Please create an account to participate in the Slashdot moderation system

 



Forgot your password?
typodupeerror
×
Python Social Networks Twitter

Python Creator Guido van Rossum Blames His Resignation Partly On Social Media (www.tfir.io) 137

"Swapnil Bhartiya, the founder of TFIR, sat down with Guido van Rossum, the creator of Python, to talk about the origin of the language and why he stepped down from the leadership of the very project he created," writes sfcrazy.

In the interview, van Rossum emphasizes that he still remains one of the core developers, and provides this update: "We're going to set up a new form of governance. We haven't decided yet what that will be. There is actually an interesting time ahead where we currently have about five of six different proposals for new governance systems, and in November there's going to be a vote among the core developers about that. And then there will be another vote that will actually determine specifically who is going to form the leadership. So we're starting out by choosing a constitution, and then using the rules set out in the selected constitution, we're going to vote for a leadership..."
He talks more about his resignation when asked if there's ever been an after-the-fact debate about decisions he's made: "Well, that certainly happens too. What led to my resignation was a form of that, where on social media -- and I've got a feeling that social media are sort of getting out of hand... But for me personally, social media definitely sort of caused additional stress. And I did not enjoy it when core developers were sort of sending tweets where they were questioning my authority or the wisdom of my decisions, rather than saying it to my face and having an honest debate about things...

"It might just have to do with the fact that I've had this role for 28 years... And all that time, I've been sort of the final decider, the final arbiter. I'm getting older, I'm not always available... I just want to spend less time feeling stressed about what is the community -- I have this attitude where everything that was being said on some of the mailing lists, python-ideas, python-dev, touched me. I felt involved in everything, because ultimately every idea would end on my desk for deciding. And I just thought that that should be a responsibiity that should either be shared or transferred... Given that I've been on the project for such a long time, and some of the currently active core developers are good personal friends that I've known for 20 years or more, I am completely confident that the more experienced core developers that we currently have, plus the newer core developers that we have, together will be able to weather any kind of storm that might come Python's way. Yes I resigned from the title suddenly, but there were a lot of responsibilities that I had already completely delegated. I mean, I barely touch the code base, I barely reviewed submissions.

At one point van Rossum compares the future of Python to that of a grown-up child, in that "You're supposed to raise your child for independence..."

So what's he doing now? " I was and still am a principal engineer at Dropbox, which is actually where I spend most of my time."
This discussion has been archived. No new comments can be posted.

Python Creator Guido van Rossum Blames His Resignation Partly On Social Media

Comments Filter:
  • I understand (Score:5, Interesting)

    by Xenna ( 37238 ) on Sunday May 05, 2019 @03:58AM (#58540702)

    He doesn't want to be a slave to his own software...

  • Yeah, I know I'm a bit off topic this time, but I've been meaning to ask and haven't noticed many Python-related stories lately. I've taken a number of Python classes, but wasn't fully satisfied. Rather than a negative approach by saying where they didn't satisfy, I'd rather describe the kind of class that (I think) I'd most enjoy:

    Maintenance programming in Python

    Anyone know of such a course to recommend? I haven't been able to find it, but I realized that most of the languages that I learned later on with

    • Could you define, more clearly what you mean by "maintenance programming"? What exactly are you trying to do? There are a huge number of fields out there only a small portion of which post on Slashdot and Python is doing great there as the next "Glue" language (like Perl is/was).

      • maintenance programming is what most programmers do - you say there are a huge number of fields out there - in fact maintaining old code is the largest field by far
      • by shanen ( 462549 )

        The real-world examples might emit more smoke than light, though I enjoyed myself at the time. As I visualize it in a class, there wouldn't be so much focus language concepts and creating noddies, and mostly exercises with instructions like "This program does foo. Change the foo function in this way." Or maybe "Where should you add the feature bar to this program?" Most nasty would be cases of "The program is supposed to do this, but does that instead. Fix it."

        There should also be exercises in recognizing g

        • Oh, and I forgot one thing,

          "Move similar things together." Move all your message sending code to one place, move all your file parsers to one place, move all your APIs about logging in to one place, move all your APIs about advertising to one place.
          • by shanen ( 462549 )

            I'm a little confused by your entry into the discussion. I think you meant "You [meaning me] forgot one thing" unless you are referring to something else you wrote in a different branch.

            However, if I understand your suggestion, I think I agree. However I think it is part of keeping things at the right level of abstraction. It becomes harder to group similar things when extra levels are involved. If you are focusing on one level at a time, then it should be clear which parts are similar and can be moved toge

      • Could you define, more clearly what you mean by "maintenance programming"?

        Tidying up the mess the "rockstar" developers leave behind.

    • I'd rather describe the kind of class that (I think) I'd most enjoy

      Well, there it is. Put together a skeleton syllabus and curriculum, get some feedback and input, and put it through a few passes of teaching it yourself. If you don't find an existing class of this sort and can find students for it, there's likely a solid need for it that you can start filling.

      • by shanen ( 462549 )

        Your reply mostly makes me think you've never actually worked as a teacher. I think you're underestimating the effort involved, perhaps radically. Actually, going beyond that, your reply even makes me wonder if you've ever had a really good teacher. I think you should clarify the foundation of your "advice" before I'll consider it more seriously.

        So I guess that means I am obligated (by my request of you) to try to clarify my own foundation, eh? Officially I'm retired now and I don't feel any urge or have an

    • by orlanz ( 882574 ) on Sunday May 05, 2019 @06:10AM (#58540902)

      Just a small thing back on point. BDFL is the "key visionary" of Python. He was the final NO on many ideas over the life of Python. I would say his greatest contribution to the language was a firm line on feature creep (the worst thing about C++ and PHP).

      It wasn't until Python 3 that he really started compromising with the community (ie: map, filter, etc). It was also at this point that Python started getting more features. Now a days, there are quite a bit more features and many I don't think are worth the additional know-how that must be kept in mental cache to review code.

      Across all programming languages, if you had to pick the most influential and hands on contributor to their language; HE would be leagues ahead of everyone else. Even Linus hasn't provided the level of control and direction on the kernel that BDFL has had on Python.

      • by jma05 ( 897351 )

        > if you had to pick the most influential and hands on contributor to their language

        I would give that title to Larry Wall. Guido and Larry treated their languages differently. Perl also seemed to center more around Larry than did Python around Guido. But that may just be my memory.

        > Now a days, there are quite a bit more features and many I don't think are worth the additional know-how

        Python is still rather slow on feature uptake though. Even the historically conservative Java/C++ seem to be adding mo

      • there are quite a bit more features and many I don't think are worth the additional know-how that must be kept in mental cache to review code.

        This is something that should be remembered in language design.

        • Why not change the subject as needed? Especially since this branch was triggered by a footnote in my original query.

          What this branch of the discussion most reminded me of was the importance of saying "No" to nonessential features. It really is important to keep a clear focus. I think this was attributed to Steven Jobs and described as a key to his managerial successes. Much as I dislike some things about Apple, the company produced some great products. (But nothing as good as the modular smart-living-chair

          • As should be apparent, I am certainly not an expert at Python (though PERL is one of my favorites and I learned it by the maintenance-programming approach I asked about), but I feel that much of Python's success has been the way the modules are separated out of the core language.

            How is that different than any other language? For example, Perl CPAN.

            • by shanen ( 462549 )

              I think your question (in the body, but maybe the Subject:, too) was rhetorical, but if not, then I would say that it's an important aspect of why some languages are more successful than others. It doesn't even matter which criteria you focus on or how you weight the criteria, the various languages don't come out equal.

              I didn't actually say anything about relative virtues of Python over other languages, but are you actually asking why I'm interested in studying Python now? I guess the main reason is that it

    • Re: (Score:1, Insightful)

      Anyone can learn Python in a week with a book. Why would you need classes in it?

      • Re: (Score:2, Insightful)

        by Anonymous Coward

        Really he wants a class on best practices for maintaining code / systems, using python.

    • Maintenance programming in Python

      A lot of startups 'maintain' their Python code by rewriting it from scratch in Java after they are acquired or go public. The big problem is that Python has weak typing, and it makes things nearly impossible to work with as they get larger. It is very difficult to figure out every place that a variable gets used, for example.

      That said, when I come around to a codebase like this and need to clean it up, here is what I do:

      1) Get the build/deploy system working and automated. In Python this usually means g

      • Thoughtful and interesting reply (though I never get a mod point to give). Relevant regarding advanced elements of the approach, though not a pointer to any existing class. More like a reflection of how it goes in the real world.

        My main reaction is actually a triggered memory from your fourth point. It reminded me of an internal email system I wrote many years ago. The boss and owner of the company was a smart guy, with a law degree and an MBA, but there was one time when we had a LONG (about two hours) dis

  • by Anonymous Coward on Sunday May 05, 2019 @04:06AM (#58540716)

    All he needed to do is kiss off social media and communicate only over official, perhaps difficult to get onto, mailing lists. Social media is made for trolling and massaging mediocre persons' precious egos.

    {o.o}

    • by Kjella ( 173770 ) on Sunday May 05, 2019 @08:17AM (#58541144) Homepage

      All he needed to do is kiss off social media and communicate only over official, perhaps difficult to get onto, mailing lists. Social media is made for trolling and massaging mediocre persons' precious egos.

      Meh, in this case I think social media is simply exposing what most people already know. Despite how much you're trying to have a constructive discussion, find consensus and reach compromises most people don't really change position very often. It's more like a tug-of-war where the other side is stupid and wrong but win because their numbers/arguments are given more weight. And the one giving the weighting becomes the lightning rod for their anger / frustration / disappointment.

      We all know the appeal of "I told you so" when it turns out the other side was wrong. When we're wrong though... I'm sure that in about 99% of the cases I have a perfectly reasonable explanation - to myself at least - as to why my position was still the right one given the available facts and my understanding of the system. Like at best you admit that some of the inputs were faulty or incomplete, so garbage in led to garbage out or they got lucky because their risky/speculative solution paid off. The last thing you do is admit you were plain wrong.

      That and it's always a lot easier to be in the peanut gallery than down in the trenches. It's easy to trash talk something when you're not taking any part in it, it doesn't mean users shouldn't be allowed to have an opinion but a lot of the time it does feel like "well you can stand there and nag or you could roll up your sleeves and help" and they'd rather breathe down your neck until you do it their way. Most of the time the people who are content are quiet, so it seems like nothing but an endless stream of complaints. Some of it is whining you simply have to ignore, if you take everything personally you'll never last.

    • I think you're missing the point of his statement. It wasn't social media that caused the issue, it was social media that exacerbated it. The issue was core devs debating things in a third party fashion rather than discussing it with him directly. And his take on it that he shouldn't be the end decider at this point of his life/career is a good call - he needed to be the final decision in the early days to see his vision survive. Now that it has a huge base and millions of devs attached to it, he's doin
  • Like any computer expert - I presume they all see what I see - I've always stayed clear of contemporary social media. I do have accounts, but they are a vector *to* my sites and presences that I control. They are for the masses - not neccesarly in a bad way - and thus lower the intellectual bar considerably. When Twitter came out, I got an account and looked at it for roughly 3 minutes and then concluded that it was of no real use to me. It was the unified messaging approache finally done right, but limited

  • by nospam007 ( 722110 ) * on Sunday May 05, 2019 @06:13AM (#58540912)

    With that name, he should have done Rossum's Universal Robots instead of just software.

  • by packrat0x ( 798359 ) on Sunday May 05, 2019 @07:57AM (#58541086)

    Python1 was written in 1991. Let us look at this in context.

    Scripting on Apple 2s: batch files
    Scripting on IBM PCs: batch files
    Scripting on Mac's: nothing, unless you paid $500 (in 1991) for the developer's kit
    Scripting on Unix: tcsh, ksh88, bash1, zsh

    Python scratched an important itch. It allowed you to write a script that could be run on various machines.

    In 2019, Bash4 (and GNU utilities) can be found on Unix, Mac, and now WinNT. The original use case is gone.

    • by Anonymous Coward on Sunday May 05, 2019 @08:08AM (#58541112)

      You forgot PERL on Unix. Which is what you used if a shell script didn't cut it.

    • by Bigbutt ( 65939 )

      I had to deal with REXX scripts occasionally and of course perl. But generally if a batch file or shell script wasn’t enough, I wrote a C program.

      [John]

    • by Anonymous Coward

      You can't compare python to those with a straight face. Plus you've left out Java.

      It's simple. Python is yet another project to adopt the CoC[k] way of life and wonder why people aren't all happy about it

      https://github.com/python/cpython/commit/2cdeced797251f218e416c04b31a00ebbac392d3
      https://github.com/python/cpython/pull/9101

      etc.

      The CoCK added in 2017, strangely in what would normally be a hidden directory, that is ".git". Interesting thing is the URL (because adding the actual text would be incriminati

      • by Anonymous Coward

        LOL, looking at those commits and discussions, no wonder he quit. master/slave is just standard computer lingo for decades, suddenly you have to keep changing it all to some weird terms that are harder to understand. Just so some SJW can stroke themselves while some poor sod has to try to make it all work. pfft

        If this is some indication of the general community, I have to salute GvR for spending close to 3 decades of his life on that. Mostly for free I guess. And then you get everyone whining when he is not

    • by Anonymous Coward

      Scripting on Apple 2s: batch files
      Scripting on IBM PCs: batch files
      Scripting on Mac's: nothing, unless you paid $500 (in 1991) for the developer's kit
      Scripting on Unix: tcsh, ksh88, bash1, zsh

      Not to detract from your point (which I mostly agree with), but wow dat apple trolling.

      Apple 2's didn't have any form of batch or scripting for the two Apple OSes (DOS and ProDOS)
      There was Basic that could chain load things, but I'm not sure that qualifies as scripting at all.

      Macs had AppleScript in various forms since System 7, and included up to this day.
      Based on the HyperTalk/HyperCard language syntax but actually made useful in that it could do things with programs and the OS.

      Also no additional "dev kit

    • by jma05 ( 897351 )

      I remember differently.

      I don't recall that Python was used much until 1.5, 1.6.
      Perl and Tcl were. But I don't recall there being much of cross-platform scripting back then at all.
      MS was later pushing VBScript and JScript in 1996 (around the time of Python 1.6 and Python was not big on Windows then).
      PSF came about only in 2001.

      Python was initially trying to be a "teaching language", not a cross-platform tool to replace Perl. But that was what it eventually became, as Linux distros increasingly adopted it as

      • I don't recall that Python was used much until 1.5, 1.6.
        Perl and Tcl were. But I don't recall there being much of cross-platform scripting back then at all.

        I remember because I was using scripts in cross-platform fashion back then. And you could in particular get TCL/Tk for Windows, and as I recall, you could get PerlTk as well. TCL was by far the most popular cross-platform scripting language, though. People built a lot of utility GUIs in it. Macintosh had something like 3% of the market at the time, so while I would be surprised if they did not have a TCL/Tk implementation, I am not familiar with it.

        Python was initially trying to be a "teaching language", not a cross-platform tool to replace Perl. But that was what it eventually became, as Linux distros increasingly adopted it as the main scripting language since it was more readable. The recent adoption spike came from data science and analytics - for the same reason, it was very easy to learn and very easy to read other people's code.

        It was meant to be a teaching language, so they chose to ma

    • by Anonymous Coward

      "Let us look at this in context."

      The IBM OS/360 environment had primitive scripting with JCL procedures. Eventually UNIX was developed on smaller DEC computers. There were also timesharing systems using TTY terminals. Then came the microcomputer revolution with flavors of
      BASIC. Qbasic was an improvement on BASIC, because it was more structured and cleaner. Eventually Perl was born as an improvement on BASIC, and Python was an improvement on Perl. My impression at the time, was that Python was born in

    • Python1 was written in 1991. Let us look at this in context.

      Scripting on Apple 2s: batch files Scripting on IBM PCs: batch files Scripting on Mac's: nothing, unless you paid $500 (in 1991) for the developer's kit Scripting on Unix: tcsh, ksh88, bash1, zsh

      Python scratched an important itch. It allowed you to write a script that could be run on various machines.

      In 2019, Bash4 (and GNU utilities) can be found on Unix, Mac, and now WinNT. The original use case is gone.

      As Python's Original USECASE has changed, so has its USE. As the problem grew apart and was resolved at the OS level, where it should have always been, Python now covers many more use cases than replacing script line bash from the 90s.

      • From TFS: At one point van Rossum compares the future of Python to that of a grown-up child, in that "You're supposed to raise your child for independence..."

        There was a need, and he solved that need. Now that need no longer exists, and Python is heading off in a different direction. This is the core reason for his resignation.

  • by JustAnotherOldGuy ( 4145623 ) on Sunday May 05, 2019 @11:10AM (#58541648) Journal

    " What led to my resignation was a form of that, where on social media -- and I've got a feeling that social media are sort of getting out of hand... "

    Oh really....and you're just getting clued into this now?

    Social media (a misnomer if I ever heard one) has become so toxic that a lot of people are simply switching off and abandoning it.

    The trolls have overrun twitter, Facebook, Instagram, and nearly every comment section on every news site. It's disheartening to see a thoughtful comment buried in a sea of hateful shit, racist comments and trollish griefer crap.

    And when you realize that an surprisingly large portion of the population can actually have their attitudes and opinions shaped by these relentless fucktard trolls, then you start to see how really corrosive it has become. They actually form their worldview from the misinformation they read and that is not repeat NOT a good thing.

    Slashdot is one of the few places I bother to comment anymore, and even it's being overrun by the APK-tards and the "hate duh jewz" horsecrap.

    After a while you just want to leave the cesspool and breath freely. I can see a time when I'll close the browser and never come back.

    • by BankRobberMBA ( 4918083 ) on Sunday May 05, 2019 @01:07PM (#58542128)

      When I got out of prison in 2016 I looked into social media and, honestly, I could not see any use for most of it. I rejoiced, though, at being able to read Slashdot again.

      I started commenting, then I started getting mod points. I started also meta-moderating, and I started to get lots of mod points. I try to use them to promote good comments, even if I don't agree with the poster.

      I browse at -1 ("to watch for abuse") even when I'm not moderating. Just to see the unfiltered reality of the site.

      Now I get 15 mods points at a time every two or three days. But what I see on the site is a skyrocketing rate of posters coming in as AC. Even just over the last two years. I often find GOOD comments submitted AC. And the prevalence of twerps and dipshits is disheartening.

      So, do you think it's inevitable that Slashdot is headed for the trash heap?

      • I routinely browse at -1 also, and it's sad to see the constant flow of comments from trolls who just want to stir shit up. They're griefers writ large. I also moderate/meta-moderate and the amount of stuff I mark "binspam" is overwhelming, but at least it never reaches the front page.

        "So, do you think it's inevitable that Slashdot is headed for the trash heap?"

        Sadly, I suspect it's inevitable. The ONLY thing keeping the site viable is the post-moderating and that often isn't enough. The owners should be ha

  • Comment removed based on user account deletion

Life is a healthy respect for mother nature laced with greed.

Working...