This is a good, levelheaded response. But it makes it seem as if the process of becoming a sponsored maintainer was straightforward.
It's not.
A number of months ago, I tried getting trimage (a gui image optimizer, http://trimage.org) into debian. Since I already had a working Ubuntu PPA, I figured it wouldn't be too hard. Packaging instructions for debian are available at various official websites, but contradictionary and they all seem to presume you have packaged for Debian before. Jargon like RFP (request for packaging) and ITP (intent to package) is thrown around frequently without explanation.
Luckily, there is debian-mentors, meant to help upstream developers with becoming a sponsored maintainer. Most of the debian-mentors site links to outdated packaging instructions, using which will mean you will be ridiculed on the way-too-hostile mailinglist. Trying to do 'the right thing' is almost impossible, as most of the feedback will depend on the personal opinion of the person responding. I suspect this is because they, too, don't know which packaging instructions to use or point to.
Trying to get my application into debian has been nothing but a massive, timesinking frustration, and I wish that wasn't the case.
A "good" response? It was very polite and informative, and I do appreciate that - but "you can package it yourself" is still ridiculous. Imagine upstream authors needing to get packaging rights from Debian, Ubuntu, Fedora, Red Hat, Gentoo and FreeBSD just to give most (most!) users a nice way to install a non-broken version of their software.
I've also brought up a counter question that people seem to not be able to answer: How do you get a bad Debian Developer fired?
It seems once they get in they're in for life. Yet, some of the worst DDs who crap on programmers seem to somehow get in charge of projects they actually hate. If they hate it, and they're breaking the software because of this, then why can't I, the original author, get them fired.
At an even simpler level, why can't you report bad DDs and have them reprimanded or taken off the project? That alone would bring out some better responses.
This significantly increases my respect for Zed Shaw.
Yes, he wrote an offensive post crying for blood, but after he had a chance to cool down, he used his own fame and audience to exhibit a well-reasoned response from someone who disagreed. That's a very adult thing to do. In some ways, it's actually better than an apology.
He reacted in the right way, but on the whole my impression is negative. I'm a bit disappointed, because most of Zed's previous rants were humorous, amusing and well researched. But in this case he didn't put any effort neither in research nor writing. He'd stumbled upon a dependency bug and, instead of reporting it, he for some reason concluded that it means that Debian is evil and everybody should boycott it. Weird.
PS. Excuse me, there was one amusing paragraph in the rant. The one in which he equaled Debian, the most independent and free software obsessed distribution out there, to Microsoft.
I didn't write it, Ersek, Laszlo did. That's how sheddingbikes works. Rather than comments, people can send responses as a post and if it's good I post it.
Sadly, it's not frequently that people send in a good response.
Like his message to Zed, it's clear, concise, and seemingly accurate. It also includes a wonderful and well commented example program in expert idiomatic C showing signal handling. I hadn't even known comp.lang.c was a going concern!
It's exactly the sort of quality code I'd expect from someone who writes such good prose. While I'm sure there are some examples of good prose writers who can't code, are there any definitive examples of good C programmers who can't write? Or is this just a "true Scotsman" problem?
I still think my point at http://news.ycombinator.com/item?id=1735344 is important. If Debian didn't throw away unit tests, a lot of these integration errors could get caught automatically.
Actually, I really agree with you here. In fact, it got me thinking about sort of a protection clause in licenses along these lines. Where you can say, "This software is BSD licensed, providing you follow the guidelines in the PACKAGING file."
Then, any OS distribution would be required to follow your wishes or not include your software. If you want unit tests, they include them. If you want all of it, or parts of it, they do it. If they don't want to or can't then they don't adopt your project.
I'm currently just figuring out how to word the license that way so it'd still work.
[+] [-] kilian|15 years ago|reply
It's not.
A number of months ago, I tried getting trimage (a gui image optimizer, http://trimage.org) into debian. Since I already had a working Ubuntu PPA, I figured it wouldn't be too hard. Packaging instructions for debian are available at various official websites, but contradictionary and they all seem to presume you have packaged for Debian before. Jargon like RFP (request for packaging) and ITP (intent to package) is thrown around frequently without explanation.
Luckily, there is debian-mentors, meant to help upstream developers with becoming a sponsored maintainer. Most of the debian-mentors site links to outdated packaging instructions, using which will mean you will be ridiculed on the way-too-hostile mailinglist. Trying to do 'the right thing' is almost impossible, as most of the feedback will depend on the personal opinion of the person responding. I suspect this is because they, too, don't know which packaging instructions to use or point to.
Trying to get my application into debian has been nothing but a massive, timesinking frustration, and I wish that wasn't the case.
[+] [-] JoachimSchipper|15 years ago|reply
[+] [-] zedshaw|15 years ago|reply
It seems once they get in they're in for life. Yet, some of the worst DDs who crap on programmers seem to somehow get in charge of projects they actually hate. If they hate it, and they're breaking the software because of this, then why can't I, the original author, get them fired.
At an even simpler level, why can't you report bad DDs and have them reprimanded or taken off the project? That alone would bring out some better responses.
[+] [-] hapless|15 years ago|reply
Yes, he wrote an offensive post crying for blood, but after he had a chance to cool down, he used his own fame and audience to exhibit a well-reasoned response from someone who disagreed. That's a very adult thing to do. In some ways, it's actually better than an apology.
[+] [-] adambyrtek|15 years ago|reply
PS. Excuse me, there was one amusing paragraph in the rant. The one in which he equaled Debian, the most independent and free software obsessed distribution out there, to Microsoft.
[+] [-] gjm11|15 years ago|reply
[+] [-] zedshaw|15 years ago|reply
Sadly, it's not frequently that people send in a good response.
[+] [-] sradu|15 years ago|reply
If HN people want to get involved with Debian, packaging and so on email me and I'll do my best to help you.
Probably the best way to fix whatever you believe is wrong is by getting involved.
[+] [-] nkurz|15 years ago|reply
Like his message to Zed, it's clear, concise, and seemingly accurate. It also includes a wonderful and well commented example program in expert idiomatic C showing signal handling. I hadn't even known comp.lang.c was a going concern!
It's exactly the sort of quality code I'd expect from someone who writes such good prose. While I'm sure there are some examples of good prose writers who can't code, are there any definitive examples of good C programmers who can't write? Or is this just a "true Scotsman" problem?
[+] [-] btilly|15 years ago|reply
[+] [-] zedshaw|15 years ago|reply
Then, any OS distribution would be required to follow your wishes or not include your software. If you want unit tests, they include them. If you want all of it, or parts of it, they do it. If they don't want to or can't then they don't adopt your project.
I'm currently just figuring out how to word the license that way so it'd still work.
[+] [-] ars|15 years ago|reply
[+] [-] cmsj|15 years ago|reply