AOL announced recently that the Navigator web browser will be no more. Navigator 1.0 started the web for most of the tech sector, and their success, and Microsoft’s response in 1994 gave me a ticket for a wild ride, working on IE 1.0-5.0 in the mid 1990s.

For a trip down history lane, check out the archive of most web browsers known to man.

Posted in web browsers | No Comments »

Rounding out this week of browser reviews: to be honest I can’t recall the last time I took a serious look at Opera - regardless of when it was, 9.02 is a much improved and simplified experience. The toolbars look sharp, the clutter and over-featured UI of previous releases is gone, and I felt invited to spend some serious time putting Opera through its paces.

The good:

• Opera has the correct UI hierarchy for tabs - they appear above the address bar, as conceptually the tab is the highest order UI element in any browser (I complained about this in my first review of Firefox). To take this all the way, tabs should be above the menus too, but there are some problems with following the model this far.
• The notes & links sidebars rock. I’m a sucker for good sidebars as I designed them for IE4 & IE5 when they were new (and I was young). Everybody else has stood pat, but the notes field says “hey, I know you need scrap paper” and the links field says “Dude, relax. Here’s an easier way to use this page as a launching pad to other sites”.
• Tight, lightweight, clean visuals. I never thought I’d see the day, but Opera has a cleaner visual design than IE. It’s not even close. They handle the new tab usability issue gracefully (no birth defect button).
• Lots of unique clever goodness. The ability to start where you left off, save sessions (your suite of tabs), undo close page (Holy shit!), hotclicking and more. These are mostly done well and at worst give you the vibe of thoughtfulness: these Opera guys have clearly thought about what’s annoying in the browsing experience and tried to fix it.

The bad:

• Shortcut keys - if you’re a market trailer, you must support the shortcut keys for the market leaders. No questions. It’s easy to implement, at least as an optional mode. Most frustrating: Cntr-H, which is history in IE and FF, closes Opera! Pressing Alt-D (Address-bar), my most used shortcut, resulted only in misery.
• Active tab confusion: its hard to tell which tab is active. There’s not enough de-emphasis of non-active tabs, and the active tab needs stronger outline in connection to the browser frame (FF does this well).
• Help! I couldn’t figure out some of their unique features, so plunged into help for an overview. Searches for “wand” & “transfers”, two labels for features in their UI, turned up nothing. I don’t expect much from help, but with the web and wikis, I expect basic coverage of anything in the UI. (Transfers = Download manager as I eventually figured out on my own). There was a welcome to Opera first run page, but I couldn’t figure out how to get back to it (Why isn’t it listed in the help menu?)
• Vestigial weird UI. My old complaint about Opera was that it tried too hard: too many features designed in well intentioned, but awkward, goofy ways. 9.02 is much improved in this respect, but there are still some weird spots. The find on page UI is all too happy to highlight every instance of the first letter you type (shocking), and then switches to selecting the first, and highlighting the rest (I couldn’t figure out why: first instance is enough 95% of the time). There are some other oddities like this (print preview is a mode, not a separate view, the scrollbar has a bright yellow mouseover effect), but fewer than previous Opera experiences.

In summary, Opera is sweet! (Download here) I preferred it over IE7 for its personality and moxie alone, but until they soften a few more rough edges, I’m staying with FF.

Posted in Design, web browsers | 21 Comments »

Something’s wrong if, after 5 and 2 years respectively, the two most popular web browsers deliver mutually low-key major releases. All the talk of how Firefox has revitalized competition in web browsers, the most used PC applications, has had little impact on this round of browser design. Everyone (Opera and others aside) is shoring up, not taking new ground.

Any review of these two browsers, efforts so similar in functionality and core design, has to be about details. There are differences, primarily in interface design, but also in fit, finish, and vibe.

IE7

Five years since the last major release, IE7 delivers on most of the innovations Firefox popularized: Browser tabs & RSS feed support the most notable. Kudos to the IE7 team for picking up the design baton and making some visible changes, and for those who criticize them, you have to catch up before you can get ahead.

But within moments of taking IE7 out for its trial run, there are noticeable visual mis-steps: the shiny polish of the grayed out back/forward buttons. The weird background gradient behind the tabs. The empty tab all the way to the right (its the new tab creator, but it looks more like a birth defect, given it has no icon, text or anything).

These are the kinds of details only UI designers would call out - but they add up in the minds of users. Its the details that create feel and vibe, and that’s where IE7 leaves me cold. I know many people worked hard to ship this thing, but their love and passion was hard to feel when using what they made. I live in a web browser all day, and like my living room, the details matter. My guess is the visual designers were tasked with being midway between Vista and XP (explaining the elimination of the command menus), a difficult middle ground to hold.

However, in the days I’ve used IE7 as my primary browser I found its workman like charms. It does what it needs to do - stays out of your way, and seems to have filled in, lack of polishes aside, the major functional gaps between IE6 and Firefox.

Innovators note: the quick tabs feature is interesting but mostly a lark. Thumbnails of web pages, as much as i tried, never seemed to help me do anything except demo IE7. The Zoom feature was nice, given my aging eyes, and improved phishing detection (which was hard to demo) and other security improvements seem to be a large part of their marketing message, but I had few comments on these anti-features.

Firefox 2.0

I did a short, underwhelmed review on their beta2, and the final release held few surprises. This is entirely a polish and plumbing release. They invested in infrastructure (installer, JavaScript 1.7, etc.) and some minor UI enhancements. The addition of spell-checking (a la MS Office red squiggles) seems minor, but is easily the underdog champion for best low hanging fruit feature in this wave of browser updates (Despite its clear value I have to lament: it’s 2006, the age of the blog, and all browsers don’t spell & grammar checking? Yikes).

While Firefox’s toolbar design is more conventional, it’s also more polished. The side gradient shading in the icons (look closely at the Reload and Back buttons) give it a warm, friendly feel. Notice how their tabs are easy to scan, uncluttered by text on blue gradients. They shifted all non-active tabs to black on gray, a simple way to make large tab sets less oppressive (though scanning non-active tabs is slightly harder now).

The summary

I’m still a dedicated Firefox user. While 2.0 is a conservative release, the product does a better job delivering on core ease of use and functionality: its an easy recommendation. Despite being a browser designer, I have mostly simple browser needs, and would recommend FF to just about anyone. If you factor in the vibrancy of their add-on community, it’s also a winning choice for power users.

IE7 is much improved - but visual design mis-steps, some awkward design choices, and lack of any compelling feature advantage makes it impossible to recommend it to anyone currently using Firefox. I’d recommend the upgrade to any IE6 user for the security improvements alone, but also for the benefits of tabs.

In general I’m left hoping hoping that both camps have their eyes on what browser user experiences could and should be like. There’s so much more user experience ground browsers need to cover. Lets hope IE8 and FF 3.0 build on their current foundations and take up the charge.

IE7 Download / FireFox Download

And don’t miss my review of Opera 9.02 - you’ll be surprised.

Posted in Design, web browsers | 6 Comments »

It’s been months since I’ve commented on Firefox, IE and the state of web browser design. I’m back: I recently installed Beta 1 of FF 2.0 and here’s a short review.

Beta 1 releases are tricky strategically: you wan’t to hold back on some big features so competitors have less time to recover, but you do want mileage and feedback on big changes. As beta releases go, this one is conceptually conservative. Especially since IE7 is late in the game, with a recent beta 3 release.

Highlights

• Easy install. Three screens and you’re in. Smoother than many final release installs.
• Auto spell checking in text boxes! A feature I tried to get into IE for years - stupid reasons prevented it, so I’m happy to see it now. (But it does hyperventilate on HTML editing (e.g. blogs) - and needs to get smarter, or have an optional toolbar button for toggling it off).
• Edge case experience improvements. FF remembers your session set up, tabs and all, on crash, and can recover. A sweet safety net.
• Beta Stable. You never know what beta means, but its held up ok. A crash an hour or so in most sessions.

Lowlights

• It’s a low profile release. Most of the work appears to be infrastructure: phishing protection, Javascript 1.7, new installer, etc. It’s good they’re striking at the root of the tree, but it’s not a user experience release.
• Microsummaries are weak. The idea is odd but interesting - but its current design demands custom, browser specific work by content providers, so the feature ships in beta 1 in a functional coma (IE4/NSCP4 had dozens of good features die soon after this exact kind of birth). The docs suggest autosummaries, but doesn’t provide any. The big question is how dynamic titles impacts recall when looking at a list of bookmarks: will they re-alphabatize? Will they stop updating if I rename them? Questions abound. This seems like a solution in search of a problem and the spec is without a screenshot: a UI design red flag.
• Hard to discover what’s new. Any beta should offer a fun walkthrough of screenshots showing me what’s new and what i’m getting for rolling the dice with my computer. I looked for ten minutes, and the best approximation I found was over at Lifehacker. A checklist of unlinked features is lame -help me, as a beta tester, or blog reviewer, feel the love.
• Not much to play with. After twenty minutes I felt I’d seen what I needed to see. There just isn’t much to explore or tinker with (as a non FF-developer). So I’ll unistall and wait for Beta 2.

Some reviews I’ve read highlight the new History menu (replacing the idiotic Go menu - yay!) and its list of closed tabs. It’s a thoughtful gesture, but it’s a hacky, Microsoft-esque UI design, in that the real solution is a better tab close model, rather than a greasetrap that captures things after they’ve fallen.

But there are other UI problems with the history menu: it still colides the history tree with the back tree. Take a look:

These two snapshots show two different histories: one for the back tree, one for the history list? Why two? Not sure - probably because back/forward follows a pruning algorithm and the history list doesn’t. But now that there’s a history menu, the conflict is more obvious (or then again, perhaps only browser UI weenies like myself catch these things). The back button is king here, so I’d rationalize in it favor of whatever it’s behavior is.

Here’s waiting for beta 2. Working on trying to get IE 7 Beta 3 installed, so stay tuned.

Posted in web browsers | 1 Comment »

Someone kindly submitted my post to slashdot this morning, and it took the site down for awhile - apologies. The first two times I was slashdotted it didn’t generate anywhere near the traffic this one earned.

I wanted to clarify a few things:

• I left Microsoft in 2003 . I did work on IE for a long time (1.0 to 5.0), but the way the slashdot post was worded many assumed I’m still employed there: not true - I work on my own as a writer and consultant.
• There is a good parallel discusson on Asa’s blog. Asa works on Firefox and is one of the folks that deserves the praise for making an excellent piece of software. I responded to a few comments over there.
• Thanks for all the feedback and commentary. I doubt I’ll ever see this many comments for a single post again. It’s been fun - thanks for writing your opinions.
• On ui design. The mistake we’re all making, myself included, is focusing on designing for ourselves. Designing for ourselves isn’t a sin, but if the game you want to win is market share, you have to work very hard to make sure your needs and wants jive with people who’s needs are less sophisticated than ours (Which is most of the planet’s web browsing poulation). Lots of folks said “my mom can do X” or “my friends can do Y” as justifications of how their experience matches everyone elses, but I think we’d all agree how fragile and anecdotal those claims are. Your mom might be a rocket scientist, and your friend might have watched you do whatever it is before they tried to do it themselves. I’m not saying I’m right, you’re wrong, or that your pants are on fire. Instead I’m saying that design arguments, ui design arguments in particular, can and should stand on firmer ground. There should be an essay somewhere called “how to have a meaningful UI design argument” (finger on nose).
• On history trails, tabs and new windows. Folks pointed out at least a half dozen different ways both new windows and tabs are used. My stated opinion was narrow: apologies. But the concerns still strike me as valid (tabs make back/forward more complex, since there are now N history trails per browser window). I’ll need some time to read through all this, do some sketching, and rethink my stance.
• On open source and design decisions. In digesting all of this, my primary thought is “how does each of these opinions/complaints/usage patterns, as diverse and sometime contradictory as they are, fold together into shaping a single design that’s of the greatest value to the most people.” I’m familiar with the extensions and how the community encourages people to modify, create and involve themselves, which rocks. It’s an innovation pool for ideas that can be considered for the core (Some might recall the Win95 and IE powertoys, it wasn’t community based, but the idea was similiar). But I flinch at using extensions as the copout to challenges to the basic design. It’s a great plan B, but as a designer, for core parts of the experience, the obligation is to dig deep enough that plan A stands tall on its own. I’m not suggesting anyone said otherwise: I just wanted to make sure we’re on the same page. It’s surprising how few mainstream users, of anything, customize. All the data I’ve ever in my career (web,software, etc.) is on the order of 10-30%, and that 10-30% correlates with advanced, savvy, early adopter, industry types: e.g. most of you reading this. I posit that most people, for most things in life, live with the defaults (I mean this about software and life in general).

I’m still reading through all the comments, so keep them coming. I can’t promise to respond to all of it (150 comments and counting, plus e-mail) but I do promise to read them all.

Cheers.

Posted in Design, web browsers | 52 Comments »

It’s a sad day and a good day. For years I’ve held onto my IE install out of love. I worked on IE 1.0 thru 5.0, and was one of the people that designed much of its UI. But my love for the past has faded. Last week I switched to Firefox: and I’ve been happy.

Why I switched:

1. IE is a ghetto. There are specs I wrote for UI features in 1998 that are unchanged today, 7 years later, in a world where browser usage has changed dramatically. I’ve watched bugs that I fought to have fixed in 5.0 become regressions, appearing in 5.01 and surviving in 6.0. Even though it’s the product I was proudest of, using it now makes me sad - it’s been left behind. I do read the IE blog now and again - smart folks are working - but there’s nothing for me to install.
2. Bookmarks work. The Favorites UI model in IE is the same one we built in 1997, when we knew most of our users had 20-40 favorites. It was made to be super simple and consumer friendly as most of the population was still new to the net. This UI is effectively broken today, designed for people that don’t exist. The Favorites menu and Favorites bar show links in different orders, the organize favorites dialog is just weird, multiselect doesn’t work: favorites is a sad forgotten place. This was by far my greatest frustration with IE, even though I’m responsible for much of the original design.
3. Firefox has quality & polish. IE 5.0, for its time (1999), was a high quality release. Really, it was. Joe Peterson, Hadi Partovi and Chris Jones fought hard to give the team time to do lots of fit and finish work. We did fewer features and focused hard on quality and refinement. Firefox feels to me like what IE 6.0 should have been (or what i expected it to be after I left the team in ‘99). It picked a few spots to build new features (tabs), focused on quality and refinement, and paid attention to making the things used most, work best. The core UI design is very similiar to IE5: History/Favorites bars, progress UI, toolbars, but its all smooth, reliable and clean.
4. They made a mainstream product. One of the big challenges in designing software is balancing the requests of earlier adopters in the community, with the needs of the majority of more mainstream users. After playing with mozilla on and off I was afraid firefox would be a built for programmers by programmers type experience. It’s not. I don’t know who in the firefox org was the gatekeeper on features and UI, but I’d like to meet him/her/them (seriously). They did a great job of keeping the user experience focused on the core tasks. If you’re reading please say hi.
5. Security isn’t annoying. . The press makes security into such a huge deal, but I’ll be honest. I don’t want to think about security at all. I’ll do what I need to, but mostly I want the system to take care of it and stay out my face. Nothing in FF makes me feel safer explicitly, I just don’t deal with as many warnings, settings and other details. I know from the PR that security in FF is better (even if only because it’s less targeted by spyware, etc.) but I’m pleased that the product doesn’t remind me of how safe I am all the time.

Problems with Firefox:

I’m a UI design guy, so many of these are UI related. (Added note: I’d used FF on and off, but since I’m now 100% some of these are complaints might fade in a month of usage. Stay tuned).

1. Find UI. Why does the find dialog appear at the bottom of the screen? I agree that a dialog box (semi-modal) can be a mistake if you’re doing multiple searches, but flipping a coin for placement (top vs. bottom), the top is a better choice for any UI, especially if it’s going to look and act like a toolbar. I can’t move it so it earns a spot on this list. However, the overall implementation isn’t circa 1992 like the IE one. It highlights, it searches on type, & it warns on unfound items - nice..
2. Download UI. Here’s a case where modeless makes sense (it’s never my primary user task), but here we get a dialog box. My first crack at this would be a one line toolbar, much like the find bar, at the bottom of the screen telling me about downloads. That’s where all the other dl status info goes. Again, despite my nits, it’s an improvement on the ancient IE implementation (which we all hated forever too).
3. Tabs and new windows. Firefox goes against IE behavior and starts each browser instance from scratch. IE intentionally brings the browser history into the new window: the bet being that users who want to continue from where they left off can, and those that want to go their home page can do that with one click. Everytime I hit Cntr-T and see a blank screen I think I’m in Word. I use tabs less often than I expected: opening new windows is often more comfortable - easier to track which window lives where. With multiple tabs (I find) the back/forward behavior becomes complex and hard to predict. Strict UI logic would put the tab UI above the toolbars, not below, but that creates other problems.
   
4. Tabs and modality. The desired illusion of tabs should be to make each tab a virtual browser. Well this breaks when you bring up a modal dialog within a tab: you can’t switch to another tab. It’s an annoyance, not a sin, but when it happens it reinforces my new window habit, and slaps my wrist on my growing New tab habit.
5. The return of the go menu. It was with great pride that we killed the go menu in IE 5.0. It was the stupidest menu I’d ever seen, since it was never used and no one knew what it did. For accessibility it was necessary, but had no rights to be a top level menu (IE has View.Go). The Go menu was probably inherited from NSCP/mozilla, but it really should be put out to pasture. And if it stays, someone needs to explain why it shows a different history list than the one in the back button drop down.

For reference: I wrote about principles of browser design here: How to build a better browser.

(Update: I’ve responded to many of the comments in a second post.)

Posted in Design, web browsers | 307 Comments »

Thoughts on browser design, from someone who worked on IE 1.0 to 5.0.

How to build a better browser.

Posted in Design, web browsers | No Comments »