Ask HN: My family business runs on a 1993-era text-based-UI (TUI). Anybody else?
No synthesized answer yet. Check the discussion below.
I wonder if that's why Target does not support making sales tax-free purchases (for resellers, nonprofits, etc.) online? Amazon and Walmart do.
Moving to a web based system meant we all had to use mice and spend our days moving them to the correct button on the page all the time. It added hours and hours to the processing.
Bring back the TUI!
I've never once seen an experienced user equal or gain efficiency when switching. It's always a loss even after months of acclimation.
The management needs to pick the right concept though, not the one with pretty and playful screenshots, but the one that focuses on the right KPIs (the 20 most common user flows need to take less than x seconds for an average user).
When I was in college (many years ago) the company I worked for used a TUI for its inventory/back office systems (terminal emulator talking to an AS/400) and once you understood the hierarchical structure and how it worked you could fly through that system because it was all keyboard nav.
Few GUI's have ever been that fast for me even the ones that go out of their way to make everything accessible via the GUI bindable.
They shouldn't start from a few pretty figma sketches and then try to make them more usable. They should start from user flows, solve how the users can do certain things with maximum productivity, easy navigation, showing the right data together on the same screen, and so on. Only in the end make it pretty.
It may be pretty but it makes me puke.
The "modern UX design" is based on layers upon layers of abstraction and "cheap" interfaces. Which one is the button ? Where is the scrollbar ?
I've literally done the before and after on this a handful of times and it's always worse off. Management will never do that, it's always design by committee, the KPIs won't be defined or will never really have teeth, it will turn into someone's vanity project, they won't even pay someone to optimize the code - quite the opposite, they'll choose to build it on something like Salesforce or some other very non-performant enterprise-y platform, etc, etc. All the TUI get these performance gains out of the box without much additional effort. The constraints of the UI are it's strength as it prevents people from adding all this bloat in the first place. When you leave it up to people, especially business users or UX folks, it will get spoiled. It's almost a law.
This is not true. Smart companies do it exactly that way. It saves them a lot of money.
Do you have concrete examples in mind that we can review?
Yeah, I guess I could say that before I tried rebinding ctrl-w and some of the Fx keys (like F12).
While using very common web development stacks a lot of developers know how to deal with.
I think the fancy new terminals also allow you to change the font.
Look at the era of teletext. Not exactly multiple typefaces, but you could express wide, compressed, italic, bold and all colors.
Sending http response, waiting for reply. Http 200 ok
and so on and so forth. Web sucks. Of course, you can have something like Jira, bur still sucks.
But your snarky reply is probably also wrong anyway, since I would be surprised if there is no business software made using game engines. Tesla is known to have used (but not anymore?) Godot for making GUIs for their in-car displays for instance. I have bought applications (not games) made using both Godot and Unity. Not saying it would be a great idea, in general, to use those engines for business applications, but people are not known for always using the best possible tools, are they, and it certainly would work in theory?
Not very familiar with Unity, but Godot has very nice GUI widgets. Same style of WYSIWYG editor like old Visual Basic or Delphi really. You drag widgets to place them on the screen, set properties, add scripts to react to different events. The entire Godot IDE itself, a very non-trivial application, is implemented using Godot's own GUI framework.
https://docs.godotengine.org/en/stable/tutorials/ui/index.ht...
I'm gonna go tell the building manager that people would like our app better and we can save on platform costs by hiring one Unity developer.
It’s perfectly possible to handle large amounts of data by copy and paste on a web browser, you just have to actually support it.
When IT was trying to sell everyone on WIMP GUI, the standard was WordPerfect. There, you had to memorize the functions of the F1-F12 keys, as well as their shifted (or Control) behavior. The only sure thing was that F1 was HELP.
We were looking for something better--but having no keyboard was the opposite extreme.
Most GUI's make you wait for the form to appear before you can type into it. That totally destroys the flow of operators.
There are GUI's that are properly designed to be keyboard driven and to allow type-ahead. Those can be truly best-of-both-worlds. Too bad they're so rare.
FWIW, dBASE IV version 1.5 does support Japanese for date format. It's one of the options for 'SET DATE' command.
Modern terminals, even the text-based console on a fresh minimal Arch linux install, is going to support Unicode probably without a ton of issues as long as you have a font installed that has the characters. The struggle then for RTL or non-Latin character sets and languages is making the UI fit the words in a way that makes sense.
Now they get promoted for changing things to appease MBAs and giving their boss something to talk about to their boss.
America has fallen, and we can't get back up.
Also no scrolling was a requirement. This was done by defining a min and max screen resolution, and designing everything exactly for that. The app was supposed to be used exclusively full screen, so no need for responsive design.
The result was a bit like a video game, very few loading delays and instant responses to user input.
Most people are running on 90s-2000s era stuff rather than TUIs.
For the most part, it works well, and is not very costly.
Check out Sage100... flexible, cheap, on prem... runs everything from job / work tickets to inventory, purchasing, financials, payroll, etc.
Aint sexy but it works!
Modernizing will roll some of that back; I would only consider it if there’s a plan to be around for the years it will take to get good again.
Like the fantasy is that the bank uses TUIs and the bank has accumulated years of knowledge and the bank doesn't make mistakes. The bank has extremely well paid staff. Joe Shmoe's TUI app looks like the bank's app, but it is unmaintained, it has accumulated years of problems, not fixes, nobody is fixing them, people who say they fix them cannot possibly be keeping up with the sheer amount of toil and bugs needed for production software. You can see this in any GitHub project, how much insane maintenance is required, for stuff people actually use and has few bugs.
When I worked ar Sherwin Williams, I got good enough with the TUI that customers could rattle off their orders while I punch it into the computer in real time.
It's absolutely crazy that a well designed TUI is so much faster. It turns out that if you never change the UI and every menu item always has the same hotkey, navigating the software becomes muscle memory and your speed is only limited by how fast you can physically push the buttons.
The program had many menu options added and removed over the decades, but the crucial part is that the hotkeys and menu indexes never, ever changed. Once you learn that you can pop into a quick order menu with this specific sequence of five keys, you just automatically open the right menu the moment a customer walks up. No thought, just pure reflex.
UX absolutely peaked with TUIs several decades ago. No graphical interface I've ever seen comes even close to the raw utility and speed of these finely tuned TUIs. There is a very, very good reason that the oldest and wealthiest retail businesses still use this ancient software. It works, and it's staggeringly effective, and any conceivable replacement will only be worse. There simply is no effective way to improve it.
Edit: I will say that these systems take time and effort to learn. You have to commit these UI paths to memory, which isn't too hard, but in order to be maximally effective, you also have to memorize a lot of product metadata. But the key is that it really doesn't take longer than your ordinary training period to become minimally effective. After that, you just pick up the muscle memory as you go. It's pretty analogous to learning touch typing without trying. Your hands just learn where the keys are and after enough time your brain translates words into keystrokes without active thought.
It's a beautiful way to design maximally effective software. We've really lost something very important with the shift to GUI and the shunning of text mode.
I've probably made that sort of UI without realizing it...
The UX is optimizing for accuracy over ease of use, and in this case is likely intentionally difficult to use.
Also I have no mental imagery of summer camp with networking much less internet. I can't comprehend dropping my kids off at a retail storefront or a church as "summer camp".
I remember my high school went big on Gradebusters software--text entry on the Apple IIe (80 column required!) of course that was all keyboard driven.
Tab tab down space down space down down down space.. that was taking attendance.
One problem with GUI is that pointer-warping is unnerving, we don't have facilities like "I clicked, now warp the pointer to the next target" but that's commonplace with text UI.
There's no TAB button on a mouse.
They are. So. Slow. Each and every one of them. Its awful
Typically the first two weeks of training revolved around new hires asking why in the world we used this system before their spirits broke and they reluctantly plunged into the deep end...kind of like being released into the matrix.
We also have a small finance team (typically around 2 employees), and finding somebody with a finance/billing background who is willing to work with TUI on Linux... that was a challenge :-)
For example, enabling a fast multi-select of rows in a longish table (or even worse, a tree) is one of the tasks that TUIs don't really excel at. Popping up a PDF or image viewer would also be great.
The TUI I'm working with runs on a pair of Linux VMs, and is accessed from Windows, Linux and Mac, so asking all our users to enable X forwarding doesn't really work.
> The TUI I'm working with runs on a pair of Linux VMs, and is accessed from Windows, Linux and Mac, so asking all our users to enable X forwarding doesn't really work.
Run another Linux VM on the client. /s
Aren't there X servers for Windows? I remembered installing one on a locked down university computer for fun.
One thing that often gets lost in the discussion of TUIs vs GUIs is that this is also true of GUIs. You have to know which icon to click, and it's not always in the same place, and not always labeled. Increasingly, functionality is hidden behind a hamburger menu, and not laid out in logical sections like File, Edit, View, etc menus.
On the system I use, every menu item was prefixed with a number. You punch in that number on the keyboard and you're in that menu. Just absolutely beautiful functionality
That seems to have slowly vanished over the last 30 years
GUIs can have keyboard shortcuts too. I'm an artist and I work two-handed: right hand moves the stylus around the screen, left hand floats around the keyboard and changes tools, summons control panels, etc. Whenever I try a different program than the one I'm used to, and have to poke at icons with my right hand because I don't know its shortcuts, I feel like half my brain's idling.
Some frequently used functions have their own special single-key shortcuts as well, so instead of having to press C-P (to open the Clip menu and then select Paste) you can just press the ' key, saving the user a key-press every time they do that.
Doesn't work with reassignable keys, either. If I wanna reassign the shortcut for "pencil" to "d" because it's easier to hit with my left hand then that's my own business.
The TUIs are a little funny, because often people can navigate and input faster than the program can render. So you vaguely see a screen change, a menu move or an overlay pop up, only to instantly disappear. The user doesn't need to know what's happening on screen, because they trust their fingers and the system to do the right thing.
Yes, but the it's not due to slow rendering speed. GUIs render even slower.
I learned this while creating some autohotkey scripts: I could have the macro click before the button was shown on the screen.
This also created some problems where users would unintentionally hit their scroll wheel after clicking and change a 1 to a 7 on the next screen before it even appeared. That wasn’t good…
Imagine learning all the keyboard shortcuts for every website you use nowadays.
For example I worked at a video store long ago that had some dos program to manage everything, I didn't own a computer and I didn't use any other software. It was still often a slow turd, and it wasn't networked with the 2 other local stores, so if I wanted to know if a customer had an account there, or if they had some stock there, I had to call.
Variable width fonts and style sheets are nice, though.
Consumer software is all about conning people who don't currently own the software into thinking it's good and buying it. Once they have it, their experience doesn't matter at all. In fact, you should actively make the software less capable - really dumb it down, to appeal to the widest audience. Yes, more white space, yes, more menus. An error message that says 'oopsy something went wrong'? What a splendid idea! Information density does not matter. Speed does not matter. Accuracy of data does not matter. Hell, even bugs don't matter, if your software looks pretty.
Business software is different. It's crufty, it's ugly, and it doesn't change. Because you're optimizing for your users using the same software for 8 hours a day for the rest of their lives.
You just have to make fast navigation and data entry a high priority. The assumptions and approaches people make today mean that generally isn't the case. With a TUI there is a hard limit on the amount of data you can load and that helps as a starting point.
- Often it's hitting a server is that is far away rather in the same building.
- you can put the web server in the same building
- Each page is very heavy with lots of JavaScript etc. loading - you can build lightweight web pages if you priority small script size, minimal dependencies,
and do profiling to keep things in check
- Data loading delays rendering - limit the amount of data loaded and displayed at one time
- Data is slow to load - an older system has limited capacity meaning it's not an option to keep inactive data around
- this means data is more likely to be indexed
- new systems have very large storage capacities, leading to a larger and larger growing db
- eventually default configuration for the database does not allow for everything active to be indexed
- solution is to keep inactive data out of active tables being queried for normal operations
- Slow to navigate with mouse - you can create a keyboard navigation system
- you can build keyboard shortcuts into every screen
- you can create autocomplete shortcut commands to jump to screensIndeed. Most developers today make their own easiness of development the priority, not anything related to the end user's experience. This is why software is so bloated and slow, but the pernicious idea of "developer experience matters more than speed" doesn't seem to be going away any time soon.
I'm going to push back a little on that. For several years, MacOS followed a strong UX convention with consistent keyboard shortcuts, menus, layout order, and more. Similarly, Microsoft started with the same, but with everything reversed. At the time, most major cross-platform apps followed these conventions.
Two periods broke these rules: the expansion of web apps and Apple's pivot towards the consolidation of everything into iOS.
First was the dawn of web apps. Faced with two opposing standards, web apps didn't know which model to follow. Business sites stuck with MS standards, while design-centric sites followed Mac standards. As those broke consistency, cross-platform apps gave up and defined their own standards.
Mobile platforms tried to establish new standards. iOS was mostly successful, but started slipping around iOS 7. Material Design was supposed to standardize Android, but Google used it for all Google products, making it more of a standard for the Google brand than Android.
The second started around WWDC 2019. At that point, Apple deprioritized UX standards to focus on architecture updates. The following year, Catalyst was literally a UX catalyst, introducing two competing UX standards for MacOS. From that point forward, Apple really hasn't had a singular UX standard to follow anymore, but they seem to be marching towards iOS standards for all devices.
I think sublime text was one of the first to bring the TUI style super-powers into the modern desktop UK, where you press some random keyboard shortcut (e.g. cmd-p in sublime) and you can instantly start typing a command.
Another thought I have had for a long time is that when GUIs like Mac and Windows were taking off, they were often described as more "user friendly" than TUIs. I always thought this depended on the kind of user. A lot of effort went into prioritising making it possible for an untrained user to use a system, but making it fast for someone with experience was no longer important.
MITRE successfully conveyed how to have keyboard-operated GUI with plenty of power user modes --all with discoverability. You might call it "fallbacks for every mode".
Its principles combined TUI (then known as CUI) and GUI as well as DSL (domain specific language).
A fully architected program having a DSL (domain specific language) had a console much like a macro recorder: as you used the menus and icons, the console would show the commands you could otherwise type in to get the same effect.
You could edit the console line and re-execute it.
In the 1990s, DSL was a buzzword but if you got it right, your GUI was a power tool.
> I'm going to push back a little on that. For several years, MacOS followed a strong UX convention with consistent keyboard shortcuts, menus, layout order, and more. Similarly, Microsoft started with the same, but with everything reversed. At the time, most major cross-platform apps followed these conventions.
I used Mac OS during that era. While in many ways it was better than the GUIs we have now, using the mouse was an absolute must, which prevented it from ever getting as efficient as a TUI. Yes there were keyboard shortcuts, but they were never sufficient to use the machine or any application without a mouse. Also they had to be completely memorized to be useful.
In the heyday of TUIs, job attrition was much lower. It made sense to create a tool that was extremely hard to use, but also extremely efficient once learned.
In the modern days, attrition is much higher, especially in retail. You need to focus on discoverability and simplifying training as much as possible. Efficiency is secondary at best.
I think what actually happened was technology changed. GUIs became possible, and therefore most designers made GUIs without much if any thought (because many assume newer == better). Now you'd be hard pressed to even build a TUI, since most know only how to build GUIs and all the tools are GUI focused.
> In the modern days, attrition is much higher, especially in retail. You need to focus on discoverability and simplifying training as much as possible. Efficiency is secondary at best.
That sounds like a post-hoc justification. Also higher attrition doesn't just happen like the weather. Deliberate decisions lead to it.
If you’ve seen any of the GUI retail tools that replaced the TUIs, they certainly aren’t designed for discoverability.
There is also trend with "modern" UI/UX to focus near entirety of effort on user's first few minutes and first few hours with a software, while near zero thought is being put on users having to use given piece of software for hours at end, day in, day out
The entire GUI was sold on it being easy to use for beginners. There's a reason why keybindings are today called "shortcuts": Steve Jobs's diktat was that the primary means of giving commands to the Mac was the mouse. Anything you can do on a Mac (except for entering text) was to be done with, and designed around, the mouse. Keyboard "shortcuts", so called as a quicker way to issue (some, but not all) menu commands, were a sop to power users but really were an addition, not essential to use of the program.
Windows, by contrast, was designed to be keyboard-accessible. Not all PCs in 1985 had mice...
And no, we don’t have the glorious keyboards from that time :( just standard 104s
And those users want more batch creation and editing.
Bloomberg Terminal basically. And then because of muscle memory, it's so hard for users to get used to another system. And then they push it onto their juniors. And then you get to charge companies $250 per head to train juniors on how to use the system, with all of its textbased commands.
Granted, that internship led me to where I am now, so I'm not complaining.
The advantage was in typing a command and most of its arguments quickly.
I hope you didn't seriously associate the price $250 with Bloomberg training. Everything Bloomberg has a lot more zeroes after it, except what they pay to new hires in Princeton.
I ROFL'd. They even bumped up our prices from $25k to $36k annually.
"Shhhhh! That's the Terminal!"
I use text-mode every day
Cannot speak for others but I lost nothing. On the contrary I gained faster hardware and faster network
What is sad is that is doesn't have to be that way. But you describe is not an intrisic characteristic of a TUI, but of using keyboard instead of a mouse. You can write a webapp that performs in the same way (modulo the resources needed of course), but it takes extra time and once it works with a mouse there little reasons to put more work for keyboard user (it is not a selling point in most cases).
The main advantage of a TUI is that it forces to write the UI in a certain way, so you get the result automatically.
Oh, I’ve been a new hire at places like this.
The day1 stuff in the system will be logically or coherently placed (e.g. alphabetically) and then the rest will be incrementally added on top.
The menu options will go like this:
1. Alpha
2. Bravo
3. Charlie
4. Turbo-Bravo
5. Charlie-Undo
6. AAAlpha
Great for day1 users that are still there and started with just 3 options but new users are screwed. The arrangement started to stop making sense circa 1995.
I’ve seen the same thing where parts shelves were sorted by manufacturer (to make re-ordering by manufacturer simple).
When they moved to automatic inventory, shelves didn’t matter anymore and if a part was made by manu1 but now manu2, they’d keep putting it with manu1 because “that’s where everyone’s gone to look for it for the part 20 years”.
All you had to know was who made that part in 2003 when auto-inventory stopped forcing them to be consistent by manufacturer to know which shelf to pick from. Easy!
New hires get screwed and over time nothing makes sense anymore.
This works because you can 'buffer inputs' as the gaming crowd says. You can hit the keys and the computer reads them one at a time and does what you asked at its pace. Often these kinds of systems do run faster than the input, but when they don't, it still works.
It's hard to do that with a GUI, you usually can't click (or tap) the button before it shows up and expect it to work... and when it does work like that, it's often undesirable.
I worked for a medium sized company who did work with Toro. We supplied many of their lubricants and they had TUI they still used on one of their machines to enter the orders from our company. It was the last of their legacy products, but worked incredibly well. We had very little issues ever with the system. Our Oracle ERP Net Suite? Had three people dedicated to making sure it ran smoothly. I still remember some of the guys I used to talk to at Toro were "lifers" who were always talking about how easier things were before all the SAAS and ERP software came on the scene.
The stories they had were pretty entertaining.
A few months before I left they switched to a "modern" GUI. It was shockingly bad. The speed of every transaction lowered. Even with optimal use it just took longer. So much time wasted.
Knowing that 1.5.6 sends you to scheduling a pickup, or 11.1 to get into the credit card application, versus hunting through graphical hell with a mouse and a touchscreen.
Wikipedia vandals these days...
At that time their 'web store' just put paid orders in a queue and a room full of humans typed the orders into the green screen which had all the actual inventory.
Probably a big chunk of businesses that developed their core systems before the PC era. I don't know if they still use it, but Avis Rent-a-car's main application used by its front-line people was a TUI like that, and the front desk people could fly around int it (like you said).
But most developers ape current trends rather than actually figuring out what would work best, so I'd guess very few user-facing TUIs are being built now.
https://hackaday.com/2021/10/06/atari-st-still-manages-campg...
I have also met some people who worked at large old insurance companies. They originally used old mainframes and TUI, and the companies still exist. They told me of various things that were done. Of course migrations happened. And interfaces were built so that modern systems could speak with the old, sometimes via terminal emulator. And of course, some old systems still in use far beyond their time.
And if anyone suggests rewriting it, fire them.
149 more comments available on Hacker News