oreilly.comSafari Books Online.Conferences.


AddThis Social Bookmark Button

Browsers that Aren't Browsers

by Giles Turnbull

If you're one of those people who can remember the early days of the Web back in the mid-1990s, you probably have fond memories of how browsing used to be.

Grey backgrounds, beveled images, faux paper-effect background tiles. Ugh. Thankfully the Web has moved on quite a bit since then, and with it the browsers we use to extract information from it.

The Web is no longer about pages, it's about data and metadata. It's no longer sufficient to simply browse it; these days we are mining it just as much.

So it follows that if we're doing more than just browsing, we need more than just browsers. And that's exactly what's happening now -- there's a growing market for "web software" -- applications that use the Web, or allow us to get what we want from it, without having to browse it in the way we are accustomed to.

The most obvious popular web software is the rash of aggregators and feed readers that has become as vital as a browser for anyone wanting to keep track of lots of different web sites.

Thanks to the WebKit technology built into Mac OS X Panther (and the forthcoming upgrade, Tiger), it's become much easier for software developers to create applications that are web-savvy. In this article, we thought we'd take a look at a few of them.

Edit your Own Web

Webstractor is an ingenious combination of browser and editor that offers a unique set of tools, valuable for anyone who uses the Web for research.

The first smart thing it does is save everything you browse locally -- including images, multimedia, and page text. Your browsing history becomes a Webstractor "document," which you save just as you would any other document on the computer.

Screenshot showing Webstractor's browser Webstractor in use as a browser.

The second smart thing it does -- once you've saved a browsing session as a document -- is let you edit what you see and condense it down.

Using Webstractor is simple. Browsing is easy, and you all know how to do that. The controls in Webstractor are a little different from the browsers we all know and love, but there's nothing too challenging to deal with.

Having done your browsing, all you need to do is click checkboxes alongside the pages you wish to make editable. Everything you've just browsed is presented in a list -- you can, of course, return to anything in the list just by clicking on it.

Here's the really clever bit. Webstractor concatenates all the pages you've selected for editing, turning them into one large, editable document. Now you can start slashing out the content you don't want, and moving around the segments you do want. It's amazingly fast and efficient, and an excellent way of distilling an hour's worth of browsing into a single, read-in-one-glance page.

Screenshot showing Webstractor in edit mode Editing web content in Webstractor.

Why would you want to do that? Browsers are great for viewing content but suck when it comes to reviewing what you've just seen. Short of continually hitting Command-S to save every page you see locally, there's no easy way to go back over your browsing session. Webstractor is ideal for research projects with a specific goal. School children, undergraduate students, and paid-up professors will find this application great for compiling notes on a topic.

Related Reading

Creating Applications with Mozilla
By David Boswell, Brian King, Ian Oeschger, Pete Collins, Eric Murphy

Imagine also being able to use it to follow a thread. The latest big news story, or a meme crossing the blogosphere; with Webstractor you can follow the thread to its end, then go back and review what you've seen. It's a compelling and unique tool for OS X.

Completed documents can be exported as PDFs, making them even better for homework or research projects.

The editing process itself is delightfully simple, and feels much like creating any other rich text document in TextEdit. Of course, web pages are designed and built in different ways, but Webstractor does an excellent job of allowing the user to select bits of tables, border elements and other extraneous bits and bobs, so they can be removed easily.

What's interesting is how using Webstractor brings about different behavior. I'm one of those people who gets easily distracted and sidetracked, and I often find myself opening a new tab in my normal browser, and initiating a Google search about something trivial to run in the background while I carry on with the original task.

Webstractor encourages concentration. More than likely, you want your Webstractor document to contain information about one specific topic, to make later editing easier. There's less opportunity to go spinning off on tangents.

Should you wish to let your mind wander, the Links Inspector tool is a good way of making a page placeholder; it contains all the links from a specified page, and allows you to navigate your way round them without losing your way, or your train of thought.

There are limitations. Webstractor reformats documents on the fly, so if you remove a chunk of web design with the aim of just leaving the interesting text behind, you may have to wait a few seconds for Webstractor to complete the new operation before you can continue. It would also be fantastic if Webstracted documents could be exported in other formats, perhaps as .rtfd or .html. Some might also argue that it's just as easy to copy-and-paste everything from a web-based research session into a plain text document, and edit it there, but that ignores some of the smarter elements of the Webstractor offering.

Webstractor is more than just a tool for collecting chunks of text; it also offers context for each chunk, and a visible pathway of browsing history. In this last respect, only the new thumbnail view in OmniWeb 5.0 comes even close.

Few people are likely to switch to Webstractor as their default browser, but that makes perfect sense; because although it contains browsing features and familiar browsing tools, it's not really a browser. Webstractor is an information-gathering tool, a means of using the Web to compile a new document. It browses, sure, but it browses different.

Browsers that Lurk in Odd Places

Steven Frank's WebDesktop application is a fun hack that uses the WebKit rendering engine to display a browser on your desktop. While still in the background it is merely there for your eyes, but bring the app to the front and the browser becomes fully functional.

Because WebKit is an integral part of Apple's Safari browser, WebDesktop should display pages just as Safari would. The only difference is that they lurk in the background, not hogging your attention, but there when you need them.

Screenshot of Webdesktop running behind BBEdit Spot the browser 1: there it is, in the background, behind BBEdit.

This is a new kind of browsing. It's not really browsing web pages, but more like being aware of them. Again, this is by no means a browser designed to cope with your day-to-day browsing tasks, but an evolved browsing tool suitable for a subset of tasks.

The examples Frank includes on his web page speak for themselves: keeping up with the news or sports scores, tracking the final minutes of an eBay auction, or monitoring stock quotes. All of these are perfectly valid things to display in a browser, but not necessarily in a standard browser window. This is peripheral data, stuff that's useful to keep to hand but not vital enough to be using up too many brain cycles. If something changes, you want to know about it. Otherwise, it can stay quiet.

WebDesktop's neatest feature is hidden away in the preferences panel, reached by Control-clicking on its Dock icon. There you'll find an auto-refresh widget, which can be set to reload the background web page as often as every minute, or as rarely as never.

Web developers might find it useful as a way of keeping an eye on changes to code; a kind of live preview, almost like WYSIWYG HTML editing, but from a distance.

iBrowser is a sort-of WebDesktop in reverse. It's a kiosk-mode browser, designed to operate in full screen mode.

It can be set to run partly transparent while in the foreground (to reach the preferences, control-click on the Dock icon and select the (oddly) empty first option), with the result that you can browse merrily away while keeping a close eye on other applications running in the background.

Screenshot of iBrowser running in front of BBEdit and Eudora Spot the browser 2: this time, you can see right through it. (Note the toolbar at the bottom of the screen.)

iBrowser imports various bookmarks and settings from Safari, and inserts a menu-bar icon that offers further controls and access to configuration options.

Screenshot of iBrowser's menu bar controls The menu-bar controls offered by iBrowser -- move your pointer to the top of the screen to make the bar appear.

Browsers that Lurk in Other Programs

Since we're highlighting the laurels of WebKit, I have to talk about the developers' ability to embed browsing capabilities into their applications. Usually this is not because they want to create a whole new browser -- people have strong views about browsers, as we've seen before, so few users are going to switch to a new one just like that. But, rather, because a certain degree of browsing (or even just web-content rendering) might be useful for some users.

The 7.x releases of BBEdit, which included the ability to preview HTML markup being created in the text editor without having to launch a separate browser, are perfect examples of this.

In other cases, browsing brings new functionality to an application. How did we used to buy music before the iTunes Music Store came along?

And no self-respecting modern RSS aggregator ships without a built-in browsing element to make the task of absorbing hundreds of feeds, and hundreds more posts, that little bit swifter and easier.

Notepad application DEVONnote (DEVONthink's little brother) includes a WebKit-powered browser element for instant access to web content. And while it doesn't offer the power of Webstractor's edit-the-Web abilities, it can grab any chunk of text from a web page, and with a Control-click, turn it into a new note. (This only works with text - include images or links in your selection, and the capture system offers options to save the link as a new bookmark-style note, or open the image.) But then, if all you want from a web research project is words, this could be extremely useful -- and it works faster than Webstractor, too.

Screenshot of DEVONnote's browser displaying Here's as browsed in DEVONnote.

Like DEVONthink, DEVONnote offers an array of ways to save and sort information, including an outliner view, built-in search, and Wiki-style links from one document to another. The browser is clunky if you try and use it for browsing, but that's not what it's there for. It lets you view web pages if you need to, and it may make things quicker than switching over to a proper browser. Like many such implementations of WebKit, the browsing element is an extra plus inside an application, and not designed to be a browser in its own right.

And if that Isn't Enough for You

Enough browsers for now? If we've sparked your interest with this article, you may like to go delving further in the hunt for more sort-of browsers. There's dozens more out there.

And if even that's not enough, the only thing left for you to do is go ahead and build your own browser. Good luck!

Giles Turnbull is a freelance writer and editor. He has been writing on and about the Internet since 1997. He has a web site at

Return to