Hacker News Reader

Comment by nvy | Hacker News Reader

The emacs out-of-box experience is notoriously poor, as evidenced by the rise and popularity of emacs starter packs, some of which have been on the hn front page of late.

I've been using emacs since 2002 or so and until very recently it needed a ton of third party config to approximate the ux and utility of what was available in other tools built in.

View on HN

AFAIU this is a discoverability problem, not that emacs isn't featureful ootb. Some of emacs's built-in features (e.g. undo in region) are even not supported at all by other editors.

So just use a starter kit. Why do people put so much importance on "out of the box experience"? Who actually cares? Out of the box only matters if you can't change what you get from the box. But you can. That's the entire point of emacs.

Emacs is the way it is because it's almost 40 year old software that still tries not to break users' configs. Do VS Code users even have configs? Most of them seem to just lap up whatever is dished out to them.

Because the "out of box experience" turns lots of users away, including future potential contributors. Emacs could be so much better if it had a larger pool of users from which to draw developers. Read the mailing lists- it's plain to see that Eli and a few others are seriously burnt out. The project is not healthy, despite outward appearances.

There's a sizeable group of emacs users who are "it works for me" grognards and for whom any change to make emacs better constitutes an assault on them as a person.

If that's you, just down vote and move on and continue enjoying living in your ensconced world of mediocre software. I don't have energy to have this argument yet again.

It would certainly be good to ensure the future of emacs. I'm not sure the set of people who care about out of the box and the set of people who would be future maintainers is a very large intersection, but it's probably nonzero. Personally over the years I've come to think emacs just isn't for everyone. Not everyone wants to be free. The Matrix told us that. I've come to accept I'll always be part of a small group of weirdos who like this stuff and that's fine. But if there's a genuine move towards ensuring the future of emacs by changing defaults then I'm all for it.

Are there any recent practical discussions about this? What's the barrier to someone simply taking something like emacs bedrock and making a snap out of it? I just checked and VS Code simply offers a .deb or .rpm for Linux. Wouldn't including an init.el make everyone happy? Or is there more to it?

Also, what is a good out of the box experience for emacs? I don't think it's code completion or whatever VS Code does. I think it's being easy to hack. That's what makes emacs great. It's not about what it does, it's about what you can do with it.

>I'm not sure the set of people who care about out of the box and the set of people who would be future maintainers is a very large intersection ...

Yeah that's the common refrain. It's just gatekeeping in my experience. A certain demographic tends to view using the terminal or using bad UIs as some sort of badge of elite status, and sneers at people who want to spend their workday looking at visually-pleasing UIs.

>Also, what is a good out of the box experience for emacs?

Well, a GUI that doesn't look like it's stuck in 1997 for starters. Off the top of my head, handling long lines without chugging, being async/multithreaded so a network request doesn't lock up the entire fscking UI, offering CUA-mode by default for easier onboarding, fixing the often absurd terminology in the docs (frames/windows/panes/yank means paste/etc.), and probably a few others I can't think of.

Have a look here: https://code.visualstudio.com/docs/getstarted/userinterface

That's just an immensely better UI, by default. To achieve even 80% of that in emacs you'd need to spend time configuring at a minimum treemacs, an lsp client, breadcrumb-mode, minimap, probably project.el or projectile, magit, one of the tab-bar modes, and a bunch of niceties like showing line numbers by default, to say nothing about a better theme.

All that stuff should be built in and there's no good reason other than "the emacs developers are overworked already" for them to not be included.

Imagine showing someone the editor above, and then showing them this. Which one do you think they'll choose?

https://www.gnu.org/software/emacs/images/teaser.png

It just looks amateurish and unpolished. It looks like it's janky and doesn't work very well. It's just unappealing in any modern sense.

I mean the VS Code one looks terrible to me. Fat icons on the left wasting space. Illegible text wasting space on the right. More wasted space in the bottom right. The text doesn't even fit in the buffers?! And this is their marketing shot?

I guess that's why I'm not in marketing.

Yeah, the Emacs one looks shit. They should definitely expunge those screenshots and replace them with ones that actually look good.

It's impossible to convey what makes Emacs great in screenshots, though. It's not the theme, it's not even magit or org-mode. I feel like someone with skills could make a good short video on what actually makes it great. Unfortunately most Emacs users are borderline autistic like me, though.

People will always be drawn to flash marketing, that's just the way it is.

I still don't get why someone (like you) doesn't just bundle a starter kit and make it easy to install standalone if you think this is the problem. Spacemacs has a shiny website but its installation instructions are `git clone spacemacs ~/.emacs.d` :/

> doesn't just bundle a starter kit

because user exposure (discoverability) will be tiny compared to fixing the OOTB in the "official" version

> It just looks amateurish and unpolished. It looks like it's janky and doesn't work very well. It's just unappealing in any modern sense.

I was with you before but now I think I might be glad that people like you get ignored by emacs developers. Do you also think that Craigslist looks amateurish, unpolished, and unappealing in any modern sense?

I'm all for improving usability (I agree with all the other points you made in this post) but I don't want my software to look "modern", simply because in general, UI design has gotten worse over the last 20 years.

> Who actually cares?

everyone who doesn't want to waste many hours tweaking the basics that have been already fixed by countless people before

> Emacs is the way it is because it's almost 40 year old software that still tries not to break users' configs

More likely it's the way it is because of attitudes like in the previous point discounting the obvious flaw of this mighty app

Reply on news.ycombinator.com