Just finished college, so I'm not too long in the tooth. I remember reading your writeup on the development of Galak-Z, and liking it so much that I bought it. I'm pretty sure it was one of the first steam games I ever purchased. I didn't end up liking the game all that much, but I definitely continued to enjoy your writing!
Unlike the Hot Dog Stand theme, the "Plasma Power Saver" theme also featured in the article actually was function over form, not just an aesthetic choice (or lack thereof). It was to reduce burn-in on the plasma displays of old portable computers, e.g. here [0].
In practice lot of applications hard-coded some elements' color while following the theme for other elements, making dark theme unusable because you end up with black text imposed by the developers over the black background you choose for your theme, and other similar issues.
As much as we love to hate on Apple's more user-hostile policies, it's only due to Apple's fiat that we've been able to claw back the smallest bit of user theme control --- light and dark mode --- from the "don't theme my app" people.
People can self sabotage by choosing a bad theme, and then they engage with the app less or even churn. Designers need to be careful to not give people rope for them to hang themselves with.
> Designers need to be careful to not give people rope for them to hang themselves with.
No, they don't. It's my system, and the look should be what I want it to be, period. What designers actually need to do is learn to respect their users, even when they disagree with the user's choices.
My teenager has their cell phone keyboard configured so all the symbols are replaced with cartoon cats. They can't type properly on it at all - I get text messages that are completely garbled - but they love that they can do this even though it actively impairs their functionality.
Some people like having ridiculously long fake nails that make it difficult to do their jobs (i'm thinking some checkout clerks I've seen who can't properly push any of the keys on their terminal), but it's their choice.
Certainly that's a good reason to force a legible version of settings, and the path to settings...
But if the user sets the system to hot dog stand, the apps should be hot dog stand. If the user wants the system text font to be wingdings, they're in for a nasty time, but that doesn't mean an app should force a different font
The issue with this thinking is that it's easier for people to quit using the product than to figure out how to fix the font. You can't beat the simplicity of doing nothing, so you need to avoid getting into this state in the first place.
Gotta keep users engaged in your app, right? Keep them onboard even if that means removing all their choices. I mean, should we even allow users to uninstall apps?
After all, the developer always knows best and all users are helpless children who need to be forced to conform and comply. Who cares what the user thinks or wants so long as we keep that sweet, sweet engagement.
If your users are not engaging with your app, you can't deliver user value to them. If you are unable to provide value to their lives because they happened to accidently changed a font that is an unfortunate circumstance where the user is losing out on value they could have had.
It's not that users are helpless, but that they just don't want to spend their time dealing with stuff they don't want to. Users like it when things "just work."
There is nothing wrong with providing sensible defaults and a good collection of pre-designed profiles to choose from (and yes, even a big, friendly RESET button).
But that doesn't explain taking away options. Users who don't want their time with this stuff would probably not use the customization options in the first place.
Also, the term "deliver value" has been badly tainted after too many companies have used it as an euphemism for "extracting value".
It's the same non-logic that advertisers use: Ads are both a service for the viewer, informing them of amazing opportunities, but also somehow the viewers must be forced into consuming that service.
I'm deeply skeptical of situations where people have to be forced into something "for their own good".
Targeting users who enjoy debugging and troubleshooting software is not the way to develop high quality software. You shouldn't be purposefully adding bugs or corrupting installations just to give people problems to figure out.
Nobody said anything about "purposefully adding bugs or corrupting installations".
What you're advocating is protecting the user from themselves, which is antithetic to the entire ethos of the hacker.
Yes, you as a developer shouldn't fuck up the user's setup. But if the user fucks it up on their own? Then that's on them. Don't limit the power users because you want to infantilize the casual users.
> and then they engage with the app less or even churn
I wasn't aware engagement maximisation is the reason we don't get customization options anymore, but it makes perfect sense.
No one used to care about this because it was at the discretion of the user whether they want to keep using the app or not. Whereas today, it's the company objective to keep the user in the app as much as possible.
I remember the miniform of media players. Because while you need the full interface while managing playlists, you only need a few button and some status when you’re playing music. I don’t think Spotify will ever implement that feature.
The colors is part of the content of the product. And it's not that it is 100% more likely for someone to leave sooner, but that it increases the probability that people leave sooner.
That’s clearly bullshit because
if the user sets a system wide theme and your appLICATION follows that theme, then your appLICATION is not going to be any harder to use than the system itself nor any other appLICATION using native widgets.
What is actually happening is designers are forcing non-native controls, in part because web technologies have infested every corner of software development these days. Unsurprisingly, those non-native widgets break in a plethora of ways when the system diverges even marginally from the OS defaults.
And instead of those designers admitting that they fucked up, they instead double down on their contempt for their users.
Also, can we please not call desktop applications “apps” in response to an article about an OS that predates smartphones by several decades.
This sounds much like a post hoc justification for having not bothered to go to the effort to implement the ability to allow anyone (power users or otherwise) the freedom to customize the "app" to their liking.
I'm amused she said they included it "in case somebody out there liked ugly bright red and yellow" and that "the 'Fluorescent' theme was also pretty ugly, but it didn't have a catchy name, so I've never heard anything about it."
I haven't really used Windows for anything serious in more than 20 years (and I recently had to mess with Windows 11 and it was terrible), but I'm not sure you'd be very happy going back to Windows 3.1.
It was a 16 bit system (it could run in "Enhanced Mode" which involves 32 bit protected mode, but in reality Windows itself, and the applications, were still 16 bit).
That means the resource constraints were very real. Even if you had a lot of actual memory in your machine, the memory that was actually available for "general purpose" was effectively a few hundred kilobytes. There was also the notion of finite (and very generically-named) "system resources", and you could see in the "About" box how many percent of those you had free. Once they were gone, you were in trouble:
USER.EXE and GDI.EXE each have a data segment (that is, heap) limited to 64K. The 8086/80286 platform architecture imposes this 64K limit. Program Manager checks the percentage of free heap space for both USER.EXE and GDI.EXE. It then reports the smaller of the two percentages.[1]
All applications ran in the same address space. A broken application meant a total crash at best, subtle data corruption at worst. Multitasking was also cooperative, so apps could hold up other apps indefinitely, or just hang the entire system.
Since it was not based on paging, to accommodate the very limited memory, entire segments could be swapped out, or even relocated, within the address space. As a programmer, that meant dealing with stuff like "locking pointers" so Windows wouldn't move your data segment under you. As a user, that could mean general slowness.
It was firmly based on DOS. So many problems that you had in DOS, drivers or whatnot, would exist in Windows as well.
There were better systems at the time that you could wish yourself back to, some number of them based on UNIX in some way or other.
But Windows 3.11 had really pretty icons. The prettiest, in my mind.
I agree, and Calmira LFN 3.3 and Microsoft Office 4.3
Would Notepad++ work?
What would you use as a graphical www browser?
I mean even with win32s and modern ssl support somehow built-in it'd be challenge.
My sway setup is everything as all black as I can get but with any accents as small and bright - neon green and eye bleeding magenta - as possible. So Fluorescent speaks to me.
I remember as a kid using 3.11 and win 95 and cycling through the themes, trying them all out for a day or two to decide which I wanted to use. You know, important decisions. Anyway, in an eternal black mark on my character I didn't even consider Hot Dog Stand.
It's actually pretty boring. When I say "accent color" I mean a single pixel border around the selected container. The waybar is text, and the text is all bright green on a black background. The active desktop has a single pixel magenta stroke around it. I've thought about turning that into just magenta text as well. Every window element I can make #000000 black without making things more confusing is.
Default text in the terminal is green, and if I select it with a mouse it's magenta. It's more of a "terminal" vibe than the win 3.1 Fluorescent vibe. I said that because they share garish colors.
Also, I'm always on the lookout for even more minimalist graphics to use in my config, if anyone has hyper-minimal things they like about theirs...
I’m using cwm with no bar. Just xclock and xbatt (an icon that reflect the battery percentage) in a corner. The only decoration is a border around the windows. Keyboard driven, but some menus are accessible by clicking on the desktop.
A non-obvious reason that I think the yellow background would've looked especially bad to people at the time, is that most people doing non-gaming on PCs at the time were using MS-DOS programs in text mode, with 4&3-bit color, where it was very unusual for the background color to be bright.
(It was technically possible to get a bright background color on PCs in text mode, but very few programs did that.)
As a kid I actually used the theme on my first hand-me-down. As the regular theme. And I remember chosing it because it was fun! It made my computer more fun.
At the end of the day, usability shouldn't trump fun. If I find it's less usable, I can switch back.
I actually used Hot Dog Stand as the inspiration for the color combination on a big internal website years ago. A "committee" was still flapping its gums about the color choice a week before a hard launch, so I simply... decided. A moderately-unpleasant but distinctive combination went into production. I figured that would finally force them to make an actual decision after nine months of meetings. I was wrong. Users seemed to like it OK, nobody complained, so it stayed.
For almost five years.
I actually finally TOLD them, "you never actually decided, so I picked the colors to be deliberately a little obnoxious so you would actually get off the pot and decide."
They were PISSED.
Not long after, I came up with a way for users (not committees of the managers of users, who usually know nothing) to choose their own preferred colors, and over 80% of them never used that feature and left the garish original I had pulled out of my butt in 1999, because now they were used to it.
Thanks for sharing! I was pleased I was able to track down Virginia and she had a clear memory of how Hot Dog Stand came to be.
Thanks for an awesome writeup! I've been reading your articles on PCGamer since highschool, and they've always been my favorite of the bunch!
Aw shucks that's nice to hear. Do not tell me how old you are because I'm not prepared for that sort of mortality introspection this afternoon.
Just finished college, so I'm not too long in the tooth. I remember reading your writeup on the development of Galak-Z, and liking it so much that I bought it. I'm pretty sure it was one of the first steam games I ever purchased. I didn't end up liking the game all that much, but I definitely continued to enjoy your writing!
I'm so glad that there's room for this kind of investigative journalism in this day and age. Kudos!
It was a fun read, cheers
Unlike the Hot Dog Stand theme, the "Plasma Power Saver" theme also featured in the article actually was function over form, not just an aesthetic choice (or lack thereof). It was to reduce burn-in on the plasma displays of old portable computers, e.g. here [0].
[0] https://retro.swarm.cz/20170331/windows-31-running-on-ibm-ps...
Back when users picked the UI colors for the apps instead of the apps picking for the users.
In practice lot of applications hard-coded some elements' color while following the theme for other elements, making dark theme unusable because you end up with black text imposed by the developers over the black background you choose for your theme, and other similar issues.
Still a problem today, with a lot of terminal programs assuming the background will be dark!
As much as we love to hate on Apple's more user-hostile policies, it's only due to Apple's fiat that we've been able to claw back the smallest bit of user theme control --- light and dark mode --- from the "don't theme my app" people.
People can self sabotage by choosing a bad theme, and then they engage with the app less or even churn. Designers need to be careful to not give people rope for them to hang themselves with.
> Designers need to be careful to not give people rope for them to hang themselves with.
No, they don't. It's my system, and the look should be what I want it to be, period. What designers actually need to do is learn to respect their users, even when they disagree with the user's choices.
My teenager has their cell phone keyboard configured so all the symbols are replaced with cartoon cats. They can't type properly on it at all - I get text messages that are completely garbled - but they love that they can do this even though it actively impairs their functionality.
Some people like having ridiculously long fake nails that make it difficult to do their jobs (i'm thinking some checkout clerks I've seen who can't properly push any of the keys on their terminal), but it's their choice.
Certainly that's a good reason to force a legible version of settings, and the path to settings...
But if the user sets the system to hot dog stand, the apps should be hot dog stand. If the user wants the system text font to be wingdings, they're in for a nasty time, but that doesn't mean an app should force a different font
The issue with this thinking is that it's easier for people to quit using the product than to figure out how to fix the font. You can't beat the simplicity of doing nothing, so you need to avoid getting into this state in the first place.
Gotta keep users engaged in your app, right? Keep them onboard even if that means removing all their choices. I mean, should we even allow users to uninstall apps?
After all, the developer always knows best and all users are helpless children who need to be forced to conform and comply. Who cares what the user thinks or wants so long as we keep that sweet, sweet engagement.
If your users are not engaging with your app, you can't deliver user value to them. If you are unable to provide value to their lives because they happened to accidently changed a font that is an unfortunate circumstance where the user is losing out on value they could have had.
It's not that users are helpless, but that they just don't want to spend their time dealing with stuff they don't want to. Users like it when things "just work."
There is nothing wrong with providing sensible defaults and a good collection of pre-designed profiles to choose from (and yes, even a big, friendly RESET button).
But that doesn't explain taking away options. Users who don't want their time with this stuff would probably not use the customization options in the first place.
Also, the term "deliver value" has been badly tainted after too many companies have used it as an euphemism for "extracting value".
It's the same non-logic that advertisers use: Ads are both a service for the viewer, informing them of amazing opportunities, but also somehow the viewers must be forced into consuming that service.
I'm deeply skeptical of situations where people have to be forced into something "for their own good".
Users who want things to "just work" aren't the entire target audience of software, and it's a huge misstep to act as if they are.
Targeting users who enjoy debugging and troubleshooting software is not the way to develop high quality software. You shouldn't be purposefully adding bugs or corrupting installations just to give people problems to figure out.
Nobody said anything about "purposefully adding bugs or corrupting installations".
What you're advocating is protecting the user from themselves, which is antithetic to the entire ethos of the hacker.
Yes, you as a developer shouldn't fuck up the user's setup. But if the user fucks it up on their own? Then that's on them. Don't limit the power users because you want to infantilize the casual users.
> and then they engage with the app less or even churn
I wasn't aware engagement maximisation is the reason we don't get customization options anymore, but it makes perfect sense.
No one used to care about this because it was at the discretion of the user whether they want to keep using the app or not. Whereas today, it's the company objective to keep the user in the app as much as possible.
I remember the miniform of media players. Because while you need the full interface while managing playlists, you only need a few button and some status when you’re playing music. I don’t think Spotify will ever implement that feature.
Funnily enough, Spotify did recently introduce a miniplayer [0]!
[0] https://community.spotify.com/t5/Community-Blog/Introducing-...
My bad, but I would never been aware of it if not for your comment.
and then they engage with the app less or even churn.
If your content is so poor that a change of colors can make people leave, then perhaps your content is not worth having.
The colors is part of the content of the product. And it's not that it is 100% more likely for someone to leave sooner, but that it increases the probability that people leave sooner.
It really isn’t, unless it’s a picture or a movie.
> Designers need to be careful to not give people rope for them to hang themselves with.
See Win 95 resolution change workflow.
This was 20 years ago. A lot of knowledge was lost since then.
That’s clearly bullshit because if the user sets a system wide theme and your appLICATION follows that theme, then your appLICATION is not going to be any harder to use than the system itself nor any other appLICATION using native widgets.
What is actually happening is designers are forcing non-native controls, in part because web technologies have infested every corner of software development these days. Unsurprisingly, those non-native widgets break in a plethora of ways when the system diverges even marginally from the OS defaults.
And instead of those designers admitting that they fucked up, they instead double down on their contempt for their users.
Also, can we please not call desktop applications “apps” in response to an article about an OS that predates smartphones by several decades.
People who are more likely to customize their app are more likely power users, therefore they're going to engage with the app more anyways.
Why would someone changing app colors to ones they specifically chose make them use the app less? There is no logic in that statement.
This sounds much like a post hoc justification for having not bothered to go to the effort to implement the ability to allow anyone (power users or otherwise) the freedom to customize the "app" to their liking.
I don’t “engage” with products that infantilise me and won’t give me the rope to hang myself with, I endure them, and only to the extent I have to.
It's not about infantilizarion. It's about delivering a product that consistently offers a high quality user experience to the user.
When that user experience clashes with a sufficiently knowledgeable user, then the user experience is the problem.
I'm amused she said they included it "in case somebody out there liked ugly bright red and yellow" and that "the 'Fluorescent' theme was also pretty ugly, but it didn't have a catchy name, so I've never heard anything about it."
Because I loved the Fluorescent theme.
Flourescent looks so of it's time, it fits right into an age of acid house, where you could go hang out in the hologram shop [1].
[1] - E Is for Ecstacy - BBC Everyman Documentary https://youtu.be/jyrhcjRc3TU?si=Qn9qG2z8wQzD-llJ&t=812
I remember playing with these or similar back then, and if you used it for awhile to just became normal.
This made me search and find this, screenshots of every win 3.11 theme:
https://imgur.com/gallery/every-windows-3-1-theme-SsVYqM1
at least half were painfully ugly
Still more readable than what Apple has released lately.
I'd rather use 3.1 with the hot dog scheme vs. windows 11...
as long as it has trumpet winsock
I haven't really used Windows for anything serious in more than 20 years (and I recently had to mess with Windows 11 and it was terrible), but I'm not sure you'd be very happy going back to Windows 3.1.
It was a 16 bit system (it could run in "Enhanced Mode" which involves 32 bit protected mode, but in reality Windows itself, and the applications, were still 16 bit).
That means the resource constraints were very real. Even if you had a lot of actual memory in your machine, the memory that was actually available for "general purpose" was effectively a few hundred kilobytes. There was also the notion of finite (and very generically-named) "system resources", and you could see in the "About" box how many percent of those you had free. Once they were gone, you were in trouble:
All applications ran in the same address space. A broken application meant a total crash at best, subtle data corruption at worst. Multitasking was also cooperative, so apps could hold up other apps indefinitely, or just hang the entire system.Since it was not based on paging, to accommodate the very limited memory, entire segments could be swapped out, or even relocated, within the address space. As a programmer, that meant dealing with stuff like "locking pointers" so Windows wouldn't move your data segment under you. As a user, that could mean general slowness.
It was firmly based on DOS. So many problems that you had in DOS, drivers or whatnot, would exist in Windows as well.
There were better systems at the time that you could wish yourself back to, some number of them based on UNIX in some way or other.
But Windows 3.11 had really pretty icons. The prettiest, in my mind.
[1] https://ftp.zx.net.nz/pub/Patches/ftp.microsoft.com/MISC/KB/...
and yet... still better than Windows 11
I agree, and Calmira LFN 3.3 and Microsoft Office 4.3 Would Notepad++ work? What would you use as a graphical www browser? I mean even with win32s and modern ssl support somehow built-in it'd be challenge.
Opera 3.62 seems to be the latest browser available for Win 3.x, though I think Dillo might also be a possibility:
https://computernewb.com/wiki/How_to_browse_the_web_on_very_...
Legend.
My sway setup is everything as all black as I can get but with any accents as small and bright - neon green and eye bleeding magenta - as possible. So Fluorescent speaks to me.
I remember as a kid using 3.11 and win 95 and cycling through the themes, trying them all out for a day or two to decide which I wanted to use. You know, important decisions. Anyway, in an eternal black mark on my character I didn't even consider Hot Dog Stand.
Do you have a screenshot? That sounds unusably terrible to me, but whatever floats your boat.
It's actually pretty boring. When I say "accent color" I mean a single pixel border around the selected container. The waybar is text, and the text is all bright green on a black background. The active desktop has a single pixel magenta stroke around it. I've thought about turning that into just magenta text as well. Every window element I can make #000000 black without making things more confusing is.
Default text in the terminal is green, and if I select it with a mouse it's magenta. It's more of a "terminal" vibe than the win 3.1 Fluorescent vibe. I said that because they share garish colors.
Also, I'm always on the lookout for even more minimalist graphics to use in my config, if anyone has hyper-minimal things they like about theirs...
I’m using cwm with no bar. Just xclock and xbatt (an icon that reflect the battery percentage) in a corner. The only decoration is a border around the windows. Keyboard driven, but some menus are accessible by clicking on the desktop.
A non-obvious reason that I think the yellow background would've looked especially bad to people at the time, is that most people doing non-gaming on PCs at the time were using MS-DOS programs in text mode, with 4&3-bit color, where it was very unusual for the background color to be bright.
(It was technically possible to get a bright background color on PCs in text mode, but very few programs did that.)
As a kid I actually used the theme on my first hand-me-down. As the regular theme. And I remember chosing it because it was fun! It made my computer more fun.
At the end of the day, usability shouldn't trump fun. If I find it's less usable, I can switch back.
Back in my day, changing the windows color pallets was the only form of video game we had.
And we liked it!
This brings up so many memories
We've now seemed to have invented glass, but in some kind of a parodic form.
I actually used Hot Dog Stand as the inspiration for the color combination on a big internal website years ago. A "committee" was still flapping its gums about the color choice a week before a hard launch, so I simply... decided. A moderately-unpleasant but distinctive combination went into production. I figured that would finally force them to make an actual decision after nine months of meetings. I was wrong. Users seemed to like it OK, nobody complained, so it stayed.
For almost five years.
I actually finally TOLD them, "you never actually decided, so I picked the colors to be deliberately a little obnoxious so you would actually get off the pot and decide."
They were PISSED.
Not long after, I came up with a way for users (not committees of the managers of users, who usually know nothing) to choose their own preferred colors, and over 80% of them never used that feature and left the garish original I had pulled out of my butt in 1999, because now they were used to it.