Making Emacs popular again [LWN.net]

Making Emacs popular again

Posted May 6, 2020 22:49 UTC (Wed) by tchernobog (subscriber, #73595)
Parent article: Making Emacs popular again

I thought that what made VSCode popular was that you can write your extensions and config files in JavaScript / JSON.

Many developers are already familiar with those in their daily projects. eLisp is not something you want to get your hands dirty in 2020 as a fresh college graduate / young professional.

Not to mention all the long-standing issues (e.g. lack of multithreading which makes the UI block). There was at least an attempt to move to GNU Guile, but I am unsure it went anywhere.

But the amount and quality of extensions (especially language server integration) available with one click in VSCode are the clear winner for me, rather than the keymaps, icons, or rounded corners.

Making Emacs popular again

Posted May 7, 2020 1:15 UTC (Thu) by dvdeug (subscriber, #10998) [Link] (9 responses)

I wonder if switching to Scheme would matter. I see Scheme as a somewhat useful language to know, whereas Emacs Lisp is its own little cul-de-sac, and not highly respected as Lisp dialect.

Making Emacs popular again

Posted May 7, 2020 11:58 UTC (Thu) by smoogen (subscriber, #97) [Link] (5 responses)

I have seen people ask about it for 30 years, and there is no 'switching'. The people who tried found that using scheme is a complete rewrite from the ground up to make a Scheme editor... and after a while just ran out of oomph of trying to make it happen. There was also an attempt by some people 20? years ago to use Guile but it also ran out of oomph. Usually because they realized that instead of the tiny project they were running into a bigger one and also running into stop energy from emacs people who will point out that your tool doesn't have X and emacs already does so why rewrite it, and stop energy from people with XYZ editor who are very happy with it thank you very much.

All in all, if someone is going to do it, they need an ego of a superhuman to power through the giant valley of despair... and usually those people find other things to work on.

Making Emacs popular again

Posted May 8, 2020 2:56 UTC (Fri) by tome (guest, #3171) [Link]

> using scheme is a complete rewrite from the ground up to make a Scheme editor

A couple years ago I think Andy Wingo had gotten guile to the point where it could interpret emacs lisp. So the great mass of emacs functionality, including extensions, implemented in elisp, would run with guile swapped in as the implementation. There was no longer a need for a complete rewrite of all that elisp code. People who prefer to do so could write new packages in scheme and take advantage of guile's greater speed and capabilities. It came down to some remaining issues in the C code of emacs. I thought it sounded feasible, but I could be wrong. Can someone tell me what's untrue about that scenario?

Making Emacs popular again

Posted May 8, 2020 4:18 UTC (Fri) by marcH (subscriber, #57642) [Link] (3 responses)

> The people who tried found that using scheme is a complete rewrite from the ground up to make a Scheme editor... and after a while just ran out of oomph of trying to make it happen.

Readable Lisp S-expressions seem like they would be much easier to implement and maintain, wouldn't they? https://readable.sourceforge.io/
It wouldn't change the language but merely add a clearer syntax. I naively think it would help.

Making Emacs popular again

Posted May 10, 2020 5:14 UTC (Sun) by milesrout (subscriber, #126894) [Link] (1 responses)

S-expressions are already very readable, they just take a little getting used to. Once you've got used to them, they're much easier to read and understand than 'readable' expressions. I don't think really surface-level stuff like this helps.

Making Emacs popular again

Posted May 10, 2020 15:18 UTC (Sun) by marcH (subscriber, #57642) [Link]

> Once you've got used to them, they're much easier to read and understand than 'readable' expressions.

Thanks for your personal impression but please have a look at the reference.

Making Emacs popular again

Posted May 10, 2020 19:03 UTC (Sun) by jem (subscriber, #24231) [Link]

I'm not convinced. Lisp is known for having almost no syntax at all, but this big specification adds quite a lot. On top of that, it is backwards compatible, so you can freely mix the old S-expressions and the new syntax.

The specification also adds "meaningful indentation" à la Python (and Haskell), with all the cumbersome interaction between tabs and spaces, and problems with copy-pasting code.

My fix to the readability problem is to indent the code properly, and split the code into separate functions as appropriate. Use an editor which does the indentation automatically, shows matching parentheses, and warns about the "silly extraneous" parentheses. It doesn't really matter if a Lisp function ends with 13 closing parentheses, as long as they are all there. You don't have to count them.

Now, if I could only think of an editor that is up to the task...

Making Emacs popular again

Posted May 7, 2020 12:17 UTC (Thu) by Sesse (subscriber, #53779) [Link] (1 responses)

Those 95% not using Emacs are not going to use it because of a switch from elisp to Scheme. Both are super-obscure languages today.

Honestly, Emacs would probably be in a better place if RMS didn't have any say.

Great support, and lots of cool new tools

Posted May 11, 2020 19:47 UTC (Mon) by tnoo (subscriber, #20427) [Link]

> Honestly, Emacs would probably be in a better place if RMS didn't have any say.

Emacs is his baby, and he cared for it his whole life. And he still does: I had a reproducible crash a few years ago. It took one hour for him to send me a first patch that fixed the bug. And after one day of discussion on the mailing list the root cause was found and fixed. No vendor does that.

For me, the amazing thing is that emacs got a lot more momentum thanks to great tools like org-mode, magit, ivy/avy/helm, pdftools, org-ref that were not around ten years ago.

Making Emacs popular again

Posted May 25, 2020 17:00 UTC (Mon) by moltonel (subscriber, #45207) [Link]

That's where a switch to guile seems most appealing : it supports EmacsLisp, Scheme, ECMAScript (aka javascript), and soon Lua. I use Emacs all day every day, but I don't like lisp, and I'm sure many potential Emacs users are turned away by EmacsLisp. Contributing to Emacs would be much more appealing if you could pick and choose between any guile-supported language.

Making Emacs popular again

Posted May 12, 2020 1:15 UTC (Tue) by briangordon (guest, #135428) [Link] (1 responses)

Atom was there first on the JavaScript point, and seems to have gone even further to make every little thing pointedly modularized and extensible. Yet VSC came out on top. Maybe someone who used both editors can speak to why. My only guess is that the first-class (and conspicuous) git integration in VSC became a killer feature for people who wanted an editor equipped to interact with the burgeoning GitHub/GitLab community but who weren't totally comfortable with the command line interface.

Making Emacs popular again

Posted May 12, 2020 2:53 UTC (Tue) by Cyberax (✭ supporter ✭, #52523) [Link]

VSCode is just faster and more fluid than Atom.

(I prefer IntelliJ editors, they are unmatched in their semantic editing support)