I was reading HTML5 for Web Designers and was reminded again of the interesting debate about autofocus and keyboard shortcuts. If you’re unfamiliar, autofocus is a JavaScript technique for automatically putting the cursor in a particular form element when a user visits a web page. This happens when you visit Google: notice that you can just start typing without manually clicking in the search box first.

The concern that many people have is that autofocus impedes the use of certain keyboard shortcuts. Awhile ago on Twitter I noticed that Juriy Zaytsev (a prolific JavaScript blogger and Prototype.js core developer)  said that autofocus breaks the backspace button, which is a shortcut for back. In the aforementioned book I’m reading, the author suggests that you should think long and hard before using autofocus, since it breaks use of the space key to scroll down the page. In both cases, autofocus causes these keys to type or delete in the selected form element instead of executing the shortcut’s intended function.

It’s a shame that these prominent folks are dissuading people from using autofocus. Autofocus really is a time-saver. The real problem is bad keyboard shortcuts.

Keyboard shortcuts that are too likely to confuse the user and produce the wrong result are a bad idea. I’d argue that using backspace or spacebar as a keyboard shortcut is almost as harmful as using p or b. Keys that indicate that the user may be trying to type something should not be used as shortcuts. Unless paired with one of alt, control, option, etc.

The good news is that although there are some very common but very bad keyboard shortcuts available, you don’t have to use them. If you want to stop being bothered by autofocus, I have found that Firefox, Opera, Chrome, and Safari all support command-[ as an alternative to backspace that works even if a form element has focus. Further, fn-downarrow is a good alternatives to spacebar for scrolling down.

Start using good keyboard shortcuts! It'll keep you sane, and let web designers make use of helpful features like autofocus!

[image credit]

Note: All keyboard shortcuts tested on a Mac only. Though I assume in Windows you have similar options.

I like the idea behind the newly-announced Chrome Web Store. It’s a store for web apps, which technically aren’t different from any other web sites like Google Docs. At first it feels like Google is just out to make a quick buck by mimicking Apple’s successful App Store. However, I think something more subtle is going on here. This is a a totally different approach to making money on the web.

Every now and then I think about making money from my side projects, but I’ve never done it. Why? Because making money on the web is called “ads”. If you’re selling t-shirts you probably don’t need ads, but if you want to make a website that’s fun or useful on its own merits, you don’t have a lot of options. Google makes 99% of its money from ads. The problem? I hate ads.

Ads pit you against your users. A user’s goal is just to use the app and have fun or perform some task. An advertiser’s goal is to get impressions. Do you hate the new ad overlays in Youtube? We all do, but Google has to pay the bills for Youtube to be free. Ever clicked a link, only to discover it wasn’t a link at all, it just triggered an irrelevant ad pop-up? They generate impressions. It’s a bait and switch, but someone is willing to pay for it, and someone has to write a check to Bob’s Discount Hosting at the end of the day.

Google has always known that people hate ads. That’s why they’ve put so much work into making ads painless. Their text ads are low-bandwidth and unobtrusive. Now and then, they can even be useful. That’s why Google is the clear leader in web advertising market share. Though now Google is asking themselves a few questions. What would the web be like without ads? And where would Google fit into such a web?

Are there websites don’t sell a more tangible product, don’t have ads, and exist only to be fun or useful? Sure. I made one, but the cost to run it is negligible. What would happen if it started generating millions of views a day, and my hosting company raised my rate? I’d need to take the site down, or find a way to make money. Inevitably, this would mean slapping ads on it. Yuck. Tons of interesting startups live in the same bubble I live in. Let’s just make a useful or fun app, and worry about money later, and only if we have to. With venture capital, you can live in this bubble for a long time. Sooner or later, though, reality sets in. Does this mean Twitter is going to have ads some day? Well, yes. They’re wrapping it up in a nice package, painfully aware of how much users hate ads, but that’s what they’re doing. That’s what you have to do.

This is why iPhone development is so appealing. You can make money without ads. Your app is just your app. I don’t understand in the least why Apple thinks creating an ad platform for iPhone apps is a good idea. They must underestimate how much people hate ads. I cringed when Steve Jobs acted genuinely excited when announcing this stuff. iPhone developers will continue to make it big when they make good software at a good price point. It’s such a better way.

As you think about all this, eventually you have to ask, why don’t web apps work like iPhone apps? Why are the choices “free” or “ads”? 37signals is pretty vocal in saying that you have to charge for your product. That works pretty well for the small-business collaboration and project management software that 37signals makes. But it doesn’t work as well the smaller the app is. On the iPhone, you can sell a nifty unit converter for $0.99. On the web? Not a chance. Google does it for free, as do countless other sites. Even if you have the best unit converter on the entire Internet that you spent weeks on developing and doing usability studies and designing gorgeous graphics for. The Internet just doesn’t sell web apps in that price range. You come across as a greedy jerk, even if you have a great product.

Can that be changed? Can we make it so that nifty bite-size web apps can be sold for a buck or two? How do you change it? Well, here are my guesses as to what you’d have to change:

• Safer and faster payment system. People hate putting in credit card information on some unknown website. They’d rather just pass than take the risk for a small app that isn’t such a big deal after all.
• Sense of ownership. A Tetris bookmark on a toolbar doesn’t feel like a product. It feels the same as all the free Tetris apps out there.
• Pay once. Periodic billing doesn’t make sense for small apps. What is the periodic equivalent of a $0.99? $0.05 / month? It’s silly. But still, not a lot of webapps work on a pay-once system. Why not? Well, at that price point, wouldn’t it be easier to just throw some ads on there? A pay-once model for apps should be simpler to set up.

I think the Chrome Web Store is an attempt to fix these problems. You buy apps quickly and safely through Google, you get a nifty apps tray in Chrome that makes it feel like you actually own something tangible, and you only pay once.

I don’t think that the Chrome Web Store is the future of web application development. On the other hand, I do think Google’s efforts will make it more acceptable to make money by selling your web apps instead of by tacking on ads. Once people get used to the idea of buying access to a tic-tac-toe website, web developers will have an easier time selling such apps outside of Google’s ecosystem. And that’s great.

It makes me wonder what Google’s perspective is on this. They’re an ad empire, and here is their product supporting a different revenue model. Are they worried about what the web would be like without ads, or even, without Google? Would Google start charging for their own apps if things went this direction? You have to evolve to survive on the web, and maybe even Google needs a plan for survival.

I’m currently developing a game using HTML5 technologies. I always knew I’d never put ads on it, so I assumed it’d just always be free. Though now I have to wonder. Can I sell it for a buck? It’s starting to look like it might be an interesting experiment.

It feels like this is a long time coming, due to the popularity of Twitter and the existence of another competitor’s similar product with the same name. But this move seems to have confused a few people. What does this mean for Wave?

While I’ve heard some opinions that Buzz is the replacement for a failed Wave, I think the opposite is true. Displacing email with Wave is a long-term strategy for a daunting task. Further, Wave is still being developed and is only in the “preview” phase. It is too soon to judge its success.

Buzz seems to be more of a transition than a replacement. Completely switching from email (even Gmail) to Wave is nearly impossible in a world so reliant on email to do, well, everything. Buzz, on the other hand, does a better job sitting alongside email (quite literally) as an enhancement. It doesn’t replace email, it expands it to do a lot of what Wave does.

More people will adopt Buzz than Wave at this point in the game, and it will nicely teach them the new communication model. Wave won’t seem so foreign anymore. And don’t be surprised if you see Buzz integrated into Wave too. Switching to Wave will be much easier if many of your existing conversations come along with you.

I’m very interested in seeing how this plays out, since email is badly broken and it is time for the Internet to get an upgrade.

It is still in the works, but I’ve been posting to it as I work on it. It uses a couple interesting WordPress plugins to do the detailed classification: specifically, the fields and tag clouds for flavors, breweries, and beer styles.

The design is not mine, but my policy is to get the site running before spending a lot of time on that part. It’ll be a great site to do a design for when I want some new portfolio material.

Wow.  I get no commitment pricing with a 2-year commitment!  And as a valued customer, I get a discounted upgrade at a higher price.

It is like no one even read this to make sure it made sense, or is just doing a terrible job trying to make me feel like their miserable offer is a favor of some kind.

This survey is:

• Short.  From just a glance I can tell I am 1 click away from finishing the survey.  The start of a survey of unknown length that scrolls off the page is far less inviting.
• Relevant to my life.  How quickly games appear in the mail is something I actually care about.
• Showing interest.  It says “Hey, did you get that game we sent you?” rather than a less personal “Rate our service on a scale of 1 to 10.”

In addition to my being pleased with it, I’m sure GameFly gets useful data from these.

This is for a site that received 269 points (413 up, 144 down) and had no prior traffic.

Unique Visitors per Day

That sums up to about 17,000 visitors, while still receiving a few hundred per day.

I also have analytics on this blog, which was discreetly linked to from jsregex. This site receives almost no traffic, and is only occasionally perused by some of my friends. Here is the data:

Unique Visitors per Day

Still a fairly small number: looks like roughly 0.3% of visitors clicked through to my blog.

Of course, I also have feedburner:

RSS Subscribers per Day

So roughly 13% of visitors to this blog went on to subscribe via RSS. That is actually fairly good. It probably helped that my most recent article was JavaScript related. [Edit: Later data showed about double what is shown here]

So now I know. It was also fairly fun going through the referrals and finding all kinds of interesting sites, from social media mashups, to blogs in Japanese, Spanish, and other languages, to rip-offs of major sites like popurls.

And if you’re one of those new subscribers: hello there! I hope you find something interesting here from time to time.

It seems awfully simplistic for a framework, but it is serving the same purpose. How far can we go, using a predefined set of CSS classes instead of picking our own when we design?  I’m not sure yet if a whole site could be reasonably made this way, and I’m not sure anyone has tried.  But it is an interesting question: stylesheets are already reusable on a small scale (element selectors rather than class and id selectors) and it is fun to wonder if we can make them even more generally useful.

Crashed plugin:

Crashed tab:

Sure beats the whole browser crashing.

To experiment with the technique, I’ve put a pink skull pattern on this site.  It could use some tweaks, but I’m going to let it sink in for a bit before I decide what to do with it.

(Note: the site design has changed since this post)