Navigation or forgetfulness? On finding purple links in your search
We here at Stack Overflow are flattered that we are your go to place for answers; in our 2020 Developer Survey, 90% of you said you turn to Stack Overflow when you are stuck. Not everybody goes right to the site for answers; we get that a lot of you find answers through your search engine of choice. Sometimes, you find your search turns up a purple line, which means that you’ve already searched for this problem in the past.
We found that when you find that the top link is purple—that is, already clicked—52% of you think, “Hello, old friend.” But another 14% see that purple link with irritation and another 16% see it with amusement. Maybe you don’t even remember having this problem before, much less what the solution was.
There are basically two things going on here: navigational search and memory imperfections. This article will explore some of the research around both of them.
In search of old friends
Early web search engine Altavista discovered that there were three primary types of search queries: informational, transactional, and navigational. Informational searches are when I want to know something, like the number of quarts in a gallon or the names of all the monsters in the original Destroy All Monsters. I’m looking for something new; I’m going to click on the link that looks like it best addresses my concern.
Transactional searches are where the user expects to get to a site to do something else. Buying movie tickets or ordering pad thai all fall under this kind of search. You’re not looking for a specific site from which to order these things, so any will do so long as it gets you sweet, sweet Thai food.
Navigational queries are intended to reach a specific site. Once search became automatically part of the address bar, it often was easier to search for the name of the site than enter the URL. I have absolutely typed “stack overflow” into a search bar and hit return, and I’m sure I’m not alone. Some estimate that 10% of all searches are navigational.
But some of these navigational searches are about returning to a page that you initially found via search. Perhaps it’s an excellent answer to a programming question that you found in a quick search and want to find again. You could enter the same search terms and look for the first purple link. Ah, hello, old friend, this is the exact answer I was looking for.
This is pretty common for sites with a lot of content, especially user-generated content, like Stack Overflow. Designing a navigation system that can bring you to a specific page without using search or bookmarks is incredibly hard.
The resistance of memory
There’s another pair of reactions to searching and finding a purple link: irritation and amusement (there’s also 18% who feel indifference strongly enough to select it on a survey). These seem to indicate that the person searching forgot that they even knew the solution previously. It may be that they didn’t even know that they had this problem in the past. This sort of forgetting—if even the information made it into memory in the first place—is extremely common.
Our memories are more flawed and imperfect than we would like to believe. In the book, The Seven Sins of Memory, the William R. Kenan, Jr. Professor of Psychology at Harvard, Daniel Schacter details seven ways that our memories fail us. For our inquiry into the case of the purple links, we’re interested in just two: Transience and absent-mindedness.
Transience refers to memories that just don’t stick. This happens mostly with new information that you’re trying to remember. Consider coming across a new function, keyword, or property. You read the docs, think of interesting applications for it, then the next time that you need to use it, you’ve forgotten what it’s called.
As early as 1885, German psychologist Hermann Ebbinghaus ran experiments where he memorized strings of nonsense syllables, then tested himself on how much he could recall. Through repeated memorization and testing, he determined a mathematical curve that describes how well something is retained over time. This curve has been refined by further research, but the basic idea remains the same: we forget new information unless it’s reinforced. Two studies in the late 50s found that subjects forget three nonsense syllables in as little as twenty seconds!
On top of that, the routine of work can help us forget the individual actions that we go through in a given day. A lot of daily programming happens within a work routine, which over time becomes just a part of the regular routine. A study in the 1950s asked workers to define their daily routine. Then, when asked to describe the previous day and a day from the previous week, the study found that workers remembered specific details of the previous workday, but seven days later, they only remembered tasks that fit into their daily routine. The more complicated the story they told about their average day, the more they remembered about any given day.
Most of the time, we search for answers during the workday, at least for professional programmers. The search is triggered by a problem, and unless we have “solving problems” as part of our regular workday story, this search is going to be unusual and likely forgotten soon after.
Sometimes, the issue is an attentional problem, not a recall problem; that’s related to the other memory sin involved in this: absent mindedness. Where transience described above prevents a memory from staying encoded, absent mindedness can prevent the memory from being encoded in the first place. Think of the times when you’ve lost your keys and you need to find them. You don’t remember where you put them. Even when you find your keys on top of the refrigerator, you have no idea why you put them there.
There’s a strong link between giving something your attention and being able to remember or recall it. For a while, attention was thought of like a movie; you watched the world and paid attention to everything in front of you. But then researchers found that we only saw—much less remembered—what we gave our attention to. It’s called inattentional blindness. This video is my favorite example of that.
The classic metaphor for attention is that it’s like a spotlight. You may be watching a movie, but you only pay attention to the things covered by the spotlight. Sometimes, you can pay a lot of attention to a single spot, other times you expand or split your spotlight and pay a little attention to a larger number of things. But you generally have a limited amount of attention that you can focus your spotlight on at any given time.
At work, you have tons of items demanding your attention at the same time. You’re writing code while fielding questions in your chat program, estimating story points for multiple tickets, and trying to prep for your next meeting. If you find a solution to a smaller problem, it can get lost, as it’s not part of the larger items that you’re focusing on. With so many demands on your attention, your memory can suffer. Research has shown that heavy multitaskers may have worse memory overall.
Sometimes that multitasking is just rapid task switching and none of the tasks gets full attention. Other times, you’re doing work that is so deeply in your wheelhouse, something you’ve done every day and comes almost automatically, that you can do it and another task without either suffering. But the downside is that you may find that you’ve passed a chunk of time in an automated haze and remember none of it.
There’s another draw on our attention at work: stress and anxiety. If you have a culture of constant deadlines, unsupportive or vindictive management, or even an anxiety disorder, your memory can suffer even more. You find a solution to an immediate problem and never encode it as a memory; you’ve got more pressing, longer terms worries occupying your attention.
There are simple things that you can do to make a solution stick a little better for the next time you run across the problem. Any time that you deliberately attend to something helps, especially if you do so in new ways. Something as simple as chatting with a co-worker for ten minutes about the issue and solution can help. You can come up with elaborate stories to create a richer memory (or tuck them away in your memory palace). That’s because memories get stronger through more associations.
In fact, the more exposure you give yourself to a concept, the better it sticks. But repeated exposure isn’t enough, you need to include more difficult study tasks, like recalling the stored information. You could use a flashcard program like Anki to beat back the effects of the forgetting curve. Tap into your visual memory and draw a diagram; even writing it down may help.
But don’t worry if you still find yourself forgetting some things. Forgetting is actually an essential part of any intelligent organism’s mental health. Like you do with your hard drive, your brain clears out any unused memories to make room for new information. That’s because the brain isn’t meant to retain information; it’s meant to make good decisions. If the memory of what you did to resolve an error isn’t going to help you make a good decision, pow, it’s out of there.
There’s also some concern that we are outsourcing our memory to technology. A 2013 study found that subjects who were asked to remember certain objects in a museum tour and photograph some of them actually had worse recall of the objects that they photographed. But this forgetting may be efficient as well; if you have a digital copy of the memory (or solution to a problem), why should your brain waste resources on storing it as well? You could just remember the path to it—the search phrase—and use that. Et voila, navigational search.
Memory and attention are places of active research right now. If you’re interested in seeing what our psychology Stack Exchange has to say about it, this question about the neurobiology of it is a good start.
If you’re willing to wait a little bit, you can get insight into that research in a book form. I reached out to Schacter and it turns out he’s updating the book. “When I wrote The Seven Sins of Memory two decades ago, it was already clear that memory lapses and errors have a significant and often surprising impact on everyday life,” he said. “I’m updating the book for a revised edition to appear in 2021, and it has been an interesting task to document various more recent real-world impacts of the seven sins, as well as our progress in understanding them.”
Whether you come upon a purple intentionally or unexpectedly, amused or enraged, we can only hope that it contained the answer that you needed. Solve problems, build solutions.Tags: bulletin, memory, stackoverflow
Thought this was going to solve a problem when you’ve already clicked on all the promising looking links and haven’t found anything. Do you know a good article about that?
For what it’s worth, I think we interpreted the survey question a little differently here, I wrote “Hello old friend” as my selection but I don’t really use search in the way you’re describing. I do use documentation or SO to remind myself of things (probably more efficient than keeping my own notes, which would end up being a giant, less searchable archive) but usually I kinda just remember a (usually Google) search query that worked, I don’t really forget that I searched for it and accidentally stumble on it again.
When I got that survey question, I was thinking of a scenario where I was searching for something that there was no good answer for anywhere, and a stumped Google resorted to giving me a page that the algorithm thought was the best, but wasn’t actually helpful.
When I get to that point I end up trying different rephrasings of the same search or inverting search terms (e.g., “How to do X in Y”, “Can I do X in Y”, “Is it possible to do X in Y”, “X does not work in Y”, etc…), there’s a chance one of those will work if the other didn’t. In that case, I kinda know that the search engine isn’t optimized for very rarely searched things, so it’s actually a little exciting to stump the search engine, it means I’m getting into things that not many have seen before. There’s also a mild “why isn’t this in the documentation” annoyance factor too, but I’d say it’s mostly just kinda exciting; that wasn’t a choice so I went with “Hello, old friend” which I kinda read as sarcasm.
The way you’re describing how you use search is what I was trying to get at with the fist part — it’s a navigational search when you’re trying to find your way back to a specific location, like the answer that you already found helpful. I do this too, mostly with food recipes. Browsers and web design has usually relied on bookmarks for these sort of stored pages, but those get unwieldy real fast.
I think you’re also describing a second situation that I hadn’t considered, where you’re looking for an answer and all you have are unhelpful purple links. It could be that the choices in the survey were a little too open to interpretation.
“Hello old friend” has also been on a couple of occasions when I searched for something and happened to land on an old answer that I posted myself and then promptly forgot!
there’s also permanent private browsing mode which does not make any links change color
I don’t often notice purple links because I’m colourblind :/
Spose i should look for some A11Y browser config…
I once googled an issue & the first hit was my own answer on Ask Different which I’d completely forgotten I’d ever posted.
“Hello old friend” indeed 😉
I’m colorblind. I don’t see purple links at all. They’re all blue to me!