure hope it _will_ happend, for otherwise darker times are beckoning.
I sincerly hope it will sort out.
Frilled,
I certainly hope that I haven’t called you a ranting idiot. You’ve ranted, certainly, but that’s what the internets are for (mostly) :p.
I’m unclear from your comment, what the actual problem actually is. And what your propose as the solution.
I’m sorry that you won’t be around for this ride.
Of course you didn’t .-)
I agree that by caring more than I should have I probably resorted to what must have been regarded as ranting. But that’s not the point. The point is dismissing stuff without using your brain first. And by “you”, I don’t target _you_, Mr. “Gu”lleen 😀
++.
Deftly written and to the point.
> Flame on!
Will do 😉 Thank you for your attention to my previous comment.
I agree that _something_ must be done regarding deprecated software, but I disagree with its simple removal from the portage tree. Here’s my solution:
I strongly believe that gentoo lacks official overlays. Not random overlays by random people. Those are fine for their owner’s experimental stuff, but not good enough to be considered official or semi-official. For example, I was told there are at least 3 overlays listed in layman with xmms. What for?
(I understand that overlays are a hot topic that seems to have cooled down as of late, but…)
I should clarify what I mean by official overlay: an overlay that would be recommended in the documentation and/or make.conf, regarding a specific category of software. There should be very few of these, because if we want a mess, we already have layman’s overlays. The overlays could be relatively independent, in both developer-wise and bugzilla-wise, but should work as well as possible with the stable keywords. The point of these overlays from a core developer point of view is to trim down portage.
More to the point, a possible official overlay should be the “legacy” overlay. This would hold things like our friend xmms, the ill-fated gtk1, and I would suppose, most gtk1 apps that are still of value to users (here I go again, I know — but I’m trying to keep a balance here).
The big feature of this particular overlay, now from a developer point of view, is that there is no support for the software in this overlay, either because upstream is dead, or because no developer cares. This software must compile, link and run, and that’s it, the remaining bugs are to be left alone. The existing patches should of course be used, but devs shouldn’t feel forced to keep nurturing those old apps. Being kept out of the tree, I believe that gentoo’s reputation wouldn’t be harmed by the poor quality of this software. It is deprecated, after all, and those that should choose to use it must be made well aware of this.
My vision for the official overlays in general is for them to reach the same category as universe/multiverse of ubuntu (which, although community efforts, are fully acknowledged in docs and GUI system configuration apps, and in practice, everybody uses them) or fedora’s DAG repository.
Sorry if this has been discussed already. Powered by rant 2.0 🙂