Infrequently Noted

Alex Russell on browsers, standards, and the process of progress.

Comments for The Importance Of Chrome


All Google need to do is make their plug-in experience more convenient than every other browser's.

Sadly, the "bad guys" killed the plugin scene when they used early plugin mechanisms to install crap-ware on people's browsers. Today, plugins are a chore because you first have to wade through tons of security warnings and little yellow bars before you even get to the bit where the plugin starts downloading.

If Google used their continent sized brain and interstellar infrastructure to create a mechanism with which signed plugins could be installed in full user-confidence with no more than one click, they'd be winning.

If they could do that without the required browser restart, they'd be untouchable.

Instead of this...

  1. Visit Site.
  2. Notice expected content is not shown.
  3. Notice annoying yellow bar saying the site either needs a plugin or couldn't install a required plugin.
  4. Click multiple times to dismiss the yellow bar and all the associated security warnings.
  5. Wait for plugin to download.
  6. Finish what you're doing on every other tab.
  7. Restart browser.
  8. Use page with plugin.

You would have...

  1. Visit page.
  2. Notice that instead of the requested page you see a nicely designed page (like the default Chrome start page, which I love to bits) displaying some information about a required plugin, the fact that it is signed by a trusted CA and a single button to download and install it.
  3. Click once and watch the plugin install.
  4. Notice that the page you requested loaded automatically immediately after the install finished.

Wow: 4 steps instead of 8, with no "secret-handshakes" or popups. It's U.I. utopia.

by Stephen Martindale at
Alex, I mean the extensions. For me, what makes Firefox an indispensable tool isn't it's stability, or it's speed, but it's extensions: Web Dev Toolbar and Firebug for example. I either missed that completely when scanning over the comic, or there's been no mention of it.

I like Chrome, I like the out-of-the-way UI and the blistering speed, but I'll not use it until I can install functionality I want on-demand.

It revs independently of browsers, implements new features without users changing their day-to-day choices, and has a (frequently exercised) auto-update mechanism. In short, it’s even more cross-platform than a browser. A plugin absolutely can (and will) help the web evolve. It’ll be gears or it’ll be Flash/Silverlight, which, written another way is “open vs. closed”.

I know very little about Gears admittedly. However nothing you have said he turns me on. I, and many of us, don't turn on auto-updates cause more often than not, they screw things up. I would have a problem with any plugin/standalone product implementing new features without me being able to veto that update.

by Manuel at
Manuel:

The stats for auto-updating browsers and plugins give me some real hope that auto-updates do, over the broader population, work. Obviously enterprises want and need control over the process, but in the main, auto-update is a Good Thing (TM) and it works (assuming the update mechanism is non-evil and continues to have user trust).

Regards

by alex at
Dean:

So you're arguing that "people don't install plugins" except when they do? And that somehow more people will switch their entire browser than will install a plugin? Either way, plugins have less friction. More importantly, I think, it's more likely that google can strike distribution deals for a plugin (e.g., Dell) than for a replacement browser.

Regards

by alex at
I'm, saying that people don't install plugins. Period. Flash is the exception because it is *pre-installed* on most machines. I thought you knew that.
Dean:

As I understand it, flash did not achieve the majority of its ubiquity through distribution deals. It got there because people wanted to see the whizzy animations, games, and other stuff they needed flash to.

by Aaron at
If Chrome doesn't have a plug-in capability, then it will simply be a two-bit player. The one thing that really separates Firefox from any of the other browsers, and that elevates it from 'good secure browser' to 'indispensable tool', is the plug-ins available for it.

Firefox without plugins is great, but not really anything special. Firefox with the plugins I want is light-years ahead of the competition, because it's the plugins that make it relevant to me. Chrome needs an open plug-in architecture.

Matt:

Not sure I understand your comment...the Chrome comic talks a LOT about their plugin infrastructure, and the plugin API across browsers is pretty much settled science. Are you referring to Firefox's lighter-weight extensions mechansim? If so, that is very much up in the air as far as anyone can tell. Should be interesting = )

Aaron:

Adobe had a distribution deal with Microsoft to include Flash with Windows up until a year or two ago (Vista and OEM XP's no longer include it, but the bootstrap is done).

Regards

by alex at
Brad:

Obviously, since you're at Google you can't just come out and say this, but I certainly does seem weird that the folks at Mozilla don't just ship Gears (and pull the same "hot patch IE if installed" trick I suggested that Chrome use). They are 100% aligned on the merits, and if Mozilla still wanted full control over the content author experience, they could simply try to create a separate distribution channel or forge a distribution deal for Gears as they have for the browser in general (say, for every Gears install via FF that also patches IE, some extra dough for MoFo).

It's insane that MoFo isn't chasing down all the angles here.

Regards

Update: changed the language to more accurately reflect what I meant to say.

by alex at
Just to add what Alex wrote about Gears, if you notice Gears and HTML 5 have been revving very fast against each other. Not only can Gears get new features across the existing installed base of the web (including IE), but those features are migrating into small standards that are vetted by the wider community. For example, Gears 0.4 just launched with a Geolocation API. The team then worked to write this up into a short spec, submitted it to the W3C, and its being revved by several others. Firefox 3.1 is planning to implement this spec, as far as I know. This is a good example of Gears, HTML 5, and web browsers working together to rapidly innovate and do standardization work (plus get actual shipping code for developers to use!).

Best, Brad Neuberg Google Open Web Advocate

An interesting thing about Flash and users installing plugins: even though Flash has an autoupdate mechanism, I've heard from Flash folks that most of the version upgrade happens almost entirely by users choosing to upgrade! This happens because content authors choose to use the new features, which drives demand, which drives users to flip their plugin version faster. So good content and user demand can cause plugins to have growth.

BTW, much of the 'received wisdom' about plugins was created in the late 90s, when bandwidth was slow and users were actually more conservative. Users in general are much more comfortable with the web, machines are faster, and bandwidth is much easier to download a plugin that might be several megabytes. In addition, the received wisdom was that plugins == proprietary in the 90s, which was true when they started. However, open source plugins like Gears changes this equation. If plugins are an effective way to help induce change on the web, why not use them to get some more movement going?

Best, Brad Google Open Web Advocate

It's not the "web app developing world", it's the consumers. People don't upgrade their browsers. And people don't install plugins. I wish they did.
*Flash is the exception and might be the solution. But it's a horrible solution.
Alex, what can Gears really offer us? It does a couple of neat things but a plugin is not going to help the web evolve.

The new gBrowser sounds great for HTML-based applications though. ;-)

Dean:

Think of Gears as a platform for evolution, not an instantaneous set of features. It revs independently of browsers, implements new features without users changing their day-to-day choices, and has a (frequently exercised) auto-update mechanism. In short, it's even more cross-platform than a browser. A plugin absolutely can (and will) help the web evolve. It'll be gears or it'll be Flash/Silverlight, which, written another way is "open vs. closed".

As long as the web app developing world continues to see "browsers good, plugins bad" as the M.O. for progress, we'll continue to be deeply disappointed (as we have been by Opera, Safari, and Firefox). I know that's an incendiary thing to say, but take a step back and think about it cooly. Yes, competition gets us a lot, but not nearly fast enough.

Regards

by alex at
Stephan, That idea was originally what IE wanted with ActiveX. Now, look at all the problems it's causing Microsoft and Vista. The real problem of plugins is that they are often used for trivial features, ways that don't really deserve to be part of a plugin. I mean, who cares if your cursor can be changed at a whim, or that the background of a page can vary based on your MP3 list?