frou_dh

Certain domains are a lot more wide open than others though. If you want to develop a command-line program that does something with files and possibly the network, then many many languages are suitable. If you want to develop a proper desktop GUI application (using mature design/layout tooling, naturally) then the field of suitable languages is cut down massively, and trying to shoehorn one in is indeed inadvisable.

There is a good IDE-style debugger available for Emacs these days: https://github.com/svaante/dape

Since it has no dependencies, I wouldn't be surprised if it gets merged into Emacs core at some point.

Also, the set of top-level things in the menu bar is not static. So even if you cannot directly interact with it for some reason, it gives you a hint that new things are possible in particular contexts. (Same goes for the 'tool-bar' that's distinct from the menu-bar)

An interesting thing about Smalltalk and Ruby blocks is that they aren't just anonymous functions/lambdas, right? i.e. if you 'return' / '^' in a block, it's the context around the block that you return from, not just the block itself? That's what struck me about both of them when I was used to thinking in basic Lisp terms.

I like having the identifiers of function/method calls being a distinguishable colour, because when looking at an overall section of code, they are the 'wormholes' to other code, and so one of the most useful things to be able to pick out instantly at a glance.

But I agree with the author that if little to nothing ends up being given the base text colour (e.g. white or black) then the design is an unthinking one (less of a design process, more of a box-ticking exercise to assign a unique colour to everything that can possibly have one).

When Python projects used that approach (setup.py files) that meant to just know what a package's dependencies were, arbitrary code had to be run. Now it's pyproject.toml

Sounds like a nice gig.

> It turns out that getting office workers to install a blessed plugin from Microsoft and navigate to a web page is much easier than distributing binaries that you have to install and keep up to date. And developing for it was pure pleasure; you got to use C# and Visual Studio, and a GUI interface builder

IIRC around that time, you could also distribute full-blown desktop applications (C# WinForms) in a special way via the browser, by which they were easily installable and self-updating. The tech was called ClickOnce https://learn.microsoft.com/en-us/visualstudio/deployment/cl.... I think the flow was possibly IE-only, but that was not a big issue in a business context at that time.

IMHO the document's introduction should explain why this double-slash // syntax is being used. It may be obvious to some readers, but not to me.

Is it a well known convention for a monorepo root-relative path?

I remember someone made an interesting point that the notable thing about traditional shell scripting languages (as opposed to conventional scripting languages) is that the filesystem is effectively the "top level scope" of the language. i.e. you can refer to files (paths) much like they are variables that are ambiently defined.

Not sure I like `command` failing silently, i.e. having to remember to check the .ok property. Most guidelines for writing sh/bash scripts arrive at using "set -e" mode unless you have a good reason not to. So when designing a new language, why not flip the default?

I do like Make so quite an interesting article, but the author seems overly impressed with their idea of using a literal ~ for the name of the subdirectory. That adds an unnecessary WTF factor to the code in my opinion :)

A counterbalance is that there's such a colossal volume of new YouTube 'content' published every day that approximately no one will end up watching an obscure video with your cameo anyway.

I guess the concern then shifts to dragnet automated surveillance of it.

Because you almost always have a specific type in mind for that function parameter, so you might as well just write it down.

I'm saying you need to bring your A-game to actually win against him in the online and particularly the Rails sphere - a formulaic petition isn't it. Even Apple found him hard to handle re the HEY iOS app.

Like him or loathe him, DHH is a serial winner. So today's petition¹ is tomorrow's fish and chips wrapper.

¹ https://github.com/Plan-Vert/open-letter

Yes - Chromium Embedded Framework (CEF) as used by some popular applications e.g. Spotify and Steam.

Personally I am completely uninterested in using this feature, but I really like iTerm2 overall.

This is just one feature of the v3.6 release:

https://iterm2.com/downloads/stable/iTerm2-3_6_1.changelog

It's quite a treat going through iTerm changelogs and finding new gems. For example, this sounds nice:

> [Timestamps] can be configured to be relative to a particular line by right-clicking and selecting "Set Baseline for Relative Timestamps"

The following is interesting too, because it seems to work on an individual cell basis and not just one overall background colour:

> [in editors and other TUIs] Detect when there is a non-default background color and extend it into the margins. In Minimal [theme], it is also extended into window chrome.

Maybe it's not so new, like "modernism" in art is 100 years ago.

There was actually an episode of the Ruby Rogues podcast released only a few days after this post that touched on much of the same stuff:

https://topenddevs.com/podcasts/ruby-rogues/episodes/distrib...

These days I honestly suspect tons of people do not think about the license they choose for their projects AT ALL and just reflexively put MIT. Like, if they were sleeping and you prodded them and said "license", they would mumble "MIT".

I wonder whether that kind of thing is actually bulletproof and doesn't end up having 2 running concurrently in some scenario.

Back when Twitter was less controversial, I remember tons of techie folks gleefully saying that they didn't bother with RSS any more because Twitter was better.

With decent RSS apps, you can generally just paste in the URL of any page (or the site's homepage) and they will take care of examining the HTML to find the URL of the actual feed.

Arguably the best is NetNewsWire, which has been around in various forms for over 20 years and is still developed today https://netnewswire.com

The more pedestrian reason that's there's demand for static typing is that when people press '.' they generally want a menu to appear and for it to be complete and accurate.

Listening to this episode, I was quite surprised to hear that even now in Sept 2025, support for classes at all is considered a medium-term goal. The "superset of Python" angle was thrown around a lot in earlier discussions of Mojo 1-2 years ago, but at this rate of progress seems a bit of a pie-in-the-sky aspiration?

That's https://opensource.janestreet.com/core/. However I think its importance is often a bit overblown. I doubt MOST people were choosing it at any point, never mind today.

It's an interesting road not taken, where there could have been much more commonality in application development between Linux and OS X.

Seems it also had more lofty goals than just making GNUstep mainstream.

Many HN users are always looking for an excuse to complain about the format or appearance of the linked sites. So much so that the following had to be added to the guidelines (but doesn't seem to be enforced that much):

> Please don't complain about tangential annoyances—e.g. article or website formats, name collisions, or back-button breakage. They're too common to be interesting.

https://news.ycombinator.com/newsguidelines.html

The main page https://www.energydashboard.co.uk/live is just as interesting as the map that's linked to.