Don’t make browsers, make extensions

There’s been a ton of speculation regarding “gbrowser”, google’s alleged browser, Netscape’s Firefox based browser, now even thoughts Yahoo might be interested. Though I wonder if that really is beneficial to anyone involved?

I’m going to make the bold statement that custom browsers are bad, making extensions are good.

There are several reasons why custom browsers are bad:

  • Casual web surfers don’t always realize “Browser X, and Browser Y are custom versions of Browser A”. They see them all as different products. It’s confusing, especially when websites block them because of their UserAgent. “I’m not using ‘Netscape’, I use ‘Mozilla'”.
  • Anyone who distributes a browser is obligated to maintenance, statistically the vast majority in a project life cycle. Especially in regard to security updates. Get them out quick. It can sometimes involve some extra work, and has minimal benefit for the distributor.
  • Self-competition becomes a factor. One thing that confuses many people about switching to linux is the simple question of “what distro?” This question, and the inability to quickly make a decision turns many people away. Windows and Mac OS have the advantage of making it very easy.
  • Over branding. Yes there is such a thing. You put a brand in someone’s face for too long, and it loses it’s significance and impact. They overlook it. How many people actually notice a McDonalds when driving around? Most don’t even see them, simply because they are more common than traffic lights. Now how many notice less popular dining establishments? Quite a few.
  • Ineffective marketing. When you share 90%+ of the code, you share features with tons of others and really have very little to market. What you do have to advertise, is somewhat insignificant. Why download a new browser for a logo? Is that even a feature? Why can’t I just bookmark your page if I like it?

The Correct Approach
I personally believe the correct approach in this arena is extensions. A great example is the brand new Yahoo Toolbar, or SpeakEasy. Why are these the right way?

  • Both leave security to the Mozilla Foundation, users can get updates as soon as they are released, they don’t have to wait for the distro’s cobranded builds to become available.
  • Users get new features as the product is updated. Don’t need to wait for the distro to update the cobranded builds.
  • Users choose branding, can uninstall it if they wish.
  • Less downloading. I change from speakeasy to yahoo, I don’t need to download a new browser, just install the extension.
  • Cross platform. It’s much less work to support Mac and Linux users when you provide an extension rather than a custom build. Get the whole audience.
  • Lower cost. It’s much less development to release an extension rather than a browser.

What do they lose?
Really nothing. You can do pretty much everything via extensions. You can create a skin, add features, overlay menus, add toolbars etc. etc. There’s quite a few possibilities.

Conclusion
Releasing your own browser, unless you really make radical changes (Camino, Galeon, K-Melon) is somewhat of an ineffective use of resources. You can accomplish the same thing, while providing better service to customers by trying to use an extension framework. Extensions by nature have less development requirements, easier to update, allow the user to have the latest browser, and give the user choice.

I personally think Yahoo and Speakeasy have done an excellent job. They accomplished their goal and really addressed the point I’m trying to make in this post. I just hope some other companies will seriously consider what they are doing, before they try and get their users to install hacked up copies of Firefox.

Extensions and Themes are the best way to customize a browser. If at all possible, try to keep within those frameworks. You’ll thank yourself later when you realize that you need little/no changes to work perfectly with Firefox 1.5 or later.

6 thoughts on “Don’t make browsers, make extensions

  1. Pingback: Gemal's Psyched Blog

  2. Robert, let’s push this a bit further: the future of open-source development is probably a model where the “basic” kernel is free, and is usually enough for 80% of the users, while some extensions bringing a real plus to the 20% others are commercial.That’s almost the only way for OSS companies to avoid being an almost pure contracting company.

  3. Pingback: xslf.com

Leave a Reply

Your email address will not be published. Required fields are marked *