The long-lived “desktop” operating system has been with us for almost 40 years. Although some of the mechanics have proven remarkably durable, contemporary computer usage is very different from the context these were born in, and it’s time to do some rethinking.
I’m going to outline the original idea, illustrate some changes in computer use, and suggest some new ways of thinking.
(I’m a longtime Mac user, so my experience and examples are specific to that platform. Maybe Windows is awesome now, I don’t know, my last real exposure was with XP when I worked at Microsoft.)
This is the dialog that I see every time I want to close the Preview app to clear my desktop of all that clutter in a hurry for a video call. The option I want is the one that’s missing: quit, but keep all the documents.
Partly this is just silly UX: the app already keeps all those unsaved images around. So I use Force Quit to quit Preview, because it does what I want. But that’s not what Force Quit is for, and anyway the reason the option that makes sense isn’t here is due to a philosophical disconnect illustrated rather precisely.
Everything that is outdated in computer desktop usability is captured by that dialog box.
What is the desktop metaphor?
Ever since the dawn of the Graphical User Interface operating system on mass-market computers in the mid-1980s, they have been designed primarily around what was then called the “desktop metaphor,” and is more usefully described as a document-centric system:
- You, the user, are imagined to be looking at the top of an old-fashioned, pre-computer desk, which is covered with documents for your work: memos, spreadsheets, pictures. Topmost is whatever you are working on right now.
- A “document” is stored in a “file,” and that file lives permanently in a “folder” (think: filing cabinet) or temporarily on your desktop.
- Because this was designed to be a sort of metaphor, each document is a file and each file is kept in exactly one place, and it’s up to you to create the folder organization system that makes sense for your work.
This should sound both familiar (yes, that’s why they’re called folders, and still use an icon that looks like a manila folder) and also strike you as a musty metaphor. It envisions computer use as a business function, and takes as inspiration the way office productivity happened a half century ago.
This design worked in the 1980s and 1990s. It spurred revolutions in productivity and usefulness, as productive work translated more or less directly to the screen, just with way more features. Need to write a memo? Ditch the typewriter and bring up Word and edit to your heart’s content. Then save it in a folder on your disk for later access. Same with your budget spreadsheet, etc. Just making some quick notes? Maybe save them in a file on your Desktop to get to it again easily. At the end of the work day, turn the computer off, close the office door and go home!
Sure, you couldn’t search your hard drive in those days, you had to remember which folder you saved which file into. It wasn’t hard though, since all the files were documents you created deliberately using the handful of programs on the computer, and disks were small and didn’t hold much. And you didn’t perceive any inconvenience, because remember, it was very much like what you did in the real world, just transposed onto a little screen.
The internet completely changed what we use computers for
There’s a whole techno-econo-social history of how we got from then to now, but the specifics aren’t important. What does matter is that if you were to fire up an old Macintosh from the 1980s (or even 1990s) today you’d find the interface to be surprisingly familiar but you’d be stuck asking: OK but does it… do anything?
That’s because the internet undergirds almost everything we do on a computer today. How many browser windows do you have open right now? (I have 37 windows open, with some 75+ tabs.) Email, calendar, cloud document collaboration, Twitter, Instagram — are mostly or all in the browser. What else? iMessages and Slack are apps that require the internet to do anything at all. None of this stuff existed in 1985.
And instead of creating documents — a sort of heavyweight work product — a lot of what we work with on our computers now are fragments: URLs and meme gifs that we copy paste between windows or chats, a PDF that we download to print out or fill in and email. Data that we copy from one spreadsheet to another. Photos from the phone that we crop so we can upload somewhere else. Video clips of toddler nieces doing something cute.
Some of these fragments become files in our Downloads or Desktop folders, because that’s the default place to land them. Still others we try to save somewhere more permanent. Many don’t ever have a home “on our computer” but pass through the computer as an ephemeral copy/paste clipboard item before vanishing into programs or cloud software like chat apps or email. We use the set of “open documents” (tabs) as its own sort of memory or to-do list.
But all of this behavior is, still, grafted onto that desktop and filing cabinet. The cognitive load produced by modern internet use on top of a 1980s metaphor is high: I have a Desktop scattered with (currently) 132 icons. I have a Downloads folder with (currently) several hundred files, most of which I could safely delete, but who has the time to decide? Ditto the “open documents” in Safari, Preview, Pages, Numbers, TextMate, etc.
Meanwhile, there’s all sorts of stuff that I can’t find — that link from maybe last week? or was it the week before? Who did I paste that to again? Or the tax form that I definitely downloaded in the last few months but I can’t really remember what it was called or whether I saved it or not. The web page I left open a couple days ago because I knew I’d want to look more closely later but now it takes me more time to actually find it than to re-Google.
Thus, two problems:
- An ever-increasing junkpile of disorganized and often temporary material, accreting unbounded in folders, on the desktop, and in every programs’ set of open documents.
- An ever-increasing set of ideas and informational fragments that are worth keeping and accessing, but either don’t live on the disk as files in the traditional sense, or live solely within cloud services accessed through a browser, or are so hard to recover that it’s easier to just Google around and search the email yet again. (This might get you a redundant download too as a bonus, now you have
Docs Final-4.zipto make you feel dumb for having gone looking five times.)
If this sounds naive as a problem statement, maybe it is. We have solutions to some of this already, and have for some time.
The ascendence of (good) search over organization
Yahoo was founded, famously, as the directory of the fledgling World Wide Web. It used a tree structure to categorize all the web pages that there were. This was OK because there weren’t many web sites in total yet and it was much easier to find what you wanted in “Jerry and David’s Guide to the World Wide Web” than it was to remember the URL to type in.
Within only a few years, this system was hopeless. The web exploded exponentially, in terms of “sites” but also it evolved ever-shifting ways of presenting and producing bits and pieces of content, such that the directory format of top-level sites was not even a great way of cataloging the world, let alone finding your way in it.
Thus began the search era, still with us today, where Google ranges across all of the content in all of its forms and then runs your search through a massively complex bunch of heuristics and gives you results ranked by “relevance.” There were others before Google of course (RIP AltaVista) but they were pretty much all bad and the reason Google is the verb is because it was then, and still is, the best at what it does.
You probably also appreciate how Google can land you on the specific thing you want, not just the site that contains it. (We have all experienced the frustration of a bad web site’s built-in search that is way crappier than just Google would be.)
This was a sea change in how we organized and thought about looking for information in the public sphere. The world could be a complete disheveled mess, but as long as our interface to it was a clean white page with a search box, the real complexity was concealed.
Does any of that sound familiar? Our computers have been applying small band-aids to our own information retrieval problems for years now. Trying and only partially managing to paper over that creaky desktop structure.
On the Mac, Spotlight (the search box that shows up when you command-space) has been around since 2005. Its job is to add search to “stuff on the computer.” It does an OK job, but not a miraculous one. It still mostly understands information in terms of local files, and relevance in terms of text match and date. It’s a lot closer to AltaVista than to Google.
Meanwhile, over in Safari, there’s a browsing history that is, technically, searchable, but it’s clunky to access and not searchable in the ways you probably want.
The clipboard (the invisible temporary place that is the destination for the “Copy” command and the source for the “Paste” command) is still a relic of the 20th century — it stores one thing, forgets everything that came before. I use the handy utility app Yoink (there are many similar) to help a little with doing slightly more complex stuff with the clipboard. Still, all sorts of images, links, and other useful fragments pass through my clipboard every day, and are then lost.
This all feels like a rather huge missed opportunity.
A computer for the modern user
I think that it’s time to rethink what a desktop computer is used for, and rebuild the UX around that. Here are some guiding principles:
- Every piece of information the user interacts with is a fragment, and every fragment is findable later through standardized kinds of interface. This includes documents and notes the user creates, data the user copies/pastes, but also things like tabs opened in the browser. The fragments have rich metadata, too, that helps them be found: not just the text in the documents or notes but also the text in the open web page and the sources and destinations of the pasted items and what things they resolve to. Imagine searching for “cat” and one top result is the photo of a cat that I copied and pasted between iMessage chats last week — modern image classification plus clipboard fragments, etc.
- Fragments are created equal but rapidly accrue value, or decay. Usage of each fragment (opening, accessing, reviewing, re-copying, editing, etc.) is tracked. Not just “last modified” date or whatever, but its whole history of access pattern. A fragment has a whole profile of revealed usage preference associated with it. That profile, along with modern deep search tech, is used to determine relevance.
- Maciej Cegłowski’s powerful “The Internet With A Human Face” highlights the cognitive dissonance between human memory (gradiated and complex and eventually faulty) and computer memory (binary: flawless or nonexistent). We should model fragment search and access after human memory, using access patterns and usage patterns as rich metadata to help the computer understand what is important and what is relevant. And what is related to what. That doesn’t mean auto-deleting documents after some period of time, but just as it’s a lot harder to Google something generic that happened a decade ago and garnered little attention since, it doesn’t need to be “easy” to find the untitled scratch spreadsheet we cooked up to check the car payment budget in 2013 (but we should be able to find it if we need to).
- Programs need to cut it out with the dreaded “do you want to review the 88 open documents” crap. This is deeply irritating and also hurts system performance. Just like applications have been auto-saving for years (huge improvement!) just close stuff after awhile if the user doesn’t even look at it. Don’t delete it, file it away with rich metadata so it can be found if necessary. Most apps on a modern desktop make you pay a real performance price in addition to cognitive load for having a boatload of passively open documents. (Newer mobile OSs (e.g. iOS) get this right — having 500 tabs open in mobile Safari, which I always do, doesn’t hurt my system performance at all because the tabs are freeze-dried when I’m not using them.)
- Sometimes, modeling human memory, things should get deleted permanently. Content the user has a hand in directly creating probably shouldn’t. But fragments from other places should: Web history already does, though usually with a hard timeline and not a decay based on relevance. Fragments from the clipboard, maybe. Downloaded files that have never been accessed, or that haven’t been saved more formally somewhere.
Think of all the stuff that is saved but not accessed recently as receding to the lower strata of archaeological time. They’re still there, but they’re buried deep. They don’t pop up and surprise you and get in your way, you have to go digging. A question arises about exactly what to actually delete, for the sake of saving space. It’s a good question, and I don’t really know the answer to it.
One answer is: who cares, disks are big, save it all, forever. (If you saw my computer’s blindly copied nested Desktop folders going back three laptops in time you’d know that it’s possible to already live this way.)
Another answer is: rather than each app providing its own lousy and time consuming “review the open documents” experience, or expecting people to spend time reorganizing their folders periodically, we might build very fast and efficient bulk delete/management UXes that lets us spelunk down into the strata, armed with the contextual metadata, and quickly wipe away large swathes of preserved junk as necessary.
There is a lot of information you care about that lives “in the cloud,” either on a web page or inside a walled-garden service (Google Docs, etc.). This is the way of the future for a lot of this stuff, and I do not wish to suggest a return to the 1980s here. But Spotlight on my Mac is not optimal if it highlights junk in my Downloads folder but not my Google Docs, right? Maybe today’s desktop makes more sense as a deeply capable air traffic controller librarian character, rather than a main center for documents.
There are also services that do various levels of the sort of thing I’m suggesting. Evernote is a great example of a brilliant tool (trapped inside long-suffering software) that indexes everything you throw into it, including text found within images, so that when its search really works, it’s pretty magical. A friend let me know about Notion, which purports to take on some related roles for teams. But these aren’t your computer desktop, the nerve center that sits at the nexus of your keyboard and monitor and local disk and internet connection. The thing that is still, at its core, the top of a walnut desk and a battered metal filing cabinet.
All of this is really a story about access to information. I am less sure what the future should be of the desktop part of the desktop metaphor. The whole thing with the piles and piles of windows open all the time has never been a frictionless experience. But now sharing your screen for a meeting is commonplace. Everyone gets to look at your messy desktop. Or you have to tidy it up real quick before the meeting. (And then we’re back to that dialog box from Preview at the top of this article.)
I also often think of my own “fragments” in terms of project context (am I working on doing my taxes? Amassing research to use for an article like this one?) and today all the various open windows scattered on the computer are an undifferentiated soup. It seems like in these ways, too, computers are not really working for us as we currently work.
I don’t know if fragments as outlined above are the right answer. I know that mobile operating systems, like iOS as noted above, have generally started afresh in their UX and have much more deeply integrated information management (and much less reliance on any notion of “files”) than the creaky computer laptops. But I am most productive and feel most in control in front of a real desktop and I’d love to see that desktop supercharged and become simple and transparent in many ways like some sort of private version of what the internet has become for public information.
I write about design, technology, history and San Francisco. I’m on Twitter at bzotto.