1. 80

Recently, I’ve seen a couple of deep-links to crass statements on mailing lists again.

None of them has lead to quality discussion. The discussion about these is predictable:

a) The discussion becomes a meta-discussion about whether such a communications style is good or bad.

b) Almost anyone just reads the mail, but not the mail it was a reply to or the discussion that comes out of it. (In which the poster might even admit that they were wrong) Mailing lists are deeply contextual. What would be an offhand dismissal on this very forum here might be an “everyone already knows my stance, so I won’t go in deep” on a core mailing list. This leads to people in the know having to explain context here on lobste.rs, which is stealing their time.

I believe both touch interesting subjects, but this form is not discussion-ready. A discussion of leadership style cannot be done based on a single email, a discussion of a leadership statement can only be done with appropriate introduction of the context around the mail.

If you feel like the need to point out a discussion or a pattern on a mailing list, please write an appropriately-researched blog post on that (or find one that does) and then submit it here.

A good example were the recent discussions about Theo’s statements on the openbsd list, which had both Rust people straggling to correct misinformation and OpenBSD people providing context that we and others missed. A structured writeup of the issues that Theo pointed out (and where he’s wrong) would have been gold, the discussion is not.

I know these links will be passed around in other places in that fashion as well, but that doesn’t mean we have to do that, too.


  2. 40

    Almost anyone just reads the mail, but not the mail it was a reply to or the discussion that comes out of it.

    What makes you think this is true? I always read the surrounding thread.

    I have found many of these links very interesting and feel they are legitimate posts for Lobsters.

    1. 6

      General experience with the discussions that burst out below. Exceptions make the rule, as they say.

      Also, in the linked example, a lot of what Theo was referring to was in topics a couple of days away, so establishing context isn’t always “click previous, click followup”.

      1. 8

        Exceptions make the rule, as they say.

        No. “Exceptio probat regulam in casibus non exceptis” means that if there is an explicit rule regarding exceptions then there exists an implicit rule from which those exceptions are derived.

        1. 2

          I saw the pattern you were talking about. It seemed part of a larger one being really eager to speak but lazy about reading. I second your proposal of no mailing list threads unless it’s an independent write-up with context included so people see big picture at a glance.

          1. 2

            I agree and disagree. I agree that if we’re going to ban deep linking into threads then sure a write up would be ok.

            But, I would rather posit a second option, which amounts to when linking to a thread we link in a comment to the specific post that is at issue with the link to the top of the thread as the article link. Thus trying to encourage everyone to read the context that surrounds the post and allowing everyone to come to their own conclusions as to the post in its original context.

            I say this as while I’d love independent write-ups, noticing how every write up or news post can skew things ever so slightly I would rather encourage everyone to read through the original source material.

            I would propose the same rule for forum posts, I wouldn’t expect write ups in all cases and think that requirement would just discourage discussion over the actual issue at hand, if any.

            1. 2

              “I say this as while I’d love independent write-ups, noticing how every write up or news post can skew things ever so slightly I would rather encourage everyone to read through the original source material.”

              This is a good point. The Dalmore memo discussions were a good example of that.

        2. 1

          Personally, I try to get the context but I’ll admit that I can get a bit lazy with it. And my reflex is also a bit to do meta-discussions on communication.

          Unlike something like a Github issue, a bit more effort is required to get the full context. Not a huge amount, of course, but enough. Not to mention double-quoting and whatnot confusing people.

          Perhaps linking to the first e-mail in the chain will force people to read through stuff? Not sure how doable that is

        3. 30

          I disagree. I find the posts useful since I’m not a regular follower of LKML/OpenBSD’s dev list or other interesting mailing lists grappling with these sorts of issues.

          1. 2

            I have not said that we should not run them as submissions at all, but that I have find the current submission style problematic.

          2. 24

            Skade, thanks for working to keep discussions high quality.

            Without direct evidence (except for your own post), I believe that Crustaceans are aware that mailing lists are deeply contextual.

            That being said, I endorse your suggestion to avoid single link posts on these kinds of topics and instead link to analysis. I’d add only that the lobste.rs platform is perfectly capable of multi-link, blog-like posts. (Ala metafilter.) Simply make a text post (as yours is) and put all the links in the text.

            1. 22

              I’d been thinking of writing up a standard reference for the discussion we repeat a few times a year when someone posts a link to Linus being angry on LKML, summarizing the topics, points, and counterpoints that are reiterated almost verbatim every time. Maybe a solid summary would help people realize there’s almost never anything new to be said and that it’s not worth their time or emotion. When one’s reply has to start “I think there could be a parenthetical to the third sentence of Standard Argument 3.4.2…” it’s an opportunity to reflect on one’s priorities.

              Declaring that deep links into mailing list discussions/flamewars are considered off-topic might be a lot easier, but it’s a heavy-handed approach. Even after 25 years of seeing Linus/Theo emails rehashed I’m not quite jaded enough to say there will never be value in those discussions. But I could be convinced by argument or another half-dozen threads.

              (If anyone reading this wants to write the Standard Linus/Theo Controversial Email Epitomy/Abridgment, I certainly don’t feel proprietary about the idea.)

              1. 4

                I would love that and would like to state that I never asked for banning. I asked for a more considerate, less lazy approach.

                1. 1

                  Ah, thanks. I got the impression that you were proposing a new rule rather than suggesting a change to norms. I’m sorry for misrepresenting your aim.

                2. 2

                  Go for it. I think this is a great approach.

                3. 15

                  I’ll add my voice to those which disagree with you. I think that a single message in a mailing list can be an excellent place to dive into an issue — far better than an article or blog post, because the entire mailing list history is right there (unlike in an article or blog post, where the author will likely only link to other instances which support his points).

                  I do think that discussions about tone & style are generally off-topic to technical discussions. Someone who says, ‘water is composed of hydrogen & oxygen, jerk!’ is correct regardless of his rudeness, while someone who says, ‘pardon me, but in my opinion snails grow up to be geese’ is incorrect regardless of his politeness.

                  1. 10

                    I agree with you et. al. too. I think there has been meaningful discussion around such posts. People often quote before/after e-mails in the thread; talk about stuff that came up years ago, etc. I mean there might not be enough context in one e-mail, but it can be a good starting point.

                    When Meltdown came out, a lot of people were asking where the BSD community stood. For a while, the only official statement for OpenBSD came from their mailing list. I think for DragonFly, it was their changelog.

                  2. 15

                    Please do, so I get the chance to learn something new. I’d rather be annoyed by off-topic discussions than not be aware of the more serious issues discussed.

                    1. 10

                      Agreed. I’ve got enough critical thinking skills to dig around in the context of a mail. Hell, most folks get looped into threads like this at work on random occasion and that’s exactly what you do. You follow the reply/response chain up to understand anything you feel like you missed.

                      1. 3

                        Not relevant to the discussion here, but people randomly looping in other people into email threads without providing at least a summary of the situation as they see it is pretty damn rude.

                        It depends on the situation of course, but I usually note in my communication why this person was added: “Adding NN to the conversation to provide input, in light of their role as customer contact” or similar.

                        Personally, I filter all emails with my contact in the CC list into a specific folder to avoid noise in my main email inbox.

                    2. 7

                      I’m a simple person. When someone submits a link to the creator of the linux kernel discussing a critical CPU issue that has a huge effect in my planning for 2018 and the very foundations of “cloud” computing, I click on it.

                      1. 7

                        I agree – if you want to talk about one of these threads, it would be better to write up a blog post explaining the context.

                        1. 3

                          I completely disagree. I don’t want a filtered, abridged version of reality. I don’t need things explained to me when when we live in an age with such incredible research tools. At most, post a top level comment on the submission with pointers to especially relevant earlier discussions. If anything, I favour a rule that prohibits summary blog posts in favour of posting original sources.

                          1. 3

                            Providing contextual information doesn’t filter, only prepare. I have not asked for a rule.

                            Also, if you want the “filtered, abridged version of reality”, you shouldn’t read this page, which in its very setup is a huge filter.

                        2. 6

                          I agree if you mean repiles somewhere down a thread. Those aren’t usually a good place to start reading the thread, and almost inevitably when linked elsewhere end up being kind of Twitter-style out-of-context quoting of an argument, not really an “article”, which is what I’d rather see at lobste.rs.

                          I do think linking individual mailing list messages can be appropriate when they’re written as more stand-alone messages. Sometimes people write messages that could just as well have been blog posts, but for various reasons posted them to a (public) mailing list instead. So those make sense to submit just like the equivalent blog post would be ok. (Often, though not always, these are messages starting a new thread.)

                          1. 1

                            I agree if you mean repiles somewhere down a thread.

                            Yes, that’s what I mean by deep-linking.

                          2. 4

                            I believe both touch interesting subjects, but this form is not discussion-ready. A discussion of leadership style cannot be done based on a single email, a discussion of a leadership statement can only be done with appropriate introduction of the context around the mail.

                            Why should we treat our fellow users like babies? If you don’t want to read the context, that is on you–let others hold the discussion.

                            More importantly, instead of just saying “please don’t”, which comes off a lot like “Deep-linking Considered Harmful” (meriting https://meyerweb.com/eric/comment/chech.html), why not structure your post around suggesting alternatives? You mention writing appropriately researched blog posts. It seems like text posts like yours simply linking to & quoting from replies giving the necessary context would work well too.

                            1. 1

                              I find the accusation of “treating fellow users like babies” quite baseless. There’s miles of land between critisising a pattern (which some people don’t agree with, which I in turn find fair) and treating people like babies.

                            2. 4

                              I’m not sure that this happens often enough to do anything. It could be an over-optimization. Sort of: trying to keep bad things from happening we might keep some good things from happening as well.

                              1. 3

                                The reason I raised my voice is just because I found the pattern problematic, I haven’t asked for moderator action or banning as some people seem to suggest here.

                                What I ask people to do is maybe invest a little time on submission of such things, instead of going for the shock value.

                              2. 3

                                I’m fine with those links, but it should be the responsibility of the submitter to supply the necessary context.

                                1. 3

                                  Maybe something about this could go in the “Story submission guidelines”?

                                  1. 2
                                    1. 2

                                      I disagree on every point except that a well-researched post is of higher value than a chaotic discussion thread. The link does usually cause me to go read the entire thread, the link is an acceptable starting point for discussion, I am capable of ignoring the meta-discussion of their communication style (though I sometimes enjoy reading that), and, finally, I don’t care what other sites do.

                                      1. 2

                                        Perhaps it would be more useful to ask people not to derail technical posts with meta-discussion about communication style and behavior. It’s a regular occurrence in this community, and not restricted to mailing list threads.

                                        1. 8

                                          Many of these submitted mailing list threads aren’t really submitted for their technical content in the first place, though— they’re explicitly submitted because they were a flamewar and people like to gawk at flamewars, so that’s kind of on-topic to discuss imo. The only particularly interesting thing about the recent Torvalds submission, for example, is the flaming. Presumably that’s why the submitter chose to include an all-caps quote, “COMPLETE AND UTTER GARBAGE” in the submission title, rather than highlighting any technical content. I’m going to go out on a limb and predict that if it had a technical title instead of a flamewar title, it wouldn’t have gotten the attention here that it did. (The little technical content the linked post has turns out further down the thread to not even be correct.)

                                          At the very least, when people are linking gawk-at-the-flamewar type mailing list posts, can I suggest tagging them with the rant tag?

                                          1. 3

                                            The only particularly interesting thing about the recent Torvalds submission, for example, is the flaming.

                                            He accuses Intel of planning to not to fix the specter bug, as in they want to provide a workaround off by default since it would impact their performance metrics and shifting the responsibility to OS vendors. That’s far more interesting than flaming and worth the submission in itself.

                                            So the IBRS garbage implies that Intel is not planning on doing the right thing for the indirect branch speculation.

                                            It’s not “weird” at all. It’s very much part of the whole “this is complete garbage” issue.

                                            The whole IBRS_ALL feature to me very clearly says “Intel is not serious about this, we’ll have a ugly hack that will be so expensive that we don’t want to enable it by default, because that would look bad in benchmarks”.

                                            So instead they try to push the garbage down to us. And they are doing it entirely wrong, even from a technical standpoint.

                                            source: http://lkml.iu.edu/hypermail/linux/kernel/1801.2/04628.html

                                            1. 5

                                              http://lkml.iu.edu/hypermail/linux/kernel/1801.2/04630.html http://lkml.iu.edu/hypermail/linux/kernel/1801.2/04637.html

                                              The next 2 emails show that Linus has misread the patch.

                                              You’re looking at IBRS usage, not IBPB. They are different things.

                                              Yes, the one you’re looking at really is trying to protect the kernel, and you’re right that it’s largely redundant with retpoline. (Assuming we can live with the implications on Skylake, as I said.)

                                              (I pointed that out in the lobste.rs thread, and that’s kind of the thing I was annoyed about)

                                              1. 3

                                                FWIW, if you look at the second email you linked…

                                                Ehh. Odd intel naming detail.
                                                If you look at this series, it very much does that kernel entry/exit stuff. It was patch 10/10, iirc. In fact, the patch I was replying to was explicitly setting that garbage up.
                                                And I really don’t want to see these garbage patches just mindlessly sent around.

                                                Linus seems to be claiming that he didn’t misread the patch.

                                        2. 1

                                          Perhaps the gold writeup is the friends we made along the way the details that come out in the comments?

                                          Maybe instead of banning the mailing list posts, ban the meta-discussion about Linus (let’s face it, it’s usually Linus) being a jerk.

                                          1. 2

                                            I didn’t ask for banning of either, I explicitely welcomed both discussions. I criticised the lazy approach.