Today, even the minimal steps of creating a desktop app have lost their appeal, but I like showing how I solved a problem, so my "apps" are Jupyter notebooks.
- CAD / ECAD
- Artist/photos
- Musician software. Composing, DAW etc
- Scientific software of all domains, drug design etcDesktop publishing.
Brokerage apps (some are webapps but many ship an actual desktop app).
And yet, to me, something changed: I still "install apps locally", but "locally" as in "only on my LAN", but they can be webapps too. I run them in containers (and the containers are in VMs).
I don't care much as to whether something is a desktop app, a GUI or a TUI, a webapp or not...
But what I do care about is being in control.
Say I'm using "I'm Mich" (immich) to view family pictures: it's shipped (it's open source), I run it locally. It'll never be "less good" than it is today: for if it is, I can simply keep running the version I have now.
It's not open to the outside world: it's to use on our LAN only.
So it's a "local" app, even if the interface is through a webapp.
In a way this entire "desktop app vs webapp" is a false dichotomy, especially when you can have a "webapp (really in a browser) that you can self-host on a LAN" and then a "desktop app that's really a webapp (say wrapped in Electron) that only works if there's an Internet connection".
Most things I create in my free time are for my and my family's consumption and typically benefit immensely from the write once run everywhere nature of the web.
You can launch a small toy app on your intranet and run it from everywhere instantly. And typically these things are also much easier to interconnect.
KDE has analytics, they're just disabled by default (and I always turn them on in the hopes of convincing KDE to switch the defaults to the ones I like).
Times have changed quite a bit from nearly 20 years ago.
These concerns may not matter to you, the developer, but they absolutely matter to end-users.
If your prospective user can't find the setup.exe they just downloaded, they won't be able to use your software. If your conversion and onboarding sucks, they'll get confused and try the commercial offering instead. If you don't gather analytics and A/B test, you won't even know this is happening. If you're not the first result on Google, they'll try the commercial app first.
Users want apps that work consistently on all their devices and look the same on both desktop and mobile, keep their data when they spill coffee on the laptop, and let them share content on Slack with people who don't have the app installed. Open source doesn't have good answers to these problems, so let's not shoot ourselves in the foot even further.
If a piece of software doesn’t have users and the developers don’t care about the papercuts they are delivering, I would argue what they have created is more of an art project than a utility.
Art works without popular appeal can become highly treasured by some.
Open source software doesn't have to be ambitious to be worthwhile and useful. It can be artful, utilitarian or a artifact of play. Commercial standards shouldn't be the only measure of good software.
Good! It's not for them! They can stay paypigs on subscription because they can't git gud!
And his point about randomly moving buttons to see if people like it better?
No fucking thanks. The last thing I need is an app made of quicksand.
The user interface is your contract with your users: don't break muscle memory! I would ditch FF-derivatives, but I'm held hostage by them because the good privacy browsers are based on FF.
Stop following fads! Be like craigslist: never change, or if you do then think long and hard about not moving things around! Also if you're a web/mobile developer, learn desktopisms! Things don't need to be spaced out like everything is a touch interface. Be dense like IRC and Briar, don't be sparse like default Discord or SimpleX! Also treat your interfaces like a language for interaction, or a sandbox with tools; don't make interfaces that only corral and guide idiots, because a non-idiot may want to use it someday.
I really wish Stallman could be technology czar, with the power to [massively] tax noncompliance to his computing philosophy.
For some things a desktop app is required (more system access) or offers some competitive UX advantage (although this reason is shrinking all the time). Short of that user's are going to choose web 95% of the time.
Ignoring the fragmentation of course; although that seems to be getting less and less each year (so long as you ignore Safari).
The impact on people's time, money and on the environment are proportional.
Does it? Have you compared a web app written in a sufficiently low level language with a desktop app?
And if we're talking about simple GUI apps, you can run them in 10 megabytes or maybe even less. It's cheating a bit as the OS libraries are already loaded - but they're loaded anyway if you use the browser too, so it's not like you can shave off of that.
What about in QML, which uses Web technologies like CSS, JS and even basic HTML? The whole KDE Plasma 6 desktop is built around these technologies now and I (and many others) consider it light and high-performance.
If you saddle up those technologies in the full browser everything then it will get larger, yes, but nothing requires you to do this, just as nothing requires providing your app as a full-fat Fedora install when a distroless container would have sufficed.
Plain Javascript can be very fast and still come at relatively low resource demands and the same is true of HTML and CSS. Many "plain desktop-native" applications often end up reinventing their own variants of HTML and CSS in the course of designing the U/I anyways.
Qt is much lighter than your Chromium-based stacks but all the waste kind of adds up.
"just as nothing requires providing your app as a full-fat Fedora install when a distroless container would have sufficed" Containers are hungrier than running stuff on bare metal...
> Containers are hungrier than running stuff on bare metal...
Containers are tremendously lightweight compared to VM. You might as well point out that running a full multiuser security-protected OS like Linux is hungrier than running on bare metal with DOS too. It's just as true, and even proportionally as true.
In any event a full Fedora container with all packages installed is going to be tremendously larger than a distroless hello-world "built" around Alpine, for instance, even though they both use container technologies. Same applies to Web technologies, you can certainly go and easily add a lot of waste using them but they are not themselves inherently wasteful.
A desktop app may consume more, but it's heavily focused on one thing, so a photo editor don't need to bring in a whole sound subsystem and a live programming system.
Counter-counterpoint: Maybe it's time to require professional engineer certification before a software product can be shipped in a way that can be monetized. It's to filter devs from the industry who look at browsers today and go "Yeah, this is a good universal app engine."
Maybe useful higher-level elements like layout, typography, etc. could be shared as frameworks.
There are many alternate histories where a different base application layer (app engine) could have been designed for the web (the platform)
Remember Livescript and early web browsers? It was almost cancelled by big tech because Java was supposed to be the cross platform system. The web and Javascript just BARELY escaped a big tech smack down. They stroked the ego of big tech by renaming to Javascript to honor Java. Licked some boots, promised a very mediocre, non threatning UI experience in the browser and big tech allowed it to exist. Then the whole world started using the web/javascript. It caught fire before big tech could extinguish. Java itself got labeled a security threat by Apple/Microsoft for threatening the walled gardens but that's another story.
You may not like browsers but they are the ONLY thing big tech can't extinguish due to ubiquity. Achieving ubiquity is not easy, not even possible for new contenders. Pray to GOD everyday and thank her for giving us the web browser as a feasible cross platform GUI.
Web browser UI available on all devices is not a failure, it's a miracle.
To top it all off, HTML/CSS/Javascript is a pretty good system. The box model of CSS is great for a cross platform design. Things need to work on a massive TV or small screen phone. The open text-based nature is great for catering to screen readers to help the visually impaired.
The latest Wizbang GPU powered UI framework probably forgot about the blind. The latest Wizbang is probably stuck in the days of absolute positioning and non-declarative layouts. And with x,y(z) coords. It may be great for the next-gen 4-D video game, but sucks for general purpose use.
You've reminded me of the XKCD comic about standards: https://xkcd.com/927/
Do you really want a universal app engine? If you don't have a good reason for ignoring platform guidelines (as many games do), then don't. The best applications on any platform are the ones that embrace the platform's conventions and quirks.
I get why businesses will settle for mediocre, but for personal projects why would you? Pick the platform you use and make the best application you can. If you must have cross-platform support, then decouple your UI and pick the right language and libraries for each platform (SwiftUI on Mac, GTK for Linux, etc...).
That's a terrible solution that preserves nothing. Try using a screen reader with an app rendered onto a rectangle.
It would have been great if browsers remained lightweight html/image/hyperlink displayers, and something separate emerged as an actual cross-platform API, but history is what it is.
Let's also remember that it's infinitely easier to keep a native app operational, since there's no web server to set up or maintain.
As a user, properly implemented desktop interface will always beat web. By properly, I mean obeying shortcut keys and conventions of the desktop world. Having alt+letter assignments to boxes and functions, Tab moves between elements, pressing PageUp/PageDown while in a text entry area for a chat window scrolls the chat history above and not the text entry area (looking at you SimpleX), etc.
Sorry, not sorry. Web interface is interface-smell, and I avoid it as much as possible. Give me a TUI before a webpage.
There are some things that NATURALLY lend themselves to a website - like doctor's appointments, bank balance, etc - but it's still a pain when, on logging in to "quickly check that one thing" that I finally got the muscle memory down for because I don't do it that often, I get a "take a quick tour of our great new overhauled features" where now that one thing I wanted is buried 7 levels deep or something, or just plain unfindable.
For something like Audacity (the audio program), how the heck does it make sense to put that on a website (I'm just giving a random example, I don't think they've actually done this), where you first have to upload your source file (privacy issues), manipulate it in a graphically/widget-limited browser - do they have a powerful enough machine on the backend for your big project? - then download the result? It's WAY, WAY better to be able to run the code on your own machine, etc. AND to be stable, so that once you start a project, it won't break halfway through because they changed/removed that one feature your relied upon (no, not thinking of AI at all, why do you ask? :-)
I understand it was just an example, but you'd be surprised how far browsers have come along with technologies like Web Assembly and WebGL. Forget audio editing, you can even do video editing - without uploading any files to the remote server[1]. All the processing is done locally, within your browser.
And if you thought that was impressive, wait till you find out that you can can even boot the whole Linux kernel in your browser using a VM written in WASM[2]!
But I do agree with your points about lack of feature stability. I too prefer native apps just for the record (but for me, the main selling points are low RAM/CPU/disk requirements and keyboard friendliness).
And if this is such a compelling value proposition for full-featured desktop productivity applications, why didn't Java Web Start set the world on fire?
It's the issue of friction. Also, good webapps are often _better_ than native apps, as they can support tabs.
> And if this is such a compelling value proposition for full-featured desktop productivity applications, why didn't Java Web Start set the world on fire?
Because it relied on Java and SWING, which were a disaster for desktop apps.
All the native apps I use support tabs, its a basic feature of the macOS windowing APIs https://developer.apple.com/documentation/appkit/nswindowtab...
Putting aside the video editing example for a bit, consider the photo editing web app Photopea, which is an excellent alternative to Adobe Photoshop. Linux is in urgent need of a Photoshop-like editor (and no, GIMP doesn't cut it), but Photopea does a decent enough job for many amateurs and even some pros. For a lot of these folks, Photoshop is one of the last things stopping them from switching to Linux, so apps like Photopea fill that gap. And guess what, Photopea works great on Android too.
Another use case is restricted environments where you can't easily find and install apps, eg immutable distros, or work computers. I use Photopea on my work PC quite regularly for light editing, because MS Paint sucks, and my role doesn't really justify going thru the hassle of getting the approvals to get an editor installed. So like it or not, web apps have their place.
How is Photopea better than GIMP? How is it better than Krita?
As for Krita, its UI is of course a lot better than GIMP, but unfortunately it's mostly skewed towards digital illustration and art creation (and it's great at it!), but less towards photo editing/image manipulation.
- Photopea has the best .PSD support of the three, which is pretty crucial for people wanting to switch from Photoshop.
- Possibly the most important feature that Photoshop users depend on these days is Content-Aware Fill and Magic Replace for object removal and background patching. GIMP lacks native functionality for this (although there are thirdparty plugins, but I haven't used them so can't comment on that). As for Krita, once again Krita lacks these tools - and most retouching tools in fact - as it's more geared towards digital art creation rather than image manipulation.
Yeah, but as a maintainer it's the opposite, isn't it? I don't have to worry about supporting version current - 3 in the Polish version of Windows because you're always running the version I've deployed in the environment I've deployed it in (I mean, yes, I'm oversimplifying given the frontend component, but that's still a much smaller surface).
I grew up reading his writings and learned pretty quickly to read them as "this is what I'm thinking right now in my life" even though they're written more as authoritative and decisive writings from an expert. Over time he's gone from SEO expert to $30K/week consulting expert to desktop app expert to indie SaaS expert to recruiting industry expert to working for Strip Atlas. It was fun to read his writings at each point, but after so many changes I realized it was better to read it as a blog of ongoing learnings and opinions, not necessarily as retrospective wisdom shared from years of experience on the topic even if that's what the writing style conveys.
So I agree that the advice in the post should be taken entirely in context of pursuing the specific goals he was pursuing at the time. The less your goals happen to align, the less relevant the advice becomes.
This, I like his writing, and am subscribed to his Bits About Money newsletter which gives a lot of info about how payment processing works, but your comment is spot on about the writings being influenced by his current work and the current place he is within his life. Hopefully everyone thinks about those sorts of things when reading articles like this, regardless of author because it's a common issue.
If development ends at a git push and users are left to build/fend for themselves (granted this is a lot of open source), then yeah not much difference, but if you're building and packaging it up for users (which you will more likely to be doing if your project is an app specifically) then the difference is massive.
No, it's a concern if you care about impact. Improving commercial profits is one kind of impact that is relevant to for-profit corporations, but there is also impact like "improving user privacy" or "helping lower-income people manage their finances with a free-as-in-beer product". This impact can be measured and the feedback can be used to improve the product according to non-profit, non-commercial goals.
There are also people who build open-source software as a hobby and couldn't give two shits whether other people use it or not. More power to them. For those people, you are correct. https://book.iced.rs/philosophy.html comes to mind.
Then there are projects like Streisand (maybe a bad example, I see it has since been archived, but it came to mind) that want to change the world in some way. Those projects very much do need to care about metrics like, how many people are downloading the software, are people opening GitHub issues, are we obscure or is our target audience talking about us, hopefully positively but if not, how can we improve that? Value must always be worth the cost (even when the code is free, it must be worth the time to download, give it a try, give it CPU/RAM, maintain/upgrade the installation) - are we giving users value or are they churning?
It might blow your mind but even non-profits hire people with MBAs (and universities offer programs for MBAs that focus on non-profit management), precisely because some organizations focus on non-financial impact.
The onboarding funnel: Only a concern if you're trying to grow your user base and make sales.
Conversion: Only a concern if you're charging money.
Adwords: Only a concern if, in his words, you're trying to "trounce my competitors".
Support: If you're selling your software, you kind of have to support it. Minor concern for free and open source.
Piracy: Commercial software concern only.
Analytics and Per-user behavior: Again, only commercial software seems to feel the need to spy on users and use them as A/B testing guinea pigs.
The only point I can agree with him that makes web development better is the shorter development cycles. But I would argue that this is only a "developer convenience" and doesn't really matter to users (in fact, shorter development cycles can be worse for users as their software changes rapidly like quicksand out from under them.) To me, in my open source projects, my "development cycle" ends when I push to git, and that can be done as often as I want.