posted by ![[personal profile]](https://www.dreamwidth.org/img/silk/identity/user.png)
rfunk at 01:30am on 27/05/2004 under email, fetchmail, free software, programming, software
rfunk at 01:30am on 27/05/2004 under email, fetchmail, free software, programming, softwareThis week I've attempted to be a bit of a minor revolutionary, but things have stalled a bit....
First, a bit of history....
I've been involved with the fetchmail project since before it was called fetchmail, when it was just popclient, a program I'd used since 1994 to move email from one Unix machine to another. In 1996, Eric S. Raymond began maintaining it. In mid-1997 he invited me to be one of the designated "backup maintainers" for fetchmail, intended to pick up the reins if Eric dropped them for some reason. (I've long held that since fetchmail is open-source, anybody could pick it up, so *designated* backup maintainers aren't really necessary.)
Oh yeah, and in 1998, Eric released his landmark essay, "The Cathedral and the Bazaar", which was based on his experiences running the fetchmail project over two years. (We fetchmail developers and users hadn't known we were part of an experiment, much less one that would change the face of the computer industry.) Besides the obvious interesting aspects of the essay, I find two things particularly interesting about it. The first is that there is a line in there directly due to me ("Later, I had to bring delivery via a user-specified local MDA back in order to allow handling of some obscure situations involving dynamic SLIP."); it shows that Eric never understood the real argument I was making, but it's still directly due to my argument. The second thing I find particularly interesting is that the "Cathedral" he uses in his analogy is not commercial software development, as later publicity might make one assume, but rather the type of development done by the GNU project and related groups working on free software.
Fast forward to 2003...
By last year, I was no longer using fetchmail, but I was still keeping an eye on the mailing list. Eric was around only sporadically; by this time he was quite the geek celebrity, working on his latest book, and when he was around it was obvious that fetchmail wasn't of much interest to him anymore.
When he finished the book, he caught up on the fetchmail fixes that had languished in his absence, and he released a series of new versions until October. Then he disappeared again. Meanwhile people found bugs, submitted patches, and nothing happened. He showed up again in January, promised a new release shortly, then disappeared again.
By spring 2004, fetchmail development was starting to look a bit like the old Cathedral model, with the elite wizard working in isolation. He had contributions from others, but there was no indication of how or if those contributions were being accepted, and no update "released before its time". Meanwhile, people were anxious to work on fixing larger problems with fetchmail, but the unresolved smaller problems were getting in the way.
Which brings us to this week...
This past Sunday I realized that it was time to take action. I gathered up some relevant materials and put them on my web site, then sent a message to the fetchmail mailing list laying out the situation and proposing a way forward. Basically, since Eric seemed uninterested in it, the community would take over fetchmail.
In many ways this was quite surreal. I was looking at various writings relating to the situation, and all of them were either written by Eric or heavily influenced by him. One not written by Eric mentioned Eric's list of designated backup maintainers -- a list of which I was a charter member:
After a day or two of discussion, things were going pretty well, and I had refined and focused the plan. I had copied Eric on some of the messages, but his mail server was down -- an ominous sign. I started wondering if something serious was going on. Anyway, even a community-run project needs leadership, and that leadership needs the recognition of the community. Unfortunately, of the four people on Eric's designated backup fetchmail maintainer list, two disappeared years ago, one (my ideal pick to take over leadership) disappeared after December, and I hadn't looked at the code in years. By now, the fetchmail bench isn't very deep; I could only come up with two other people that seem able to step up, and one of them declined despite being very helpful in the process.
The Sorcerer Returns... Maybe
On Tuesday, Eric's server was still down, and someone suggested calling him (his number is available in the WHOIS entries for his domains). Despite having exchanged email for eight years and living in neighboring states, we'd never met or spoken before. But early Tuesday evening I called him up, and discovered that he'd just gotten back online after a four-day DSL outage, and he'd just sent some of us an email giving very brief first impressions on the situation. He said he was about halfway through reading all the email on the subject, and that I'd managed to move fetchmail to a top priority (at least for the moment; he still has some secret project going that's been taking his time). He verified my impression that he was losing interest in fetchmail, but said he still had enough interest left to want a part in any transition. He emphasized that he recognized his responsibility to hand it off to a competent successor. He also agreed with my initial feeling about who'd be the best person to take over the project, and shared my great disappointment when I told him we hadn't heard from that guy since December. He wasn't entirely comfortable with any of the remaining options for running the project; fetchmail has been his baby for eight years, and due to CATB is strongly associated with him, so of course he's very picky about who he blesses to run it. Of course, fetchmail is open source, so anyone could take over without his blessing.
My impression was that he'd read the whole thread, then tell everyone his thoughts. But there's been no word from him since the phone call and brief email Tuesday. I'm holding off on my takeover plans in deference to Eric, waiting to find out more of his thoughts, but if he disappears again we may have to proceed without his blessing.
So I may yet be a revolutionary. Or not.
First, a bit of history....
I've been involved with the fetchmail project since before it was called fetchmail, when it was just popclient, a program I'd used since 1994 to move email from one Unix machine to another. In 1996, Eric S. Raymond began maintaining it. In mid-1997 he invited me to be one of the designated "backup maintainers" for fetchmail, intended to pick up the reins if Eric dropped them for some reason. (I've long held that since fetchmail is open-source, anybody could pick it up, so *designated* backup maintainers aren't really necessary.)
Oh yeah, and in 1998, Eric released his landmark essay, "The Cathedral and the Bazaar", which was based on his experiences running the fetchmail project over two years. (We fetchmail developers and users hadn't known we were part of an experiment, much less one that would change the face of the computer industry.) Besides the obvious interesting aspects of the essay, I find two things particularly interesting about it. The first is that there is a line in there directly due to me ("Later, I had to bring delivery via a user-specified local MDA back in order to allow handling of some obscure situations involving dynamic SLIP."); it shows that Eric never understood the real argument I was making, but it's still directly due to my argument. The second thing I find particularly interesting is that the "Cathedral" he uses in his analogy is not commercial software development, as later publicity might make one assume, but rather the type of development done by the GNU project and related groups working on free software.
Fast forward to 2003...
By last year, I was no longer using fetchmail, but I was still keeping an eye on the mailing list. Eric was around only sporadically; by this time he was quite the geek celebrity, working on his latest book, and when he was around it was obvious that fetchmail wasn't of much interest to him anymore.
When he finished the book, he caught up on the fetchmail fixes that had languished in his absence, and he released a series of new versions until October. Then he disappeared again. Meanwhile people found bugs, submitted patches, and nothing happened. He showed up again in January, promised a new release shortly, then disappeared again.
By spring 2004, fetchmail development was starting to look a bit like the old Cathedral model, with the elite wizard working in isolation. He had contributions from others, but there was no indication of how or if those contributions were being accepted, and no update "released before its time". Meanwhile, people were anxious to work on fixing larger problems with fetchmail, but the unresolved smaller problems were getting in the way.
Which brings us to this week...
This past Sunday I realized that it was time to take action. I gathered up some relevant materials and put them on my web site, then sent a message to the fetchmail mailing list laying out the situation and proposing a way forward. Basically, since Eric seemed uninterested in it, the community would take over fetchmail.
In many ways this was quite surreal. I was looking at various writings relating to the situation, and all of them were either written by Eric or heavily influenced by him. One not written by Eric mentioned Eric's list of designated backup maintainers -- a list of which I was a charter member:
Even when the primary maintenance continue alnog the benevolent dictator model, it's not uncommon for assistant or backup maintainers to be explicitly designated (Eric Raymond has done this for the "fetchmail program he maintains, and Linus Torvalds seems to have done the same for the Linux kernel.Later in that same chapter, I read:
Sometimes, someone just gets too busy, or too tired, to continue being responsible for a program. When that happens, it is incumbent on the maintainer to abide by yet another convention, well summarized by Eric Raymond, "When you lose interest in a program, your last duty to it is to hand it off to a competent successor."And here I was, dealing with Eric himself having apparently gotten too busy to continue being responsible for fetchmail.
After a day or two of discussion, things were going pretty well, and I had refined and focused the plan. I had copied Eric on some of the messages, but his mail server was down -- an ominous sign. I started wondering if something serious was going on. Anyway, even a community-run project needs leadership, and that leadership needs the recognition of the community. Unfortunately, of the four people on Eric's designated backup fetchmail maintainer list, two disappeared years ago, one (my ideal pick to take over leadership) disappeared after December, and I hadn't looked at the code in years. By now, the fetchmail bench isn't very deep; I could only come up with two other people that seem able to step up, and one of them declined despite being very helpful in the process.
The Sorcerer Returns... Maybe
On Tuesday, Eric's server was still down, and someone suggested calling him (his number is available in the WHOIS entries for his domains). Despite having exchanged email for eight years and living in neighboring states, we'd never met or spoken before. But early Tuesday evening I called him up, and discovered that he'd just gotten back online after a four-day DSL outage, and he'd just sent some of us an email giving very brief first impressions on the situation. He said he was about halfway through reading all the email on the subject, and that I'd managed to move fetchmail to a top priority (at least for the moment; he still has some secret project going that's been taking his time). He verified my impression that he was losing interest in fetchmail, but said he still had enough interest left to want a part in any transition. He emphasized that he recognized his responsibility to hand it off to a competent successor. He also agreed with my initial feeling about who'd be the best person to take over the project, and shared my great disappointment when I told him we hadn't heard from that guy since December. He wasn't entirely comfortable with any of the remaining options for running the project; fetchmail has been his baby for eight years, and due to CATB is strongly associated with him, so of course he's very picky about who he blesses to run it. Of course, fetchmail is open source, so anyone could take over without his blessing.
My impression was that he'd read the whole thread, then tell everyone his thoughts. But there's been no word from him since the phone call and brief email Tuesday. I'm holding off on my takeover plans in deference to Eric, waiting to find out more of his thoughts, but if he disappears again we may have to proceed without his blessing.
So I may yet be a revolutionary. Or not.
frustrated
There are no comments on this entry.