Conversation

In trying to set up a physical workspace for the first time in decades (workbench, instruments, tools...) I'm struck by just how much computer UXes have evolved from already-bad workspaces to pure consumption spaces.
9
151
The desktop metaphor was already quite terrible, built on top of the document metaphor. The "app screen" metaphor is far worse, since it obscures the filesystem and doesn't even give you a flexible folder-like organization scheme (grouped apps don't count).
1
53
Think of all the things that live on a computer: 1. Things that are reasonably close to "documents" 2. Cloud connections 3. Applications that are like information sensors (like the Activity Monitor on a mac) 4. Processors (esp. for things like images) 5. Actuators (drivers)
1
23
The sheer variety of things that live on a computer (and here I'm talking about physical devices you use as a computer, not cloudy distant abstractions) makes the "desktop" or worse, "screens" insufficiently expressive to capture the capabilities embodied.
2
23
My mac "dock" contains a bunch of folders, a bunch of drivery/sensory things, folder-proxies for cloud (dropbox).... why not a richer metaphor like a documents area (filesystem = physical cabinet), a sensor area (like say instruments in a lab), work areas like soldering station
2
18
In the grad school lab I worked in, the center was dominated by a large, heavy optical table with a couple of large experimental rigs (a control-moment gyro rig, and a noise-cancellation rig). They were bespoke work constructs and the centers of attention.
1
13
That's how I think of my work: a bunch of varied sets of commodity capabilities surrounding a couple of bespoke configurations of things. And this is not just research labs. An ordinary home kitchen is organized this way, around the stove-top, fridge, and oven as anchor "rigs"
2
20
A desktop metaphor is like a library. A screens-of-apps metaphor (what do you call this? it's not really even a desktop) is something like a restaurant menu or a living room. Optimized for consumption, not work.
3
28
We talk a lot about repairability of devices, which I actually don't care about much. Computers have become more sealed, less repairable, but also more reliable and less in need of repair at the physical level.
Replying to
This can be reasonably justified as a good tradeoff between reliability, cost, form factor and reliability. Just like cars. The hardware is getting higher-tech, less actually repairable by average people, and less in need of repair. But the same cannot be said of software.
1
7
With software, things have gotten MORE awkward, messy, in need of maintenance/repair etc etc. In the 80s, when the command line was still the UX for everybody not just for unix geeks, there was a match between capability and what I'll call "workability"...
1
19
The command line is like a primitive work interface as opposed to consumption. It's like a campfire in relation to a modern kitchen, or a basic household toolbox (screwdrivers, wrenches, hammer...) in relation to a full-scale lab. If it had evolved as a *work* interface...
1
13
... it would look very different. So there has been highly asymmetric evolution. the computer as a consumption device has evolved 3-4 generations towards increasingly frictionless digital shopping basically. But as a production device, it is basically stuck in 1988.
2
31
This is why, over the years, even as my methods for managing physical workspaces (GTD, setting up offices/desks/kitchens, and now a home lab/workshop) have gotten more mature and sophisticated with the evolution of my own thinking, my digital methods have not.
1
12
I'm still struggling with janky link management... most people now seem to solve that via tab proliferation in the browser. (This is one reason Roam was so impressive for me... it improved my link management)
2
16
I still can't easily organize my "tools" into say "work areas" for "image processing" vs. "spreadsheet work" the way I can separate an electronics workbench from a kitchen. Automation is still at "keyboard shortcuts" level. There is no intuitive way to group tools/processes.
3
21
The browser is a whole other level of interface, but the limitations there are more forgivable, and there has been more of an effort by at least some services to use the browser well as a workspace rather than a consumption space.
1
10
I recently tried OnShape (browser-based full-featured CAD), the first time I've played around with CAD in 30 years. Quite impressive, though the interaction paradigms haven't evolved that much, the fact that it's now all in the browser with good version control etc. is neat.
1
6
This feels like say, if the kitchen evolved in capability from a camp fire to a modern kitchen, but we still tried to do everything with a poky stick and huffing/puffing/fanning. Shell scripts, keyboard shortcuts, and emacs are not the answer even for geeks, let alone home cooks.
1
12
One of the problems is that unlike consumer interfaces, which have given rise to the whole specialized profession of UI/UX designer, producer interfaces in the early, fast-evolving stages of a technology are built by the producers themselves, not a specialized separate profession
1
10
I'm reading a lot about 17th century science now, and it's amazing the degree which early scientists were also the early instrument makers (Galileo, Huygens). The separation into tool designers vs. producer-users didn't happen till the 1670s or so.
4
27
Now the thing about the computer as a producer's tool is that it is so powerful and capable, it hasn't stopped evolving rapidly since Day 1, which means it's never plateaued into a zone where a tool-maker class (a producer-side UI/UX class) can take over fully.
1
11
Ie, the only people who can move this forward are people who use computers for production at the edge of its evolving capabilities. One subset (the command-line purists) has decided not to try. Instead, command line prowess becomes the status thing. Is there a subset trying?
3
10
It's not been entirely static. I think version control and package management have been huge conceptual and implemented leaps in production UI/UX thinking. But it's so limited...
1
7
And unfortunately it's almost all limited to actual programmers. People who use the computer as a tool for other purposes, like say image manipulation or data analysis without coding... they haven't been able to design/customize/evolve/innovate their own tool environments much
1
21
Actually it's worse: it's not even just programmers. Even most experienced programmers aren't deep in enough to do more than say customize emacs or their IDE. The actual evolution is driven by programmers' programmers -- systems programmers.
1
11
Like Git, arguably the biggest innovation in producer-side UI/UX, is a programmer's shop tool built by the ultimate programmers' programmer, Linus Torvalds. These people use computers for one very narrow kind of production work: making better computers.
1
12
We need Linus Torvalds level people trying to think up better, broader producer-side systemic UX metaphors. Like a "workshop" or "kitchen" or "lab" metaphor for the computer.
5
26
And I'm thinking here mainly of full-fledged ones, like a laptop or desktop, but also for phones and tablets and voice-control devices.
1
2
Jarvis is a voice assistant metaphor, but has the skills of an extremely capable lab assistant/tech or shop machinist (gender aside: he's a male voice, unlike most voice assistants, who are descended from office secretaries and have female voices)
1
3
But Jarvis is not quite right, since he's a digital assistant for a primarily *physical* space, and controls robot arms and prototyping areas and stuff. We need Jarvis-grade expressivity for *digital* environments. Like think a massively more powerful Clippy that actually works
1
10
This train of thought started with me organizing my home office into a sort of lab-maker space. The project crept up on me, but once I realized I was doing it, it was easy.
Quote Tweet
Thinking about desks I've had over the years. I had an actual L-shaped desk for home office for several years (I think 2004-09) and 2 cheap desks in L-config for several more years. Single desk for the last year. Now going back to L with a twist: second leg will be a workbench.
Show this thread
1
2
A workspace designed for writing and consulting is very simple. Just a desk, a laptop, room for some papers, maybe a whiteboard, a bookshelf with commonly referred to books within reach. Maybe a mic/video rig if you do podcasting and stuff.
1
2
It's also very different in very obvious ways from a space where you can do simple repairs, mechanical/electronics projects/soldering etc. Once I realized I was headed in this direction, the decisions were obvious:
1
1
1. What kind of workbench to get, 2. Where to put a 3d printer 3. How to hold work (small desk vise?) 4. How to store small parts 5. How to ventilate work area 6. Creating enough length for an optics project Then I thought... hmm what would a lab *computer* look like?
2
4
And here I realized I had *no idea* ... my first thought was "maybe I should build my own computer" but though that might be a fun project, it's not actually salient to what it means to have a "lab computer."
1
3
Back when I was teaching an undergrad lab course as a grad student, a "lab computer" meant a regular PC with a data acquisition/control board attached and some instrumentation software for the students to do their experiments. There were programs like LabView, LabTech etc.
1
Show replies