Emacs maintainer steps down | Hacker News Reader

Emacs maintainer steps down

lists.gnu.org

241 pointszeveb71 comments

Open articleView on HN

Note: this happened almost a month ago now.

See these threads for the discussion on the new head maintainer:

https://lists.gnu.org/archive/html/emacs-devel/2015-09/msg01...

https://lists.gnu.org/archive/html/emacs-devel/2015-10/threa...

I recognize that the role of head maintainer is in part a political one, but yow.

Snippet that I believe exemplifies what you mean (since it's a bit hidden):

  It would be wrong and harmful to give MacOS an "equal footing". 
  Our goal is to replace nonfree systems (and nonfree software in
  general), not to enhance them.

https://lists.gnu.org/archive/html/emacs-devel/2015-10/msg00...

I browsed through the email thread briefly and it looks like a lot of "political ideology vs technical purism" drama... Seems high stress.

Looks like after some hearty back and forth it settled down: https://lists.gnu.org/archive/html/emacs-devel/2015-10/msg00...

Although it seems they ended up on an "agree to disagree" finale: https://lists.gnu.org/archive/html/emacs-devel/2015-10/msg00...

> it looks like a lot of "political ideology vs technical purism" drama…

That's pretty much the history of the FSF and the GNU projects considering they were created for political/ideological reasons (the Unix in "GNU's Not Unix" is nonfree/encumbered AT&T UNIX, and the ideological underpinnings are what allowed Linux to overtake BSD during "the troubles" of the early 90s). The pulls ideological and technical pulls have oft conflicted.

the ideological underpinnings are what allowed Linux to overtake BSD during "the troubles" of the early 90s

As I understood it, also to a degree witnessing it as it happened, this lawsuit spiked BSD at the critical period: https://en.wikipedia.org/wiki/UNIX_System_Laboratories,_Inc.....

It certainly took Linux a long time to get in the ballpark of quality that BSD started out with at the gate, including the standard TCP/IP stack most everyone used, including Microsoft (initially, at least).

Oops, HN cuts the canonical URL's last character, a period; try this: https://en.wikipedia.org/wiki/USL_v._BSDI

...and attitudes like that just result in many people not using any free software at all, when they could instead be running free programs on a non-free platform.

It's important to read the full context. All that snippet means is that the GNU Emacs maintainer should be fine with adding new features to Emacs that are not portable to non-free platforms like Windows and Mac OS X. That is a perfectly reasonable requirement: Non-free platforms shouldn't hold back free platforms, because portability is not the ultimate goal, creating the best free software possible is the ultimate goal.

That's a very positive spin on the message. The way I read it is that developers and maintainers are encouraged to make Emacs work better on free than non-free platforms.

This is, sadly, self-defeating. If Emacs were widely recognized as the best editor (or whatever of its myriad functions you use it for), "works best on GNU/Linux" could advance the cause of free software. As it stands, though, "only works well on GNU/Linux" instead just means that fewer people will be using it.

Or https://lwn.net/Articles/629259/ , referencing the long-standing issues caused by forcing the use of GCC rather than LLVM (and trying to satisfy someone who doesn't understand why people need to process source code and ASTs for anything fancier than indentation and syntax highlighting).

Holy shit. This is incredible. I love Emacs but it's so sad that the stewardship of this amazing piece of software is in the hands of someone who would rather keep the feature-set substandard because of vague fears that someone might steal the AST's from his precious GCC.

I wonder what would be the problem with someone using libclang to generate the same AST information in a much more convenient fashion. Will it break the GPL in some way?

People can and do use libclang for AST info in the same manner that you're talking about: https://github.com/Sarcasm/irony-mode

Almost didn't click on that link because it looked like a joke link (Sarcasm and irony!!). Glad I did and found something new :-)

Stallman's had a long history of hating on Apple, up to and including organizing a GNU boycott of them back in the 90s.

The more I learn about Stallman, the more I admire him.

That was a very interesting and long discussion that you linked to in the second link, in this day and age it was quite refreshing to read an adult and polite debate where disagreements were discussed without resorting to name-calling and insults.

Emacs being such an important part of the FSF it makes a lot of sense that its maintainer should be conscious of this fact and prioritize it accordingly, even if sometimes that means not making Emacs as good as it could be (see the clang/gcc discussion in the linked thread).

Totally agreed. Rare thing nowadays...

Heh, I like how Richard Stallman begins his messages: https://lists.gnu.org/archive/html/emacs-devel/2015-10/msg00...

I dunno, I think that's some bullshit doublespeak right there. He conflates emacs and the gnu project, while claiming to want the best possible emacs.

   The GNU Emacs maintainer's responsibility is to take charge of Emacs
   on behalf of the GNU Project, and produce the best possible GNU Emacs
   -- which means, the one that advances our aim the most.
   [...]
   It includes making sure dynamic loading resists GPL violation. [1]

Non bs would be to say, eg, "the task of our maintainer is to build the best emacs possible subject to the political goals of the gnu project."

Note that I'm not sharing my opinion of whether this is good or bad; I just dislike bs.

[1] https://lists.gnu.org/archive/html/emacs-devel/2015-10/msg00...

edit: I take it back; you're talking about the message to the NSA? I skipped over everything in brackets while reading...

I don't think you're justified in reading in your own definition of "best", when RMS defines it himself right there in your quote. If you're prepared to agree that the best possible chocolate mousse is probably "good" in ways that are incomparable to the best possible steam locomotive, then all you have in your comment here is misleading rhetoric.

Kind of funny to then sign off with

> I'm not sharing my opinion of whether this is good or bad; I just dislike bs

the thing is, words have meanings. Writing

   we want to create the best X (*)
   [...]
   * NB: not the best

   we want the best X (redefinition of best)

is bs

Finding mangling language to deceive dumb is orthogonal to my opinion of the political goals of the gnu foundation.

How is it a "redefinition" of best? What was the original definition? Unless you add some qualifier, "best" for a software project doesn't have any kind of objective definition. For a company it might be "most profitable", for a newbie it might be "easiest to use", etc. There are a ton of criteria that can go into it. RMS helpfully explains in broad terms what in his opinion would constitute the "best Emacs".

In the context of avoiding widely desired features in order to preserve the political goals of the gnu foundation, it's straightforward to understand.

"Best" means "most optimized for some utility function". It always requires the utility function to be defined, and there's always an infinite array of conflicting utility functions that could be used -- that's the nature of "best".

Stallman is, quite properly, clearly addressing the relevant utility function. This may differ from your preferred utility function, but that doesn't make RMS's definition a "redefinition of best".

He's being straightforward and honest, which you label "bullshit doublespeak".

I have seen many demos and such where Steve Jobs and the like talk about how they want to make the best product possible for users and so forth. According to you, they all have saying "the task of our company is to build the best product possible, subject to the overriding concern - profits for the billionaires who own a majority of our stock".

I have never seen this. They don't say this. That's the real bullshit doublespeak. Stallman is honest and straightforward about his goals, and you use that honesty to not only attack him, which is fair enough, but to use his honesty to say he is being dishonest. It is you who is being dishonest. Why aren't you attacking companies who don't say "Our #1 priority is making profits for the billionaires who own a majority of our stock, the quality of our products reflects this concern". That's the truth, that's the agenda, but you say nothing about this. Stallman is honest about his agenda, and you use his honesty not to attack him because you disagree, which is fair enough, but use it to call him dishonest, which is ridiculous.

Read it carefully: When he talks about Emacs, he says Emacs. Then he says: “produce the best possible GNU Emacs”. You might have just skipped over the GNU here, because that is also the name of the software, but it matters in this case. If your realize that, his message is extremely clear.

(And this also how I read his message the first time, before I read your post)

Stefan has always seemed fantastic. Thanks very much!

The thread makes hilarious reading for those who don't care so much about the FSF's ideals. E.g.

> The liberation effort of the soviets died from these kind of treatment.

My understanding from skimming those two links is: John Wiegley is considering/being considered for the head maintainer, but the decision hasn't been made yet. Is that correct?

Looks like he withdrew himself from consideration (2nd reply paragraph):

https://lists.gnu.org/archive/html/emacs-devel/2015-10/msg00...

Then he withdrew his withdrawal:

https://lists.gnu.org/archive/html/emacs-devel/2015-10/msg00...

Stefan's stewardship resulted in a much-improved Emacs. He did a very good job.

I'm really sorry to see Stefan go. The rate at which Emacs improved while he was maintainer contrasts quite strongly with the stagnation of some years before (6 years from Emacs 21 to Emacs 22).

Infrastructure changed a lot. major emacs packages are now upgraded everyday (I use melpa "rolling release"), so different than the emacs distribution every x years. Not to reduce Stefan role in emacs quality and liveliness. Seeing the praise he received says a lot.

BTW, there is also MELPA Stable https://stable.melpa.org

How does that work with long-running emacs processes?

I generally keep a daemonized emacs or two running the entire duration of my uptime. If package A depends on a certain behavior of package B (but doesn't eagerly load it), if I start emacs, load package A, wait some duration X, package B gets updated, then I load package B, will A break?

If so, is that a failure of A for not eagerly loading B, or a failure of the dependency system for not letting A declare that it wants a particular version of B, or something else?

Sorry, I have not really dived into package.el all that much and the documentation is somewhat fragmented since everybody customizes it the way they want.

I don't know much either, I had folder issues many times (package-v1 still loaded trying to open .emacs.d/elpa/package-v1/some-file even though package-v1 has been deleted and now only package-v2 is on disk.

I remember a few elisp level error after upgrades.

I've been using Emacs for 20 years, I realized. If you think about all the different things that come and go so quickly in this field, that's a pretty amazing run.

Thanks Stefan!

Same, come to think of it -- almost 20 to the day. In a world of modular software and months-old "unicorns," it's good to remember the monolithic, decades-old tortoises and bristlecones that remain and sustain.

Same here. About twenty years. Started programming with emacs and cvs on DEC Ultrix machines. Learned UNIX the hard way. Good old Dave sitting in the corner, hacking away on imake, and when asked about something answered "Try man something. Don't know what man is? Try man man".

Been on and off with emacs, but never really got into elisp. Only after having to learn Scheme/Racket for a Fluxus project, I read elisp with different eyes now. Really whish I got into Lisp earlier.

I feel the same. Not getting in to Lisp and Scheme decades ago has been my biggest regret, as far as technology goes.

22 for me. Damn I suddenly feel old.

You guys are kids. I used MINCE (Mince is Not Complete EMACS) in 1981.

You beat me. MicroEmacs on an Amiga....

I've been using Emacs for only four years. I would love to have a peek at you guys' .emacs files.

https://github.com/duncan-bayne/emacs-setup/wiki

Around that time I used gosling's emacs. I remember switching to gnu emacs a few years later and having to port all my .el files. Fun times.

> You guys are kids. I used MINCE (Mince is Not Complete EMACS) in 1981.

If we're going to get into I-can-top-that territory here <g>, I used some version of Emacs and Brian Reid's SCRIBE on a DEC-20 machine in 1980-82. (I was a law student; the AI guys at UT Austin's CS department let me have an account to experiment with word processing for the law review -- a grateful shout-out to Dr. Mabry Tyson, a grad student at the time, if by chance he reads this.) Then The Final Word [1] on a Compaq PC clone to do camera-ready copy for my first book. At my law firm I wrote an Emacs keyboard emulator for Word Perfect for MS-DOS (and posted it on CompuServe), then another one for Microsoft Word for Windows.

</obnoxious>

Now if I could only do more than pitiful coconut-headphone programming in Emacs Lisp ....

[1] https://goo.gl/d9MaKH (Google Books archive of 1983 review in PC Magazine)

You were a law student, then a lawyer, and found time for all that? Bravo!

Donate to the FSF! :)

(I mean you don't have to obviously, but I think it's a good remainder for all of us that Emacs wouldn't exist without it. I've just bought a shirt myself!)

For me, 18.59 was the last real solid release.

More than 10. Emacs is my most used software.

7 years there. No regrets (almost). Use it for everything except Java

I used a port of Emacs created by Dan Bricklan (MIT) of Software Arts (Cambridge, MA) for the Prime minicomputer in 1980, so 35 years. Have used it ever since as my daily programming editor.

I would encourage anyone who clicks the link to read through the thread. Pretty heartwarming to see the goodbyes, a definite mid-day boost :)

The tone of that thread definitely makes contributing to emacs development an appealing project.

I've made some trivial contributions (enough to require copyright papers, barely), and can confirm that it's a smart, positive community. There are the usual few kooks, and you absolutely must work within RMS's strict ideas on software freedom, but it's nowhere near LKML hostility and dysfunction. The regulars, including RMS, operate in and assume good faith.

Is LKML generally hostile to newcomers and not assuming good faith? I heard it had issues, but not this one.

They are only hostile to newcomers not willing to learn what is important for the project. Things like bug free code, code that does what is supposed to do in a way that does not interfere with other kernel parts, ...

That's a sentiment that every OSS project maintainer should take note of.

Good to know there are people stepping down off of open source project leadership roles without throwing a tantrum!

Emacs is a GNU package. While you are correct to call it "open source", its maintainers prefer to call it "free software", in order to emphasise freedom.

Better to call it a GNU package, to emphasize the GNUness. GNU's definition of freedom is quite specific, and not my definition. Better to call it libre, as there is less confusion.

It's no more specific than any definition that forbids others from taking freedom away.

Heading over to vi? :)

liVIn la VIda loca

All this over rtags[0] not in melpa/core? Typical RMS drama, but I see his point, and he reacted quite defensively. Not as aggressive anti-clang/apple as before.

0: https://github.com/Andersbakken/rtags

I remember in my C++ class around 1997 the professor was saying emacs was more a way of life or operating system than just an editor. He was only half-kidding.

In the past 18 years I imagine the functionality may be even more comprehensive, if that is possible?

I guess this will impact the Emacs 25 release?

Reply on news.ycombinator.com