this post was submitted on 25 Feb 2024

19 points (59.2% liked)

Unpopular Opinion

6186 readers

130 users here now

Welcome to the Unpopular Opinion community!

How voting works:

Vote the opposite of the norm.

If you agree that the opinion is unpopular give it an arrow up. If it's something that's widely accepted, give it an arrow down.

Guidelines:

Tag your post, if possible (not required)

If your post is a "General" unpopular opinion, start the subject with [GENERAL].
If it is a Lemmy-specific unpopular opinion, start it with [LEMMY].

Rules:

1. NO POLITICS

Politics is everywhere. Let's make this about [general] and [lemmy] - specific topics, and keep politics out of it.

2. Be civil.

Disagreements happen, but that doesn’t provide the right to personally attack others. No racism/sexism/bigotry.

3. No bots, spam or self-promotion.

Only approved bots, which follow the guidelines for bots set by the instance, are allowed.

4. Shitposts and memes are allowed but...

Only until they prove to be a problem. They can and will be removed at moderator discretion.

5. No trolling.

This shouldn't need an explanation. If your post or comment is made just to get a rise with no real value, it will be removed. You do this too often, you will get a vacation to touch grass, away from this community for 1 or more days. Repeat offenses will result in a perma-ban.

Instance-wide rules always apply. https://legal.lemmy.world/tos/

founded 1 year ago

MODERATORS

[email protected]

If you're making software for actual end-users, you HAVE to give it a goddamn GUI, or else you suck, your software sucks, and nobody is going to use your damn software. (lemmynsfw.com)

submitted 6 months ago* (last edited 6 months ago) by [email protected] to c/[email protected]

66 comments fedilink hide all child comments

You see this shit SO much more often than you would think. And the infuriating thing is, it seems to be most common among programs that are INCREDIBLY complex and sophisticated.

It'll be like this:

"What does my program do? Glad you asked. It simulates stress patterns in glass and ceramics, after they come out of a kiln. You can specify any melting temperature, adjust the composition of elements in the glass, and the ambient temperature of the cooling and tempering stages."

"Wow, can you show me how it works?"

"Sure! "

"O-oh. Do you have any plans to add a graphical user interface?"

"HAHAHAHAHHA, no. That's never happening. And here I thought you were serious about using advanced software, and being an intelligent person."

Obviously, that last part is just kinda implied. But sometimes, when users request a GUI, the goddamn developer will kinda get in their face, like that.

They always fall back on the position of "well, I developed this shit for free, for your ungrateful ass. So you can build your own fucking GUI."

But the thing about that is...no. And fuck you. I shouldn't have to be two-thirds of a fucking developer, in order to use the fucking software.

If you can figure out how to simulate molecules, or draw 3D stereograms, or translate hieroglyphics, or any other RIDICULOUSLY COMPLICATED SHIT, making a graphical user interface should be nothing to you. You should be able to do it in a fucking afternoon.

IT DEFINITELY SHOULD BE THE EASY PART, FOR YOU.

All the rest of us, who aren't programmers? We envy programmers, and their ability to really connect with computers, on that deep logic level.

If we could do that shit, we would. But a lot of us have tried, and we realize it's not a good use of our time. We can do cool stuff with software, but it's just not ever going to be worthwhile for us to struggle through the act of creating software.

Also, I hasten to add that I have put in my time, using command line interfaces. I used DOS, I used BBS systems, I have used modern command-line-only programs. I know how to do it, but I DON'T WANT TO.

I don't want to have to memorize commands. I don't consider a GUI workflow to be some kind of weird luxury. It has been a basic part of modern software, for around 40 years at this point. Literally get with the program, guys.

If you're serious about making software, get your shit together and implement a fucking GUI from the very first release. Nobody ought to be taking you seriously, if you refuse.

top 50 comments

sorted by: hot top controversial new old

[–] [email protected] 2 points 6 months ago

Idk man. Neofetch is pretty good and that doesn't use a GUI.

[–] [email protected] 1 points 6 months ago

Eventually AI will just generate a GUI for you. In the meantime, ShellGPT will write the shell commands for you. But for the most part, GUIs are a weird luxury and you should just learn some bash, it’ll make you feel better!

[–] [email protected] 18 points 6 months ago

They always fall back on the position of "well, I developed this shit for free, for your ungrateful ass. So you can build your own fucking GUI."

But the thing about that is...no. And fuck you. I shouldn't have to be two-thirds of a fucking developer, in order to use the fucking software.

This took it from being "unpopular opinion" to "I'm just a fucking asshole and my opinion doesn't matter."

Or in other words, fuck you, you ungrateful cunt.

The dev making the ceramic stress testing whatchamawhatsit probably has zero skills in making a GUI. The next thing you'd be complaining about is that there's some bug in the GUI, because you're a complete moron and keep doing something stupid. Something the dev didn't ever think someone would do, because they're not a moron, and don't think like one.

They're making the tool because they have a passion for whatever it is they have a passion for. Not because theyre trying to appease your moronic simple mind.

If you're so passionate about GUIs, learn to make a fucking GUI for a command line tool. You fucking tool.

[–] [email protected] 13 points 6 months ago

If you can figure out how to simulate molecules, or draw 3D stereograms, or translate hieroglyphics, or any other RIDICULOUSLY COMPLICATED SHIT, making a graphical user interface should be nothing to you. You should be able to do it in a fucking afternoon.

In a word, no. Being able to build an engine doesn't mean you know the first thing about how to design a car. It's a totally different skill set.

I work with PhDs who code all sort of amazing physics engines and then I design the GUI apps around them. That's a full-time job right there (I'm living proof of it), and I wouldn't expect them to understand it any more than they would expect me to understand all the physics.

When you write some sort of procedural tool, you are in complete control of the program flow from start to finish. In a GUI app, the user is in control most of the time. That's awesome if you're the user, but it means a lot more what-if scenarios you have to account for, since users are notoriously unpredictable. And if the task your command line was performing takes an appreciable length of time, you need to spawn it off into separate threads or subprocesses and worry about all the synchronization logic you must get right. This is a programming minefield for anyone who has done it, especially when you need said threads to interact with the GUI, as GUI frameworks are notoriously not thread-safe.

Anyway, what I'm trying to say is designing and implementing GUIs is non-trivial, unless maybe you just want something like an installer wizard that runs you through 10 dialogs to gather info for a command line and then runs it.

[–] [email protected] 3 points 6 months ago

Nope, don't agree. I definitely write software, but the UI for the end user is (usually) exactly one button. What would I need an GUI for?

[–] [email protected] 7 points 6 months ago (1 children)

I actually kinda agree with this.

Even if a commandline tool has more power, utility and flexibility than a gui tool, guis let you drill down into the core workflow of what your software is meant to do. At a minimum, it lets you segment what your regular users are expected to do from what your power users could use your software to do.

Additionally, if you intend for your software to be used by non-software adjacent users, a commandline interface is just asking for people to get lost/confused.

At work, we use ROS on some of our systems and while the commandline tools are simple to someone who works in the ecosystem, knowing what to look at when things go wrong is tricky.

Even a simple gui in tkinter that shows statuses or shows a list of topics and lets you print them out is leagues above the commandline when it comes to how much I need to be involved in other peoples' problems.

It is a luxury to be knowledgeable in software concepts and I think software devs/power users forget that often.

[–] [email protected] 0 points 6 months ago

I mentioned VisiCalc to a couple of people, in this discussion, and I think it REALLY needs to be remembered more often, and more clearly.

I keep seeing a lot of people replying to various parts of this discussion, with varying degrees of this opinion:

"Anything you can do with a GUI can be done without a GUI. You basically just THINK you need a GUI, because you're a lazy non-software person. You just need to get used to using CLI, and when you do, you'll thank me, because you'll be able to do even more powerful things than you ever could, with your GUI crutch."

Well, spreadsheets blow that entire argument out of the water, instantly.

The whole thing about spreadsheets is the fact that they are a GUI. The original VisiCalc spreadsheet application BARELY had a GUI. It was all text characters. It was barely running in realtime. But it was a realtime GUI, and that was the entire reason that it was useful.

Note that you could make a command-line-only spreadsheet program. As in, an entirely hidden spreadsheet, where you use commands to define the fields, enter the data, and then use other commands to do operations and query for the results. But that's entirely destroying the whole point of the spreadsheet, as a tool. With VisiCalc (and all modern spreadsheet programs), you can see all the fields at once, change them in any order, and watch as any other relevant fields are affected by the different numbers.

As many people have said about the spreadsheet revolution, it allowed people to "play around with the numbers." For most users, that was what elevated the microcomputer from a mere calculator to a game-changing, paradigm-shifting, universally necessary tool, for anyone who did any kind of business.

Before VisiCalc, you might as well just use a digital calculator, to help you with your business stuff. Just using command-line programs to do calculations wasn't really any faster than a desktop calculator. And those desktop calculators usually had hard-copy printers built into them, so that you wouldn't ever have to re-list the queries you'd made, to refresh your memory.

But, again, the realtime GUI spreadsheet blasted all that apart, and gave users life-altering tools. It wasn't just a matter of "oh, look. We dumbed down the interface, so that stupid Bob the Business Major can finally use software." It was a real step beyond what command-line tools can do.

Now, think about my example of a molecular simulation program for melting, cooling, and tempering ceramics. Isn't that process even more complex than doing sum and division operations on some dollar values, rounded to two digits after the zero? Doesn't it logically follow that you'd want a GUI for a molecular simulation tool, so that you could PLAY AROUND WITH THE NUMBERS, in the same way that you would, with your business expense sheets?

That's the whole point of the software doing simulations, right? On the one level, it's simply easier to poke around at finite-state machines that model known physical and chemical properties than it is to do a bunch of experiments in a real kiln. That's the first stage of why simulation is good for progress. But it really does go beyond that. I would argue that a GUI in a simulation lets you see things from angles that you would never have thought of, before you started tinkering.

If you have to think about how you're going to query the system, on a command line, then read out the data afterward, you're going to run up against the limitations of the human imagination. Some users will think of every query they ought to make, without any prompting, and with less trial and error. Others will never think of certain aspects of the simulation, unless and until they can SEE the parameters laid out in front of them, in a graphical, preferably realtime fashion.

Obviously, not everything can be realtime. But graphical interfaces always allow better viewing of simultaneous and parallel functions.

It's not just a matter of "we'll throw some of these non-techie users a bone, and make this stuff easier for their lazy asses." There ARE fundamental advantages to a GUI, which truly to go beyond simple "user friendliness." Succinctly, a GUI user is not the opposite of a power user.

[–] [email protected] 3 points 6 months ago* (last edited 6 months ago)

I don't have experience to add to the discussion, but it seems like this is a genuinely unpopular opinion from the comments, so it gets my upvote

[–] [email protected] 8 points 6 months ago

By all means, provide a GUI.

But being able to do it all from the command line means you can pipe the output of another program directly into your program, and then take that output and feed it into yet another.

The program you describe could be slotted into the control systems of the kiln, cooling, and tempering ovens.

GUIs break that connection. As soon as you start using a GUI, the ability to automatically pipe outputs to inputs is broken. The user is now the link between the programs, manually transferring data between them that could be performed automatically.

All the rest of us, who aren't programmers? We envy programmers, and their ability to really connect with computers, on that deep logic level.

A good GUI will leave the command line visible, and interact able, so that the user can see the program's input and output syntax, and gain that sense of "connection". A good GUI will teach the command line, so that by merely using two programs, the user learns what they need to connect them together.

[–] [email protected] 1 points 6 months ago* (last edited 6 months ago)

If you mod Skyrim or Fallout: chances are you use the Script Extender... Which runs in a cmdline window without a GUI.

[–] [email protected] 5 points 6 months ago

This software is not for you. It's for the author. It's in furtherance of their research, or their career, or their interests. It is only incidentally available to you.

That's it, that's the whole story. You get a GUI if the authors wanted there to be a GUI.

If you join a community around the product - which would require being nice, willing to engage, understand, not make demands - you might be able to give feedback and occasionally help steer the direction of the product.

[–] [email protected] 21 points 6 months ago* (last edited 6 months ago) (1 children)

People who do things they like for free are called "hobbyists". These hobbyists, unsurprisingly, only do what they want to do. And sometimes they don't want to make a GUI.

The fact that they share the results of their hobby for free at all is just a bonus.

[–] [email protected] -3 points 6 months ago (2 children)

I honestly do get that. And I realize I am overstating my case, in a way that basically makes me an asshole.

I think the problem is often rooted in how projects advertise themselves. The small dev team is like "WE'RE PASSIONATE ABOUT GETTING ________ INTO THE HANDS OF USERS, BECAUSE WE KNOW THE ONLY OTHER SOFTWARE THAT DOES _______ IS A 27-YEAR-OLD APP THAT ONLY WORKS ON A SPECIFIC REVISION OF WINDOWS 95."

But then the damned app is command-line only, and it just feels like it was all a tease.

[–] [email protected] 6 points 6 months ago

This raises a point though. A command line tool from '95 will likely recompile and run just fine with maybe a warning or two, while the GUI app is no longer supported because GUI frameworks are notoriously fickle things that go obsolete all the time.

[–] [email protected] 7 points 6 months ago (2 children)

But if you're this passionate about the stress levels of ceramic whatever, presumably you've trudged through countless dense works of academia. After that how big of a step is it to learn to type a few keywords into a command line? You're not required to learn a whole programming language.

[–] [email protected] 4 points 6 months ago

I've dealt with that kind of software that was a paid for product.

Part of the issue is that the software is that it is very specialized, so there isn't a big demand for the software. At that point, it is usually cheaper to train the users instead of making the program easier to run.

You also run into problems where a GUI can cause a false sense of confidence in the work. People end up trusting the program completely instead of using it as a tool that could cause problems.

[–] [email protected] 0 points 6 months ago* (last edited 6 months ago)

That's a hard argument to rebut. The only thing I can really refer to is VisiCalc. The OG spreadsheet program. The original "killer app."

The whole thing about a spreadsheet is that it's a realtime pseudo-GUI. It's not a command-line-interface thing, and that's VITAL TO ITS FUNCTION, AT THE END-USER LEVEL.

Of course, the computer can do any of the spreadsheet calculations by means of querying a CLI. But the thing about a spreadsheet is that you just change one element on the screen, and all the other elements auto-update, in response to the altered field.

That functionality was a major-ass deal for doing some summing and dividing, so you don't fuck up your payroll and your expense reports. It increased productivity. It reduced clerical errors.

But you're over here, like "well, can't you just use the command line to enter the new parameters for your molecular glass simulation?"

My argument: isn't that shit just a little more complex than a spreadsheet with some basic monetary figures on it? If a GUI is a game-changer for juggling a few dollar-figures in a monthly budget document, it's EVEN MORE CRITICAL in any application where both the data and the calculations are more complicated.

The GUI paradigm isn't a frivolous luxury, which is only made for those of us who are just too lazy to get savvy with the CLI game. GUI has real and exclusive benefits. The more complex the data entry, the sillier it becomes to just say "you don't really NEED that GUI, you just want it for convenience."

[–] [email protected] 11 points 6 months ago* (last edited 6 months ago)

Conversely, I think that every program should have a way to operate it from the command line and the GUI can just needlessly get in the way a lot of the time `¯_(ツ)_/¯
You write a command once and you can do whatever you want with it, repeatably, forever. It's fantastic! You can then pipe the results into a completely different program all without needing to do much of anything! Want to run it on a schedule? Easy! Want to send the exact parameters you used to someone else? Easy! Want to copy and paste the exact output? Easy! Want to get a daily email with the results? Easy!
Some things are better with a GUI, but, plenty don't need it.
With all that said, I understand why a dev who didn't build their program with CLI in mind wouldn't want to go in afterwards and add it in, especially if they don't use/like CLI interfaces.

I've seen similar arguments to yours that every program should have be packaged as an exe, because fuck those fucking programmers thinking I should need to install python to make their shitty programs work. The devs are like "I don't use Windows and wouldn't even know how to package it as an exe and even then, I wouldn't even be able to test that it works, python just works for everyone" and then the guy just continues to go off on a similar rant to yours about how ridiculous this is.

Besides, even if it had a GUI, the complaint would be "why don't they make their software do the things that people want it to do‽‽‽" and then going off about how easy this {insanely complicated thing} would be to add in and EVERYONE {nope, just you} needs this function in order for this software to be useful and the dev is just a POS for not adding it.

Like, I get your point if you're paying for software that doesn't make it clear it doesn't have a GUI, but if you're just on github using someone elses software that they made for themselves and thought other people might like, why are you bitching? Just move on to another piece of software that can do what you want and has a GUI (I'm going to guess you can't/won't because those cost money and you wouldn't be able to belittle and harass the actual software creators and that's half the fun for you).

[–] [email protected] 3 points 6 months ago

Not all end-users are the same though. Clearly, you are not in the target user group for this software. If a GUI was that sought after, someone could create that as a standalone project. Upvoted for the genuine unpopular opinion

[–] [email protected] 3 points 6 months ago

GUI components aren't (usually) interoperable, with other programs, they aren't flexible, and they aren't scriptable at scale. They are good for new people and laypeople, but they are hard to write (lots of requirements for input validation) and you have to prescribe a specific workflow. I would argue after being on both ends of the discussion that the one guy you DEFINITELY don't want to write a UI for a complex scientific program is the guy who wrote the program: it means to use it he expects you to think like him.

I would suggest the other perspective: any academic or research tool that DOESNT have CLI parity available alongside a GUI (if it has one) shouldn't be taken seriously.

[–] [email protected] 3 points 6 months ago (1 children)

I'm a senior software dev and have built a shitload of tools and scripts that are command line only. I've been told that I'm a really great developer, and I have a knack for figuring things out, even when there's no documentation.

For the life of me, I absolutely cannot wrap my head around frontend shit. Like, I can do markdown just fine. HTML? CSS? Any of the popular frontend UI frameworks? I just seem to have some sort of conceptual and motivational block that prevents me from figuring out how to make those accursed fucking frontends.

It's such a different form of software development. I wish I could wrap my head around it, but I've accept that I'm not ever going to be a frontend dev. I just write server side stuff and scripts. I apologize if you ever have to use my software :(

[–] [email protected] -3 points 6 months ago

I appreciate your perspective. I'll also fire off a question at you, which I've also posed to some of the other participants in this discussion:

Why is all of this so different, when you shift the frame to game development? Modern game engines are built to allow EXTREMELY fast implementation of realtime GUI elements. When you go to study gamedev, you're tying GUI elements to backend code in literally the second hour of the coursework.

So why aren't game engines used for productivity and specialized-use-case software?

[–] [email protected] 10 points 6 months ago (1 children)

Congrats on the true unpopular opinion lol.

I'll add one thing that I don't think has been said, adding gui adds a lot to the dependencies and complexity of maintaining a program. Keeping it in the cli keeps the size of the program down, and ensures it will work longer and on more devices since it's dependent on less libraries that could change.

Because of that, if a cli program gets a gui it's usually a separate project, like how yt-dlp has front-ends like this.

[–] [email protected] -2 points 6 months ago* (last edited 6 months ago) (2 children)

I asked another person this, but I'll ask you, too:

Why is all of this SO INCREDIBLY DIFFERENT when it's a game being developed, instead of a program that does non-game stuff?

When you go to learn game development, every student is being taught to use an engine to manipulate graphical elements in, like, the second hour of study. Why not use engines (even the same game engines) for non-game applications?

Is there some real reason not to do this, especially for small projects? Am I missing something fundamental?

[–] [email protected] 9 points 6 months ago (1 children)

Front end and back end development are two very different things. I can write extremely complex software, but I couldn't design or code a GUI to save my life.

They're two very different skill sets.

[–] [email protected] 0 points 6 months ago (1 children)

I suppose there really does need to be a renaissance of people trying to contribute on the frontend side.

I do feel some real motivation to get my shit together, learn Godot, and start shoving small open-source projects into it. I could make a difference.

[–] [email protected] 3 points 6 months ago (1 children)

Exactly, be the change you want to see in the world. You've identified a problem that you feel strongly about, just like the developers of the esoteric software. If GUI development is something you can get into, go for it (plus, good front end devs make a lot of money).

[–] [email protected] 1 points 6 months ago

Agreed.

I am still curious, as to why this hasn't been a solution, already.

For that matter, why isn't there any kind of industry standard "non game engine" that caters to productivity, academic, and specialized applications, in the way I'm describing?

[–] [email protected] 2 points 6 months ago (1 children)

Games are already doing a lot of complex things, and pretty much all games are inherited graphical (other than text adventures I guess).

And a game engine is useful since it helps you do a lot, but it's also a huge dependency with a lot of moving parts. This can be a problem even for games, but games do so many things that having an engine often outweighs the downsides.

Using ytdlp as an example again, that just has to connect to the Internet, and write to a file. Using a game engine for that would be major overkill and make the project way bigger then it has to be, as well as increasing the barrier to entry for developers who want to contribute. Python can do all those things but is lighter, smaller, more flexible, and has a much larger userbase.

[–] [email protected] -2 points 6 months ago

Using a game engine for that would be major overkill and make the project way bigger then it has to be

But what if that makes the difference between users being able to actually use the software, versus...just literally not using it?

Keep in mind that I'm talking about end-user software, from the very start of the conversation. Not backend-only stuff, not tools that are only ever going to be used for batch process stuff. I'm talking about stuff like my example of a simulation program for glass and ceramic casting.

Wouldn't it make absolutely perfect sense to use a game engine for something like that? The user won't give a fuck if it takes 2 gigs of disk space, and has a loading screen on startup. They are going to be spending real hours of time, using the application to generate actionable information.

The Python solution that would be lighter, smaller, and more flexible isn't any good if nobody actually implements it, right? And, from where I'm sitting, it seems like the version that uses Godot or Unity would be so much more actionable.

[–] [email protected] 4 points 6 months ago (1 children)

I know. Its like these foundation guys. Learn to build the whole fucking house if your serious about being in construction!!!

[–] [email protected] -3 points 6 months ago (1 children)

Rebuttal: "I know I said I could build you a house, but walls are hard. Like, you don't even understand how complicated modern walls are. It's not like walls were, 20 years ago. You're gonna have to suck it up, with this carport structure, with just the roof and the floor."

[–] [email protected] 6 points 6 months ago

But nobody said they’d build a house. They said they’d setup the plumbing and then you expected the walls all on your own.

load more comments