Ross Scott

Ask questions or topics to discuss here for the next videochat with fans on 5:00pm UTC on July 5th at twitch.tv/rossbroadcast. More Game Dungeon and Freeman's Mind (I think) coming! This is a blog post. To read the original post, please click here »

Yeah I actually tried a plugin like that in the past and the results were fairly hideous and it still let some white through, but it's been a while since I've used it also. I can see the potential for a dark theme if you really can clamp down on all colors above a certain brightness level, ideally it's something I should have tested out more, but I only had so much time. Thanks, I'll try to add these as an option, maybe mention them in the followup video. For me it's not the radial menu I want exactly so much as rapid access. I've actually used some radial menus in games and applications before that I HATE (they'll track the mouse movement in a non-intuitive way, where you have to move the mouse left or right like a dial and up and down don't behave the way you think they would. I imagine ones designed for the OS are more intuitive though. That's great hearing that it's increased your workflow speed, that's what this is all about! (well, that and things looking nice) Right on, though I doubt it's a replacement for the whole shell, but rather the start menu, although that's something I was having trouble finding a replacement for that which could be launched anywhere, so thanks a bunch!

No, it doesn't exist, it's not even close. So it's always a tradeoff of what do I hate not having the most, which shifts in my brain. As for trackballs, those are a tease, because a lot of trackballs have an EXCELLENT shape to the frame, but I'm complete garbage with trackballs. I would want a mouse in a trackball frame.

Okay, I think I found the problem. I thought I made a post explaining it in detail, but it looks like that never happened, so I understand why you think I'm sounding irrational and irritable on this. Let me spell it out clearly then, I think it will clear up the confusion (I'm hiding the text so it doesn't flood the chat): No, damn it. BY ALL MEANS propose GUI solutions, that was the whole point! I plan to use your keymouse link in the followup video as it is! I didn't know about that! Just maybe be careful before making accusations, that's all. Try to focus more on new ideas / software, that sort of thing. Getting into a giant debate is unlikely to solve much, whereras I think more brainstorming would be way more constructive considering the state of the GUI today. I WANT people proposing more ideas, screenshots, etc.

My mouse is starting to die on me and that made me frustrated trying to find a new one. I ended up making a graphic of a mouse I'm pretty sure no one besides me wants, but some of you might find some individual parts of it appealing. I wasted time making this, but it felt cathartic. This is a blog post. To read the original post, please click here »

I'm not really looking to get into an argument about this, but you're accusing me of being contradictory, which I don't appreciate unless it's accurate (like what I said on memorization, fair enough). So please, specify with no uncertainty, what did I say regarding targets where I'm being contradictory? I want easy to hit targets with low travel time. Sorry if I wasn't clear here. I'm not seeking to keep my current workflow. I want something BETTER. Moving from 7 to 10 is the catalyst. 10 struck me as shutting down more options, since it's hostile to custom shells, and updates will wreck some 3rd party programs with no consideration for customization. As for being incoherent, I can be guilty on that one. GUIs are very complex, I'm juggling dozens of thoughts on this in my head, so sometimes it doesn't always come out clearly, even though there actually is merit behind the thought. 2 responses to this: 1. Something that has the same functionality as what I have now, but is LESS efficient, I see as a downgrade. The suggestions I saw you mentioning all struck me as accomplishing the same goal, but in ways that take more time. It's more than just whether it can do what I was doing. It's about whether it can do what I need as well or even BETTER than what I had before. If somebody wants a new convertible because their old one has damage, and there is a new convertible available, but it only tops out at 45mph, they're not going to see that as a good answer. 2. As for bragging about having an option a lot earlier, I could have been mixing up my responses to other comments I saw, I forget. Some people have pointed out that me using old software is ludicrous, whereas when I set this up, the options I wanted simply didn't exist by standard means, so I went with what did work, and didn't always keep up with what changed, often because new OS's would wreck options. Part of that is me being behind the curve, but when I see how old software was sometimes ahead in some ways to what's even available NOW, it makes me frustrated with the entire situation. In other words, look at it like this: Old software: -Could do functions A, B, and C New software: -Can do functions, A and B. Does function A even better than the old software, does function B more poorly. Can't even do function C. Listen, you say you're trying to help me, but you're either concern trolling now, or you're unaware of your own bias on this to say something like that. Yes, I'm not an expert, and yes I'm seeking help from others. But insinuating I provided no proof or effort on this topic? Why do think the video was so long? Here's what I did do: -I showed my existing desktop and a few optimizations it had, even though I recognize it's very flawed -I showed time trials how launching frequently used programs was faster in my method than the Windows default -I proposed a pie menu system for rapidly launching software -I recognized the potential of a custom keypad that's ergonomically designed for hotkey deployment -I found research demonstrating that flat themes are objectively slower for users. -I proposed mouse gestures as a way of doing things more rapidly and suggested some fundamentals for coming up with styles -I gave a list of programs that I think are taking us in the right direction for customization as a starting point, suggesting possible implementations -Demonstrated objective advantages for not having a close button in web browsers. -Gave specific changes that should be made to an existing pie menu launcher for maximum efficiency Think what you want, but if you watched my video and then say I'm making the easiest criticism possible and making no effort for trying to advance the GUI, then that's a troll move, whether you mean it that way or not.

My main issue with Rainmeter is that it's all on the bottom z-order wise. So if I have a maximized window, most of those options are gone until I show the desktop again (at least that's my understanding of it). Alternately, if they showed up on top of Windows, I don't know how you would get the GUI elements to hide themselves for full screen applications. I think it will still be good for some things though, but I look at it from the perspective of both how nice it looks and how practical it's going to be for long term usage. It of course depends on the person's preferences however. While that wouldn't work for me personally due to my freakish typing style, this is exactly the kind of outside-the-box thinking I'm trying to encourage. That's awesome as a proof of concept. Some thoughts that come to mind: -While faster, the instantaneous nature of it can be a bit jarring, this could be the sort of situation where extremely fading or morphing might make the transition nicer for the user. That, or removing the wheel background since that throws up a lot of white against the background, whereas just seeing icons pop in and out might be less of a "shock" visually. Alternately, if somebody wanted to go max eye-candy mode, the second layer of icons could start appearing out-of-focus underneath the selection as you got closer to it with the mouse cursor, coming into focus as you fly over the selection (and fading / blurring out the former layer). That would take more work though and I don't even know if it would look better, just a thought I had. -I think the key for something like this would be to have it heavily customizable, with just a suggested template people could modify. People could choose how many selections on the wheel, skin it, have different wheels trigger via different hotkeys or mouse buttons, etc. -For your final destination, I notice if you fly past it, it becomes unselected, I would have it stay selected as long as you're in the corrected direction to allow for maximum sloppiness from the user. -It might be better to just have a button to bring it up, but after you've pressed it, it's no longer needed (or press another button to cancel it) Anyway, don't think I'm trashing it, it's neat seeing it in operation. It looks close, but I couldn't tell from the options if it does quite what I was talking about. Is it possible to trigger the next menu simply by dragging the mouse to it, or does it require you to click on it? That makes a substantial difference speed-wise. I wasn't sure if that's what "Turbo mode" was or not. I can try and test it out when I have more time.

Oh I think I'll be able to spend a lot less than that. It'll be tricky predicting my average though, since I bet a lot of them will get 2-3 seconds, but others could be up to a minute. The short ones I think will outnumber the longer ones. No idea if I'll get it done in one sitting though.

Yeah, this is a good point actually, since I was bringing up 2 different aspects of the GUI, design AND looks. Even if things are just as inefficient as regular windows, if it LOOKS the way you want it to, you're simply going to be happier using a computer longer (or I would be if it was a design I liked). It's definitely a form AND function problem, more power to you if you can nail down the "form" part for yourself.

No, I think you're misunderstanding me. Ideally, I want accuracy to barely matter. Take that pie menu I demonstrated. Say I want the upper left, so I flick the mouse that direction. Maybe I flick it 10% of the screen, maybe I flick it 60%. Ideally, it won't matter. It detects I moved it to the upper left, so it draws a new menu where I am (or returns me to the center automatically). In other words, the GUI adapts to my movement and intentions, not the other way around. I don't see what's contradictory about this. When you press most buttons or a D-PAD in a videogame, does it matter EXACTLY how hard you pressed it, or just that you pressed it? It's a similar concept. This is also why I got excited about mouse gestures. It opens up a wide world of possibilities where accuracy isn't important. I think you see what I said as contradictory, because you're looking at this in terms of hitting a target of a certain size across a certain distance. I'm talking about the potential of targets where distance and size are irrelevant. In other words, Fitt's law doesn't even apply for that kind of manipulation. It transcends it! Seeing how that concept isn't even USED in most modern GUIs is why I think it's woefully underdeveloped. I see that as a huge potential breakthrough. I would compare to the early days of FPSs, where keyboard aiming was dominant. The mouse rapidly displaced that because it was SO MUCH BETTER. I see that kind of potential in where we are now with GUI concepts. In my eyes, it's like we never left keyboard aiming, just made lots of refinements. Now for more conventional GUI menus where that's simply not an option, I want what's fastest and feels best. That's where that Fitt's law thing people are talking about come into play. Tiny targets take longer to hit. Targets far away take longer to hit. You want everything semi-big and close-by, which is probably why I thought of that pie menu system. Truth be told, the Fitt's law thing I was never even thinking of consciously (nor aware of when I made the video), it was totally a subconscious understanding on my part. Yes, I think you get it. This video is essentially: "what are the answers?" "What are better ways of doing things than we have now?" "I can imagine MANY different possibilities, but I haven't seen much information on this to know what's best, hopefully someone out there has all kinds of ideas on this, but hasn't had much of an outlet" and I tried to show a FEW efficiency shortcomings as proof that what we have now is indeed, not the penultimate in GUI design. Instead, a bit many comments seemed focused on me not providing solutions. From my perspective, it's a little insane that's on my shoulders to begin with, but I plan to slowly figure out something for myself in the absence of other ideas. I'll detail what I plan to have / what I think could be better in a followup later on, but it's a little disappointing seeing how narrow the range of ideas have been. "Criticizing a brainstorming session" is exactly how some of the commentary comes across, however, I was inflammatory in the video, so I expected a certain amount of that.

I'm a moron for forgetting about that, my bad. Still increases the travel time and involves tiny targets (even smaller than the popup menu I have in Litestep), but you're right, it exists.

Itch.io had a gigantic game pack for sale a week or so ago, that ended up coming to 1700+ games. The majority of this stuff looked like games I wasn't interested in, but there were some promising ones. I was going to attempt to filter through these on a livestream to see which ones I was interested in, and which I would pass on if people are interested. This will of course be to my own tastes (you can see those in my games list video), but I'll explain my rationale. In any event, some people may want to check the video afterwards at the end and see what I whittled down the list to. I originally thought I would be able to do this in one sitting, but that was back when it was "only" 700 games. I'll just go at it for a few hours and see how many I get through. If you're interested, it will be at twitch.tv/rossbroadcast at 5PM UTC on June 20th. This is a blog post. To read the original post, please click here »

No, it's like I said in the video. Be able to do as much as possible with just one hand, then do even more with 2 (left hand for hotkeys, I even said in the video a custom pad could do a lot). I actually wasn't criticizing hotkeys, just the current implementation of them. It's when you need BOTH hands on the keyboard and / or poor ergonomics that I think they aren't a great solution. In other words, I want to minimize the switching and I personally do too much with the mouse (sometimes where the keyboard would be a poor choice also) to make the keyboard the primary control (I think this is the case for most Windows users). For others, the mouse is only sparingly needed, so the keyboard makes more sense to them. Both camps don't like having to switch hands more than they have to.

My understanding (which was apparently incorrect) was if I'm editing a text file, that's only in system memory. The copy on the disk is still safe. If a power failure occurs, unless it was in the middle of saving and writing to the file, the original copy on the disk is still safe, all I'll lose are any edits I've made in memory since my save. Instead, the ENTIRE file was written blank. Don't worry, I'll probably move to Notepad++ at some point. And of course theme look is subjective, I was hoping to see a variety which might give me ideas of what I'm looking for. It is true, I think yours don't have enough contrast between the background and the text and some of which look poor against a bright windows frame, but I appreciate the input. This is a thought I had after making the video I think has a lot of potential and may or may not happen. I was envisioning something different from Eagle Mode though, I think that goes TOO far with zooming and kind of reminds me more of microfiche, which can get confusing. I appreciate the outside-the-box thinking though. I plan to talk about zooming in the follow-up later on. Yeah, I'm up against decades of conditioning, it might be a fool's errand. But I'm looking for that gap between the cracks to make things better. I'm actually more optimistic now than before I made the video. That's a great analogy. I'd argue it's more than just the lack of standardization, it's kind of the Windows 8 Metro problem as well: The GUI simply can't accomplish all the functions expected of the OS, so of course it commonly gets dismissed. If it was the case where there was no GUI standardization, but every desktop manager did almost EVERY function a typical user could need, I think that would be far less of a problem. You have a lot of excellent points in your whole post, but your example is a great one to look at. First, that would not be fast for me, just because I simply don't deal with non-typical characters very often, like *,$,&, etc. I can touch type, but I realized, for many of those, I still have to look down, but hey, that's an example where I'd be faster if I adapted. Second, I make plenty of typos. The longer the command is, the odds of that happening go up. While they're rapidly fixed with backspace, it's still an average slowdown that's going to accumulate an average minor slowdown for me (even typing normal sentences like this). Finally, I look at this from 2 ways: 1. How would I do this right now in a file explorer? 2. How could that be improved if there was more free reign with the GUI? With way 1, it would be Right click in the space, hit new folder (or hit a new folder button set aside for it). Then hit a tab to organize by filetype, select all txt files, move to folder. I say if there's only a few files and no scrolling, the GUI will be faster. If there's a LOT of files, the CLI will be faster. Either way, I admit it doesn't feel great on the GUI, it feels clunky and sluggish With way 2, I'm seeing lots of potential for optimizations: -I could have a mouse gesture assigned for creating a new folder (a gesture specific to my file manager), so I don't have to go through the slowdown of hitting a tiny icon to create a folder, or right clicking and scanning through a submenu, I just do it instead. -I could have a hotkey or gesture that pulls up a submenu of icons representing all the different file types contained in that folder. I could then select .txt and then the window only shows me .txt files in that folder, or another hotkey or gesture that selects all of them. -I could make a quick gesture to copy what I've selected or use a hotkey -The folder tree could be represented visually somewhere, additionally I could have an ongoing history of folders I've been to or created. -I could then hover over that folder, then hit the paste hotkey or gesture and it's done. For Way 1, the time involved highly depends on how many files there are. For Way 2, I could imagine these actions all occurring within 2-3 seconds. regardless of how many files there were, and less likelihood of a mistake (for me anyway). I think this time would put it as competitive with the CLI. With file management, I think right now, the GUI loses ground if there's a MASSIVE number of files AND there's a wide variety of filetypes in one directory, due to scrolling required, but I think changes could work to mitigate that (like in my way 2 method + I have more ideas I'll put in the followup). I think the CLI starts losing ground the longer your commands or names get. So if I need to move files from a subdirectory 8 levels deep with long names to a completely different one multiple subdirectories deep, my understanding is that slows things down. With the right GUI, that's almost irrelevant. EVERYONE ELSE: I saw a lot of great posts and I'm reading all of them, I just don't always have anything special to add

See this is where we have to look at the larger picture for the user. Anything that can be launched with a keypress can also be launched with a shortcut or a mouse gesture. You just have to set up the script for what you want ahead of time, that's true in both cases. So in my eyes, if they're on equal footing, you have to look at how often you need that key, how easy is it to hit, how close is it to others that you use, etc. Plus you mention launching a video setup, well that's just your foot in the door. Once it's LAUNCHED, then I probably need to be navigating it via mouse for maximum speed. So my hand is heading back to the mouse, unless this is a one-handed hotkey launch. I mentioned in the video I saw potential with custom keypads with ergonomic though behind them. Really, I'd say that's not incompatible with anything I was saying in the video. It's once you start needing TWO hands on the keyboard to make the most of hotkeys that I think average users (or anyone who needs the mouse for maximum efficiency / convenience for software they regularly use) start losing a lot in terms of overall efficiency. Someone linked me to this, I haven't had a chance to test it, could be super powerful if it's using AHK: https://www.autohotkey.com/boards/viewtopic.php?t=12078

Okay, sorry, I realized I wasn't clear with my wording, I actually wasn't arguing people should be UNABLE to have desktop icons if they want them, what I meant was this was ugly as a default and a poor way of doing things. Of course, if people prefer that, they should still be able to have it. To be super clear, I want people to be able to customize their GUI in ANY way they want. As for "don't use them", yes, that is POSSIBLE, and it's going to be an ongoing battle for pretty much as long as you use a computer AND with no alternative provided. Every time you install a new program, it's going to dump icons there. Some installers give you the option not to do that, others don't. You can never escape desktop icons permanently by default. Additionally, there's no other alternative given if you want to rapidly access everything that WOULD go to the desktop. Windows is designed to have desktop shortcuts. If you NEVER want to see them AND install new software, that's just not an option under normal means. I can click in what would be the taskbar area also. I could also configure it to have them come up literally wherever the cursor was, but that would have taken more brainpower to figure out how to do, I was lazier here. There are 4 inefficiencies to what you're proposing v. my method: 1. I have to travel down to the bottom left of the screen each time 2. I have to click twice as opposed to once 3. Even when clicking once, I have a higher travel time to the folder in ADDITION to the travel time to the start menu 4. I can't fit as many programs in the same amount of space Maybe some of these can be fixed, I don't know. For the record: my method could be a lot better, but my options were limited. I think I might need an additional piece of software for this purpose specifically I'm not defending what I have here. I said in the video I thought this was an example where circle dock would be superior, because it would involve less travel time than the start menu. I've been meaning to use something better than Notepad and for a while I did, but then I had a power failure and lost all contents of a file I was editing (I thought that shouldn't be possible, since it would be stored in memory and not alter the source unless I was saving), this made me more wary of 3rd party programs, but I should probably give some more a try sometime. Again, it's something else I haven't gotten around to. I was hoping people would post themes showing the nicest looking notepad-like colors, fonts, etc. but that hasn't been the direction the commentary was gone exactly. Hey, you mention themes, what are the nicest ones for Notepad++? This has minor inefficiencies, since it involves 2 hands for maximum efficiency (if you use keyboard-only you have to hit the tab key repeatedly to swtich to the task you want) and involves a little more travel time. This was an example too where I had access to better options long before Windows did, I've had a dock system like that off and on for maybe 17 years. Well first, I've had this since 2002-ish, so I would have had to wait 13 years for Windows to give me what I wanted. Second, I admit I could be an idiot here, but could you tell me how to enable it? It doesn't do that by default. I just did a test on 10 and tried copying a bunch of files, the second job started immediately, wasn't queued. I saw the option to PAUSE a transfer, that's it. I admit, I'm either stupid on this or Windows 10 does not support queued file transfers.

This is part of the reason I plan to make a followup. I didn't adequately explain that what I have NOW is NOT my vision of what I WANT the GUI to be like. The only aspects of my current GUI I defend are shortcuts off the desktop, my hotspot launching option, and rapid identification. I apologize if I made it seem like my current GUI was the way to go. Now, if you're attacking me on the pie menu system and gestures though, that's fair, I'll take that. I WAS advocating that I think that could lead to a lot of advancements. It's not that I won't try it, it's that I'm not seeing the speed increases for non-coders. If I spend most of my time doing multimedia, that means I'm on the mouse a lot and it's unavoidable. So that means switching back to the commandline to type is an additional delay in time. Again, this is partially about EFFICIENCY. I acknowledged in this thread that if you're typing on the keyboard most of the time, the keyboard could be more efficient! That was a consideration I hadn't made before. As for irrational zealots who want to take away my mouse, have you seen the Youtube comments? I'm not against using the keyboard, I'm against slowdown. If I'm working in multimedia and web browsing and many other varied programs, using the mouse for maximum speed is inevitable. So that swapback needs to be taken into consideration, otherwise my TOTAL speed goes down. But disregarding that, give me some non-programming scenarios where the CLI is simply faster and I bet I could come up with some GUI concepts that would be about as fast. There might be a few where there the CLI edge it out, but I'm guessing not a lot. I think it would have be something extremely specific before a versatile custom-designed GUI would fail.

To the "just use CLI / keyboard" people. I may put this in a followup video, but I think I realized what the issue is. I work a lot in multimedia. I'm editing images, audio, video, doing web browsing + I tend to be visually oriented. By all means prove me wrong, but I'm guessing this is all stuff you don't want to do with the CLI / keyboard-only interface. My guess is the biggest proponents of it are coders. So for you, the times you need the mouse could be in the minority, so it makes sense for you to push the keyboard as far as you can. For me, unless I'm typing emails or scripts, it's reversed. The times I need the KEYBOARD are in the minority, since so much of what I do relies on visual manipulation, which isn't the keyboard's strength. Even then, I'm hardly in love with the mouse. I wasn't kidding about wanting that Minority Report interface for video editing. So before somebody else tells me to buckle down and go CLI, maybe consider the last time you edited pictures / audio / video using only the keyboard to better understand where I'm coming from.

When I do a followup, I was going to talk more about possible improvements for a file manager. This may be halfway there, I'm not sure of its functionality. I really like the more visual style of it, I guess my question is if this is something could be auto-generated or not. Even if it can't, it might be good middle ground if you have a LOT of commonly accessed stuff. Yeah, I need to get something better. I actually had one replacement earlier that was working out okay, but then I had a power failure and it erased an open file. That was a bit of a shock for me, since I thought that could only happen if I was doing a save or something and all normal editing took place in memory, but apparently not! I forgot the name of what it was, but I uninstalled it and have been meaning to get a replacement, but haven't gotten around to it yet.

I'm too busy for that kind of fun, but I could maybe make the jump if the end result was worth it. I have a similar attitude on GUIs actually. I don't want to fiddle with this stuff forever, I want to put in the work to get it right, then forget about it for 10 years or more and bask in how much nicer the experience is. As for keyboards, care to weigh in on which method you think is best?

Yeah, it's not QUITE my look, but I like a lot going on there, I tried to include a BeOS shot in the video. See, I disagree with the "can't ever" sentiment. I think that's a lack of imagination (though to be fair, if there's one thing I'm good at it, it's imagination). I think gestures actually COULD get you there and potentially even faster, but they would really have to be thought out. It feels so weird to me, it's like we're existing in a time before martial arts even exist. Anyway, I'm glad to see some sanity. I recognize that CLI can do a lot, but I think there really are scenarios it just isn't as good as even a present-day GUI. That's why I tried to show in my video examples of selecting many specific files files from a line up to move, jumping from multiple subdirectories as fast as possible, using a frontend to change options for DOSbox on-the-fly, etc. (actually that later one I could potentially see being as fast in a CLI with the right utility, but certainly not by default). If I could express one point to the CLI die-hards here, it's that no, the CLI is not 100% faster than the GUI in all situations, even with what we have now. I think with the right tools, it could be MUCH faster, but I'm still trying to figure that out. The people who insist it's better in ALL situations I think are making it more of a religion rather than trying to look at it objectively (or hell, prove me wrong in my examples above). I think it depends on what you do too. I'm not a programmer. I'm very multimedia-heavy with my work. If I'm not typing sentences for scripts, emails, and searches, I really don't need to be on the keyboard that much. Heading to the keyboard is usually the greater inconvenience for me, whereas if you code all day, the mouse might be the bigger pain for you. Don't get me wrong, I'm not in love with the mouse either, I was getting tempted by those motion control gloves + keyboard.