Bit-cycling : Internet Explorer

Little things you may not have noticed in IE8: Part 7 – Stop and Refresh buttons

justsean — Thu, 11 Sep 2008 18:05:00 GMT

Move stop-and-refresh back to where they were in IE6

This was going to be my first post in this series, but the Vista Team blog stole my thunder with a much more detailed post covering lots of the chrome customizations, so I left it for last. :) Be sure to check it out the Vista blog.

PS. A quick reminder: I don’t work on IE, and I haven’t for nearly a year. So this all qualifies as my personal opinion, and not the opinion of the IE team. I found all of these feature just by playing around with IE (okay, well, some of them I knew where to look -- but I didn’t necessarily know they had made the final release)

Little things you may not have noticed in IE8: Part 6 – Authenticated RSS

justsean — Wed, 10 Sep 2008 18:33:00 GMT

Authenticated RSS Feeds

When we first released the Windows RSS platform a number of people asked us about authenticated feeds. As I said in those posts, it was a hard cut for IE7, but I’m really happy to see that it’s here for IE8 :)

To set up an authenticated feed, just subscribe to it, and then go to the properties for the feed (on the right in the feed view), and you’ll see the “settings” button for the User name and password. Click, fill in the details and go.

Of course, anything that uses the RSS platform will get this new feature for free, including the Feed Headlines gadget that ships with Vista, and Windows Live Mail. Outlook uses its own download engine, so this won’t help if you were trying to use Outlook.

Trivia o’ the day: One of the key drivers for making this work was the Web Slices feature. It turns out there are far more web pages behind authentication than there are feeds, and since Web Slices are implemented using the Windows RSS Platform, it was natural to implement support for authenticated feeds for all types of feeds.

Little things you may not have noticed in IE8: Part 5 – Tab duplication

justsean — Tue, 09 Sep 2008 18:23:00 GMT

Clone your current place

One of the nice things about the “New Window” (Ctrl-N) feature dating back to early releases of IE, is that it always duplicates the current open window, as well as all of its state – in particular, the window’s history.

When the “New Tab” feature was created back in IE7, I remember discussing with Aaron how IE could recreate the New Window behavior for Tabs. Unfortunately, it didn’t make the cut for 7, but I’m thrilled to see it make it into 8 (along with a ton of other tab-related new features).

The valuable thing about Duplicate tab (Ctrl-K) is that it allows IE8 to support both of the two most common search-and-discovery patterns on the web. The two patterns are as follows:

Breadth-first
Let’s say you’re searching for information on IE8. You head up to your address bar and type in “ie8” (which, we’ve learned, will take you to your search engine). Looking at the results, you see a four or five likely sites, so, using Ctrl-click (or middle-click), you open several of them in background tabs. Then, once you have several candidates open, you switch over to them one-by-one to explore them.

This model of search-and-discovery is fully supported by the tab features in IE7 (and other browsers) today. But, there’s another model.

Depth-first
Interestingly, prior to tabbed browsing, this was the predominant model for finding information (using New page). Let’s say that you are searching for dinner menus at cuban restaurants. So you do a quick search for “cuban restaurant menus”. You then click on the first likely result, and navigate through the site until you find the menu. Now you’ve found it, you want to go back to the hub and find another, without losing the page you’re on.

This model is supported by the new Duplicate tab feature. With this feature, you hit Ctrl-K to duplicate your current tab, then hit back to get back to your search result page, and click on the next link to explore. The page you were on is left alone on its own tab until you are done with it.

The depth-first model is used when you are looking for specific pages deep within a site that may not be easily categorized at the top level. Prior to IE8, it wasn’t possible to do search-and-discovery using tabs.

This small feature reflects a theme of IE8 that should become clear over time: There are a lot of great major features which you’ll read about blogs and the news, but deep down, you’ll find that what the team did during planning and development was to research and hone in on common patterns of using the web and build in features – large and small – to make those patterns easier or more refined.

Little things you may not have noticed in IE8: Part 4 – OmniBox™

justsean — Mon, 08 Sep 2008 18:02:00 GMT

Use your address bar for searching

A little-known fact is that IE has always had a combination address-bar/search box (this goes back to at least IE6). Well, to be fair, I thought this was fairly well-known until reviewers starting going nuts over the Omnibox™.

If you enter a single word with no spaces, IE has always tried to find a site with that term, and then, if it failed to find one, it went to your search provider. But if you entered two or more words, it went straight to your default search engine. A little-known way to do single word searches quickly is to start typing into the address bar with a question-mark and space (as in the example above), which forces IE to always send your text off to your search provider.

Some people wonder why a separate search box is necessary. The search box’s primary value is to provide a prominent reminder that the browser can help them with the common internet task of searching. From a design point-of-view, however, the search box allows the browser to assume that the primary intention is searching for new things (with a secondary intent of searching for something you’ve searched for before). In IE8, because of this knowledge of user intent, the browser provides several features optimized around internet search, ranging from the quick pick feature to visual search suggestions. The address bar, by contrast, is optimized around navigation – getting you to sites that you know about already, or have visited before (so the drop-down includes your typed URLs, your history, your favorites and feeds).

In IE8, it was important to the team to continue the tradition having a unified address bar/search, but also to enable address-bar-searchers to get access to the cool new features that were being developed for the search box – like Visual Search Suggestions. The UX team built a great shared component known during early development as the GUP, or Grand Unified Picker (Bruce, the development lead who named it in jest, was unhappy that the name stuck, so he eventually forced it to be renamed. :), which allowed both the the address bar and the search box to provide a single drop-down experience across a number of different data sources – some local (like history) and some remote (like visual search suggestions), with a mix of text and images.

The IE team has pulled this off this complex piece of work beautifully. As noted above, the drop-down in the address bar is optimized for navigation scenarios. But, if you start typing in the address bar with a question-mark, IE knows immediately that you intend to do an Internet search, so it immediately swaps in the search-optimized drop-down (which also includes history search after the search-engine supplied suggestions). [Note: it should also do this as soon as you hit two words in the search, but it doesn’t do this in the current Beta. I don’t know if it will be addressed for the final release].

Bonus tip:
The address bar also includes built-in reminders of, and quick access to, some common navigation shortcuts. Click on the arrow at the bottom of the address-bar drop-down to see them.

Can’t resist section:
OmniBox™ + AwesomeBar™ == Smart Address Bar :)

PS. Chrome’s Omnibox gets one thing right – the autocomplete behavior. I’ve said before that IE’s autocomplete is one of my favorite features, and I’m sad that it’s missing from this beta of IE8. They also have a hard-to-use (but visually compelling) implementation of keyword search shortcuts similar to something we planned early in IE8, but which was one of the early cuts (IE has had this feature at least since IE6, but it requires delving into the registry to enable).

Little things you may not have noticed in IE8: Part 3 – Feed Search

justsean — Fri, 05 Sep 2008 18:14:00 GMT

Searching the future

This morning I wanted to find out where I read about that “Skymarket” initiative. Off to IE, type in “skymarket” and it shows me not just Long Zheng’s original post, which I had read but also Joe Wilcox’s commentary, which I hadn’t read yet.

What happens is that IE is searching across not just my history, but also across the titles of feeds and feed items – so I can find not just what I’ve already seen, but also the new things from people and sources I find interesting.

A little interesting side-note: the ability to search favorites, history and feeds is provided by the Windows Search indexing platform, which is really cool. It really showcases the value of having fast, efficient indexer built-in to the platform. Without it, the team would have had to install its own indexer – and having two indexers running on the same system is a recipe for performance problems.

Little things you may not have noticed in IE8: Part 2 – Formatted View Source

justsean — Thu, 04 Sep 2008 18:00:00 GMT

Formatted View Source

You don’t have to be a pro developer to love this one. Since the very beginning, IE has relied on the built-in Notepad editor to show page source. Notepad, it should be noted is not a source viewer.

In IE8, for the first time, the IE team itself took on ownership of the core development experience, making a deep investment in built-in developer tools. In the past, the developer experience belonged to the Visual Studio team (and the professional developer is still well-served by using that product for end-to-end development), but with the experience with the intern-developed Developer toolbar in IE7, the team realized that there is a place for a lightweight, built-in set of development tools.

A great side-effect of the deep work is that the team developed a great source viewer, and hooked it up as the default “View Source” app.

Bonus tip: If you have a better app (Notepad++ has been my standby for years) for source viewing/editing, don’t forget you can always change the editor to anything else by using the Programs tab in Internet Options. For my needs, however, I don’t think I’ll need anything more.

Little things you may not have noticed in IE8: Part I – Search Quick Pick

justsean — Wed, 03 Sep 2008 11:30:17 GMT

IE8 has a lot of big new features (see the IE blog for more of that), but what I like about it is the small things. This is the first in a short series of my favorites little features that I haven’t seen mentioned much (or at all) elsewhere:

Search Engine Quick Pick

I am so thrilled with the way this turned out. Early in IE8 planning, we spent a bunch of time looking at how people use search engines. One key observation was that users tend to have one (or maybe two) “main” search engine (say, Google, Yahoo or Live), and any number of “vertical” or specialized search engines.

We wanted to make it as easy as possible for users to be able to switch between the main search engine and these vertical engines. Thus, the Quick Pick menu was born. I still remember sketching it out roughly on a piece of paper in Sharon’s office (Sharon is the main Program Manager for the search features in IE8), and then seeing it mocked up by the awesome designers (Ben is the man) a few days later, adding in the “Find” button – which solved a common discovery problem we had with the Find dialog in IE7.

The team executed it perfectly – it takes two clicks to re-execute any search on any of your sites. Even more important, the site you choose doesn’t persist (i.e. your main search engine will be the default again in a new window). This is because Quick Pick is designed to help you search those specialized search engines, and you don’t want Amazon or Youtube to be your default, just because you happened to search on it once (ahem, other browsers :).

Bonus feature: Watch what happens to the text box if you do a search using the search box on Live search, and then, on the search results page, you change the search term using the box on the page. It works with any search engine (as long as the engine doesn’t have redirects in the query URL like, for example, the Ebay one). The idea behind this feature was to support a more complex variant of the pattern described above (searching on multiple engines): (1) search on A using the IE search box, (2) refine the search a couple times using the search box on the page, then (3) two clicks in the IE search box to search somewhere else using your refined search term.

I guess they don’t like us anymore

justsean — Thu, 28 Aug 2008 20:25:13 GMT

I don’t think Yahoo likes us anymore. IE8 Beta 2 ships, and they start pushing Firefox on their home page.

Here’s the funny thing. They didn’t really want to work too hard on the page. Here’s the page they used for IE7, followed by the page they are using to push Firefox 3:

(Click the images to go to the live pages)

Cake Delivery Services R us

justsean — Wed, 18 Jun 2008 23:42:01 GMT

Yesterday morning I got a call from my old buddies on the IE team asking me to help out with a cake delivery. I was happy to help out, and Gerry’s Cakes (www.gerryscakes.com) in Menlo Park was able to pull together an awesome 3D cake with only a couple hours notice. Mozilla’s offices are 2 minutes away from the MS campus in Mountain View, so it was a short drive over to drop it off.

Al Billings wrote up the delivery on his blog, and it got picked up on Techmeme this morning (and on Lockergnome, Ars Technica, and Gizmodo – read the comments, they are hilarious).

Apparently though, according to those in the know, all this pales into comparison, when, earlier today, the cake attained the official status of “dope” earlier today when Trent mentioned it on PinkIsTheNewBlog.com.

Former Intern makes Good

justsean — Thu, 07 Jun 2007 15:24:26 GMT

The tall white dude in the middle of this picture is none other than Chrix Finne of IE RSS team internship fame.

Picture courtesy www.harvard.edu Harvard Class Day, June 6, 2007.

You may now commence to make the obvious Clinton and intern comments.

On names and codenames...

justsean — Fri, 09 Feb 2007 20:02:57 GMT

The naming of products at Microsoft is something that never ceases to fascinate the world.

The latest thing to get the blogosphere's proverbial knickers in a twist is the re-renaming of Windows Live Mail to Windows Live Hotmail. Well, guess what... I actually have no opinion on this re-rename. Why? Because while Live Mail is orders of magnitude better than Hotmail Classic, it's still not overly imaginative as a product. So, until the product is something I can really get excited about, it doesn't really matter what it's called.

Note: I did have a very strong opinion about the renaming of Windows Live Local to Windows Live Maps.I won't go into the details, but let's just say I'm a lot happier with the new name.

Anyway, on to my real topic: codenames.

Over on InsideMicrosoft, Nathan Weinberg wrote an post recently, titled "Sinofsky's on to something: Codenames must die." He discusses the "problem" with codenames in some detail, and proposes a solution:

Microsoft has this problem all the time. People are still referring to Windows codename: Longhorn today, even though the final name of Vista was announced 18 months ago, and that Vista is nearly a completely separate project from Longhorn. Windows Mobile 6.0 is probably going to be called “Crossbow” for the next year or so. Many analysts and Microsoft employees have complained of products that have better codenames than final brand names. The next version of Windows is on its third codename, first BlackComb, then Vienna, now Windows 7.

I honestly don't understand this railing against codenames. Who cares? So what if the digerati are still referring to it as "Longhorn" or "Crossbow." It really doesn't matter.

How about the people who still refer to Mac OS X 10.4 as Tiger, and to the previous one as Jaguar? (In fact, Apple had some fun with the box art for Mac OS X 10.3). Or what about the code name for OS/2 3.0, "Warp" (from a long history of Star Trek related OS/2 codenames), which eventually became a part of the product name.

Do you really think, for example, that if Vista had been called "Windows 6" during the development, the problem that "Vista is nearly a completely separate project from Longhorn" would have been avoided? We would still have had PDC03 at which we told people about the amazing new things in Win6, and then we would have dropped it all and replaced it with new amazing thing by PDC05. Nothing about the codename helped or hindered in that scenario. And for what it's worth, Longhorn was the codename for Vista for a long time after the event known as the "Longhorn reset" (all we did was start referring to the version prior to the reset as "Alpha Longhorn."

In fact, the codename thing can often be a savior, because if too much "baggage" gets attached to one codename, you can drop it and start fresh. This happens over, and over again. Blackcomb became associated in people's minds as "the next big release" and teams would start planning the grand new projects they would deliver in Blackcomb. So jettisoning the codename and picking another was the only sane idea.

Or, how about Kodiak -- the version of Exchange that would switch to using Yukon (the product eventually known as SQL Server 2005) as its datastore? That name was tightly focused on the SQL Server datastore project (not the least because Kodiak is a town in Alaska located fairly close to the Canadian territory of Yukon), so being able to jettison the name (and the associated baggage) and move to another is a good thing. What if it had been called E8 -- would that have helped, or would Exchange be getting the same heat that Windows gets about the difference between the early visions of Longhorn and the later visions of Longhorn?

In short, it really doesn't matter what you call your product. No matter what you name it, it get a set of attributes attached to it as visions turn into plans and plans turn into code. When we rename Windows 7 as Windows Coffeetable (you heard it here first, folks), there will still be tons of people calling it Windows 7. It's just a habit. 2 years later -- it won't matter any more. So, just pick something, and move on.

As a minor anecdote, a short time after we started working on IE7, there was an effort afoot to come up with a codename for IE7. In particular, we wanted a codename-series (a series of codenames related in some way: Whistler, Longhorn, Blackcomb (Windows); Platinum, Mercury, Titanium (Exchange); etc.).

Eventually, we settled on beach-related codenames (because that's where people surf -- get it?), and the name Rincon was selected for IE7. Rincon is a famous surfing beach in Puerto Rico, which has a famous road leading to -- the roadsign is here, so you can see for yourself why it was attractive. :) I believe Tony was responsible for that one. Random trivia: the codename for the version after Rincon was going to be Maui (this was "decided" a year before IE7 even shipped).

Well, as it happened, it never stuck. IE7 had been used internally as the name for so long that we never really got any momentum behind it. I offered to be the champion for the name, but only as long as management agreed to take the team to the associated beach when each product was done. That one didn't fly.

Oh, well. At least IE is already ahead of the game on the current "productname + number" naming scheme trend that Sinofsky favors.

Atom content and summary and IE7

justsean — Fri, 12 May 2006 14:58:00 GMT

I took a moment to point IE7 (on Vista) to this feed.

Looks good to me. :)

Reminds me -- I really need to post that documentation on how the RSS platform handles different feed formats and different elements of different formats. I'll see if I can get that done this weekend.