Hmmmmm. Maybe I was taking the wrong approach in my previous post about tabs. Incremental improvements won’t get us to the desired state of think-about-it-wiggle-your-fingers-and-you’re-there-ness.
Let’s go back to first principles and see if there’s a different way to think about the tab problem.
When you want to get back to a tab that you have open, how do you identify the tab in your mind? I’m not asking how you would describe the tab to someone else — I’m asking, at the moment that the thought occurs to you that you want to go to a tab, what does that thought consist of?
- Visually — “The one with a picture of a panda on it.”
- Spatially — “The one way over on the left.”
- By keyword — “The one about pandas reproducing in captivity.”
- By site — “The one on the BBC news site.”
- By activity — “The one I opened as part of my morning news reading.”
- Some other way?
I suspect that the answer will vary between people, of course, but also that each person identifies tabs by more than one method, depending on the tab contents and how they’re being used. Do some introspection — I’m very curious to hear what you notice about your thoughts. For example, I’m the type of person who relies very much on visual and spatial cues in most things that I do. I’m very map-oriented. But I find that visual and spatial cues in the browser are too weak to be of much use to me when navigating tabs. My tabs are not consistently in the same places, and most of them look very similar because the pages I use most are almost all text and no images. So I rely mostly on the textual content of pages or the activity that they’re associated with.
Here’s an even deeper question. When you want to go somewhere, does your brain even distinguish between wanting to go to a tab that’s already open vs. wanting to go to a site? Does going to an open tab feel like a different action from opening a new page, or do you think first “I want to go to page X” and only then stop to think about whether you already have it open or not?
I’ve got no answers here, only questions.
August 21, 2008 at 8:51 pm
If I’ve got a small number of tabs (and much of the time I don’t have more than 5-6), then it’s the site and/or the page title I’m looking for.
When I have a larger number of tabs, it’s usually because I’ve opened a dozen or so to flip through lots of new threads in a web forum, or lots of search results. Generally in that case I don’t do a lot of looking for specific tabs, but when I do it’s by site and by position (e.g. I want the tab that I opened about 4th from my list of forum threads, so that will be 4th in the set of tabs consecutively opened on that site).
Thinking about it, I am sometimes using position to find a tab because when there are many tabs open, I cannot scan the page titles.
The current new tab switcher looks pretty, but because it brings up the tabs in a different and unfamiliar order, I’ve stopped switching tabs with keys altogether and reverted to clicking….
August 21, 2008 at 9:32 pm
My T.I.T. (tab identification thought) can consist of all of those, depending on what I’m “working”: a site (e.g. multiple bug reports, comparing user reviews), a topic (many sites opened from search results and Places), or returning to an earlier task (e.g. a long essay I opened from my morning feed then put aside).
Yes, going to an open tab is a completely different action from opening a new page. A tab I’ve opened but never switched to is more like a new page because I have no “state” on it, just intention.
I could imagine a whizzy animated hierarchical timeline showing my clicks in different windows and tabs, but are you overthinking the problem? I use Ctrl-PgUp/PgDn and if I still can’t locate a tab I click the ‘v’ for the tab drop-down, and sometimes I look in History > Recently Closed Tabs. I don’t expect my browser to make up for my disorganization; but feel free to surprise me
August 21, 2008 at 10:28 pm
Almost always by topic. “email” “planet” “that one blog by that guy” “that dom bug i’m working on”.
The exception is when I’m doing a task that involves two tabs and I’m switching back and forth. Then I don’t even form a complete thought in my head, it’s just “ok, time to switch!” and my fingers hit the right keys for me and I switch.
The spacial location of tabs is useless to me *except* that when I do want to see if I already have something open, I look at the tabs first, then navigate. (That’s why the new Ctrl+Tab behavior has been really frustrating to me.)
August 21, 2008 at 10:53 pm
I’m an old man of habit and little variability so I think of my tabs spatially (my muscles have a spatial memory?) since they’re almost exclusively opened from groups of bookmarks.
I know that when I’m reading through my group of design blogs that they’re already ordered from left to right by importance (some combination of how often they refresh content, how good they are, how much I care if I get to them daily, etc.) and so I just read left to right through my tabs.
That applies to a dozen or so tab groups and I’ve essentially memorized the relationships between all the tabs in a group. In my “work” group of tabs, it’s Bugzilla, Planet, Bonsai & HG queries, two Mozilla Google Groups, and Gmail. If I’m in Gmail and I want to go to Planet, it’s just spatial or muscle memory because I don’t really even have to think about it — I just know to click on the tab over to the left that’s got Planet in it.
In some cases, I replace one set of tabs with another. In others, I’ve got multiple windows that organize my tabs by activity (news and tech, work, personal interest blogs, family, etc.) But with tabs segmented by activity, and anything new in that activity opened to the right of my right-most permanent tab, I just don’t have to think about it much.
August 21, 2008 at 11:09 pm
The train of thought generally goes as follows: When my brain gets bored with one page it wants to go somewhere else. If I have an unread tab open I will just close the current tab. If I don’t (or I have forgotten) I will move my focal point up and either open a site from my bookmarks or type a URL into the location bar. So I would say yes, my brain does differentiate between switching to another tab and opening a new URL.
I will tend to switch tabs i two different ways:
1. Use/look at the tab strip (and select using old behaviour Ctrl+Tab or mouse)
– spatial location of the tab
– favicon
– page title
2. Use/look at the page canvas (and cycle through using old behaviour Ctrl+Tab):
– visual: how the page looks
Independent of how I switch tabs, I will have a back-of-the-brain record of “interestingness” and read/unread state of the background tabs. (”There are several interesting tabs over to the right, especially one near the end.)
The proposed Ctrl+Tab behaviour breaks both of my two tab switching methods. (It breaks spatial order and visual – the preview is more or less useless, completely unrecognizable and makes your head jump through hoops.)
August 21, 2008 at 11:11 pm
“The current new tab switcher looks pretty, but because it brings up the tabs in a different and unfamiliar order, I’ve stopped switching tabs with keys altogether and reverted to clicking….”
Me too. Alternatively, moving my right hand to hit Ctrl+PgUp/Dn, which is just as time consuming.
August 22, 2008 at 8:42 am
I often use tabs when looking for information as a bread crumb trail to get back where I started from. For example I do a search for subject X in Google. I scan the first page and find something that looks interesting, I then use ‘open in new tab’ to open that link. Depending on my confidence level of that page being my desired result I will then go to that page or proceed to open new tabs. But it does not stop there. I will then go to one of my newly found pages and see some interesting information. Because this site it unfamiliar to me I will open the link in a new tab, and then go to that tab. If this tab is a dead end I will then use the tabs that I open to go back to where I want to be. (yes, I could use the Back button, but by the time I figure out that this is not where I want to be I have to go; Back, Back, Back, Back, Back…. You get the idea.)
I keep doing this until I become completely lost in tab hell and close everything except the original search page, where I start the process all over again.
So my point to all of this is that given this workflow what seems interesting to me is not a tab system at all, but instead a map system. I would think, some way of viewing my tracks that I browsed in some type of “node to node map” would be much more useful. Of course it would need to logically group the nodes and sub nodes, and it would need to be easy to access and quick to paint on the screen.
So all in all I think the whole idea of tabs is useful, but clumsy at best.
August 22, 2008 at 8:56 am
I didn’t think this would be my answer but, thinking about it, I almost never “think about” the tab I want spatially – unless perhaps I know it’s next to the current one.
I do a lot of “Open in New Tab” and so often have a trail of 20 or 30 tabs which represent different points in my recent browsing history where I’ve branched off. Occasionally, when I run out of memory, I go through the windows and close a load of them. This means that particular important tabs never have a consistent location.
I’ve also found myself, if I know I will want to switch multiple times between two tabs, opening the second one in a new window, because using Alt-Tab to toggle between two windows is actually more reliable, consistent and understandable than using Ctrl-Tab, whose current workings in Firefox 3 Final I don’t have a good mental model of. Perhaps this has changed in 3.1a?
I almost always switch tabs using the mouse, because that’s the only way I can reliably find the one I want. Which means that I’m using a spatial method even though I almost never think about it spatially – which is slow. I do end up scanning the tab bar and trying to puzzle out the titles very often, which is sub-optimal. I’ve even wondered whether it would be better to have tabs down the left side of the screen, leveraging the fact that normal screens are wider than they are high but web pages are longer than they are wide, to give you more room for tabs and more title text space.
And, of course, close-button-per-tab reduces the amount of space there is for a title.
To answer your question positively, I think I think of tabs by site/task – “the blogpost I’ve started”, “the wiki page I’m editing”, “the Amazon page where I’ve found that book”.
August 22, 2008 at 11:23 am
My method of finding the correct tab depends on how actively I’m using them. If I’m switching between tabs, or have accessed it recently, my use of the ColourfulTabs extension comes to the front as I know that I want “the red tab” or “the green tab”.
If I have’t used the tabs recently I’m a visual person, and a “tab bar” like NetNewsWire works well, which has a pane down one side of the window with thumbnails of all the tabs. I can easily tell from the small picture which one is the right tab to go to, or if I need to scroll further because I have too many open.
August 22, 2008 at 1:37 pm
Interesting question. There’s definitely a separation between 7-8 tabs I have open all the time (FB, intranet, GMail, netvibes, webcomics etc) and the ’session’ tabs—up to 20, probably. I know where my everyday tabs are; FB is always far left, and when I want to go to one of these I think of the topic and glance at the favicons and click.
When it comes to ’session’ tabs, I actually have to admit that I organise tabs by dragging them around. So if I’m looking (as I currently am) at furniture, I move around different options so that the order of tabs follows my thoughts, if that makes sense. So the thing I need to keep near the surface of my mind stays far right, while if I’m saving something for later I move it more tabs to the left. That may be quite unusual.
My random browsing will also be grouped by the kind of thing I’m thinking about (e.g. sidetracks, looking up resources), and when I think of a particular tab, I think about the topic, then I think where the topic is. Then I go there by clicking.
I never use Ctrl Tab unless there are very few tabs open—the limit is probably 5 tabs.
I must also admit to liking the way that IE opens link-opened tabs adjacently. There’s probably an extension that makes that happen for Firefox, but I’ve never looked.
August 22, 2008 at 3:47 pm
I have 2 ways of thinking when I switch tabs.
1. Go to the next tab (to the right) or the previous tab (to the left). The right-to-left ratio is probably 5 to 1. This is what I want to do 95% of the time that I switch a tab.
2. Switch between 2 (or 3) specific tabs. For example: webmail and topic I’m writing about in the email. Another example: new blog post and 2 or 3 sources. This probably consumes only 5% of my tab switches.
August 22, 2008 at 10:59 pm
Well, to be totally honest, I feel that the shortcuts for tabs work very well for me. I do however agree with the points in your previous posts about a visual tab change. But I don’t like the idea of changing between tabs to do with the most recently opened. I like flicking right and left…not older and newer.
August 24, 2008 at 2:44 am
For me, it’s mostly spatial — I have the “long term” tabs towards the left, the “browse” tabs towards the right, because that’s how they open. So for example, Gmail and LJ is generally on the left. Because of the Olympics, I also have the medal count tab towards the left.
When I browse a forum or read something with links, I just middle-click all the interesting-sounding links. When I have 5-10 piled up to the right, then I go one tab right (in other words, the earliest-middle-clicked), and read through it. When I close that tab, it automatically jumps to the next one. So on and so forth until I’m done. I might have 2-3 tabs still open because the info is cool, so I’d bookmark it or decide to keep it open longer (in other words, relegating it to the “left” pile)
When I need to do something new — start a new search or something, I’d open a new tab, and everything that results will be to the right of that tab because of the “recent-opened-goes-right” phenonmenon. So when I’m done, I just close everything to the right of, say, the initial Google search.
If I want to go back towards something that I *know* is open — as in, something I’ve already browsed and decided to keep open, I’d look towards the left and look at the page titles.
In my method of tab using, sometimes I’d run across new tabs that I’m like “What is this, and what was I reading that linked to this?” Other times I’d run across old tabs that are like “Oh yeah, I remember this. I forgot how cool this is.”
So I guess I’m spatial, but it’s because firefox currently associates space with a sequential time order.
August 24, 2008 at 3:43 am
Oh, forgot to add–
- when picking tabs from the row, I use my mouse and click.
- when just browsing and closing tabs, I cmd-w
- when I’m working through a bunch of open tabs from following links, I don’t really look at the tab bar at all. I mostly do so to find old tabs.
- there are times when I want to switch to the next tab over/back, but I don’t use ctrl-tab because I don’t like how it actually switches the window, and that disorients me. I’d prefer that ctrl-tab scrolls through the tabs at the top, and when I release ctrl, the tab I selected is opened.
I know it’s weird — I’m mainly a visual learner, which means I remember things by their images and what’s written. (For example, the only way I can remember names is if I see them written). However, I don’t like having images associated with my tab browsing. I feel like it’s data that I don’t need– I already have a mental image when I see the name of a tab (a composite of colors and pictures, and the “gist” of the page) , I don’t need the computer to load up another one. On the other hand, this is why it’s so important to me to be able to read the first few words of a tab–because I assign a lot of meaning to them.
(This is sometimes problematic because aside from the “browse” tabs that I close in 5 minutes, I don’t like to scroll through my tabs).
August 24, 2008 at 7:25 am
Since I use the Tabs Open Relative extension, my tabs are always organized by relatedness. Any new pages that one page spawns are always next to it, and I never have to dig around to find them. More over, these groups of tabs are always grouped by historical order, with oldest being on the left. Finally, the spatial order reinforces my mental model. The only time when I don’t immediately know the location of a certain page is when there is a long list of almost-identical tabs. In that situation, I think thumbnails would help. Generally, though, that extension makes almost all of the problems of finding the tab mute.
There’s another extension that helps in this regard: Tab History (which preserves history for newly-spawned tabs). It means that I can always press Back, never have to worry about keeping extra tabs open just in case I need to get back to them, and context is always preserved.
So, as has been pointed out in several of these discussions already, if you make the tab order actually mean something, and minimize the amount of tabs necessary, all these different ways of remembering tabs will reinforce each other, and tab-hunting will be minimized. I hope that made some sense.
August 26, 2008 at 1:05 pm
[...] into it in MediaWiki markup. Pretty simple, but kind of useful. It also fits in with some of the ongoing discussion on Planet right now about how to improve the display and access of tabs. (Aside: I still [...]
August 30, 2008 at 2:13 am
Umm… Firefox already has Find. Why not just give it the ability to search across tabs?
September 1, 2008 at 3:53 am
Thanks for your feedback, people!!
Brent: The ability to search across all tabs was the first thing I worked on when I got to Mozilla, as a way of teaching myself extension development. I never polished it up enough to be releasable. But yeah, it’s a really important missing feature. I’m gonna reimplement it as a Ubiquity command — no need for added check boxes or search fields in the chrome this way.
September 2, 2008 at 1:12 am
I think spacially about tabs, but usually only when I’m closing them – “Close all tabs to the right of this one”. This is because I tend to have core sites (say a feed reader), then open up new interesting tabs to ‘read later’. I clean them up at the end – it’s not specific “this tab is next to this tab”, but rather big buckets – “these tabs I want to keep open, these tabs I will close after I have finished”.
“Site” isn’t fine grained enough – I usually have tabs open on the same site in different spots and I want to switch to a specific one – I think “this page/section on this site” rather than “this site”.
The favicons really help me identify tabs when I want to switch to one with my mouse.