Top 10 Web Dev Blunders of 2008

It’s inevitable: We’re all going to screw up at least once when it comes to the highly volatile world of web development. Downtime, stupid UI mistakes, miscues on communication, bad product launches… they happen to the best of us. Literally. Over the past year, many of the top websites in the world have committed serious development no-no’s. It’s almost comforting to know that the best development teams in the world aren’t perfect either.

Instead of viewing this article as a way to bring down the big guys a peg or two, let’s look upon it as a teaching tool. Learning is all about making mistakes and finding ways to correct them. It’s nice when we can occasionally learn from other’s mistakes, instead of screwing up ourselves.

Here’s a list of the 10 biggest “Whoops!” moments of 2008 in web development.

10. Ecommerce Sites Down On Black Friday

Here’s every online retailer’s worst nightmare: A flood of people flock to your site begging to spend large amounts of money, and your site is down. Unfortunately, this past Friday it became a reality for many retail sites on the Web, large and small.

Black Friday is the biggest shopping day of the year in the U.S., and the same is true for online shopping. Many of the major sites like Amazon, Sears and Live Cashback were all brought to their knees, forcing shoppers to go elsewhere for their Black Friday shopping.

It turns out that this widespread outage is a perennial event. Black Friday comes every year on the same day, yet many online shopping stores can’t figure out that they need to have major backup in the server department. As online shopping continues to grow, this epidemic is going to increase every year, unless ecommerce sites can plan in advance for the traffic surge.

9. Onslaught of Twitter Clones

2008 appears to be the year of the Twitter clone. Twitter has take the web by storm these past couple of years, and consequently many web developers have tried their hand at making a clone of the popular web service. And it would appear that there isn’t an end in sight to these Twitter clones.

While making competing websites that are unique in some aspect (a la Plurk and Identi.ca), making an exact replica of a service is never a good idea. Nevertheless, many Twitter clones have sprung up the past year, and nearly all of them lack any differentiation from Twitter, and worse; nearly all aren’t as good as the original.

One should think long and hard about building a clone of a website. Unless there is a major difference between services, it’s highly unlikely that the clone will find any traction at all.

8. MobileMe’s Awful Launch

You know a product launch is bad when the head of the company admits that the launch was poorly done. Such was the case with Apple’s MobileMe launch, in conjunction with the new Mac operating system Leopard.

As Apple launched its much-anticipated new iPhone 3G, it was also announced that a brand new syncing service MobileMe was going to launch in tandem with the new iPhone 3G, iPhone 2.0 software and the App store. MobileMe is a syncing service that allows Mac and PC users to sync their contacts, calendars, mail, photos and more, all from a web interface.

Much of the first day of the launch was filled with slow performance and downtime, weird problems like random user logouts, and the nail in the coffin: The service couldn’t sync with calendars and contact entries for the entire first day. It was, as Apple CEO Steve Jobs put in an internal email, not Apple’s “finest hour”. Since then Apple has fixed their launch bugs and have credited all MobileMe users two extensions to their subscriptions totaling 90 days of free usage.

7. 30 Hours of Gmail Dowtime

Gmail is one of the most popular email providers, with many of the users from the Google Apps service. Many businesses use Gmail (either for free or on a paid plan), so it became quite noticeable when the service was out for a whopping 30 hours for some people. What made the outage so unfortunate for Google was the fact that many of the people affected were small businesses who weren’t able to access their email for over 24 hours, losing profits in the process.

6. Digg Scripts

Digg saw some of its top users (and many others) banned throughout 2008 for using scripts to automatically vote on their friend’s submissions. While many believed that it was unfair not to give these active users a second chance, others argued that a zero tolerance policy must be the standard. Regardless of which side of the fence you sit on, few disagree that Digg scripts damage the community.

Why is this a web development blunder? If you’re going to make a Digg script, you better be certain it’s not blatantly easy to detect. NETTUTS suggests the coders behind these scripts put their time to better use, for example, by coding some apps that actually help the community rather than harm it.

5. Facebook Beacon

When Facebook launched their innovative new ad platform Beacon, it received extremely harsh criticism. It turns out that Facebook users didn’t like the idea of everyone knowing their purchase history on the Web. There are numerous horror stories of people buying things on Facebook’s “Beacon partner sites”, including one guy who bought an engagement ring at Overstock.com and immediately had all the details of the purchase pushed to his Facebook profile, ruining the surprise engagement.

Facebook later added opt-in options to Beacon, but most of the damage had already been done. In the end, Facebook beacon resulted in a couple class-action lawsuits for Facebook and the partner sites associated with the service. The lawsuit claimed that Facebook silently harvested data, not only from Facebook users but also from non-Facebook users on the Beacon partner sites. Also, Facebook admitted to collecting data even after users opted-out of their purchases being shared. Ouch.

4. Cuil’s Sub-par Launch

With all the hype surrounding stealth-mode Cuil, the innovative search engine had plenty of anticipation built up around its launch. Founded by ex-Google employees, Cuil promised to be a better search engine than Google, with more relevant results based on word associations, and a better results interface. Cuil had the makings of a very impressive search engine, and (finally) a possible Google competitor.

However, the Cuil launch proved to be a huge disappointment. For starters, the site was down for hours, which is forgivable given the massive amount of traffic the site received. More importantly though, Cuil didn’t live up to its over-hyped billing. The search results weren’t even close to Google’s relevancy, and their indexing bot was crashing websites.

3. Amazon S3 Outtage

On February 15, one of Amazon’s three geographic locations went down for a few of hours, killing the Amazon S3, the cloud storage system. Many web startups like Twitter, SmugMug, and even the New York Times had error messages on the site. Many of the most popular sites on the Internet were affected by S3′s downtime.

What irked many about the downtime was the lack of communication from the development team. It took the Amazon S3 development team one hour to verify the outage, and they didn’t respond again until the problem was fixed a couple hours after that. Many S3 customers found that unacceptable, considering entire businesses were shut down during the duration of the outage.

Transparency and lots of communication are key when a development crisis happens.

2. Twitter – The Popularization of the Fail Whale

As Twitter continued to grow over the year, the popular microblogging site had more and more sightings of the infamous Fail Whale. Thanks to Twitter’s awful uptime, the error page sightings became so frequent that Twitter users started to take a shine to the sleeping whale. It wasn’t long before the Fail Whale had it’s own fan club, merchandise, and even a Twitter account.

Not many web developers can claim that their error page logo has a fan club, and not many would want to either. The error page is the last thing that any development team would want their users to see. Fortunately, Twitter’s error page is fun and catchy, so they’ve made the best of the situation. Yet I doubt the Twitter folk would want their legacy to be the popularization of a sleeping whale.

1. Site Meter Crashes For IE Users

August 1, 2008 was a day that Site Meter developers will remember for a long time. It was the day that they effectively shut down a good chunk of the Internet for Internet Explorer users. Essentially, any site that had Site Meter tracking installed caused the site to issue “Operation Aborted” messages to Internet Explorer users.

While the blame lies partially with Internet Explorer (they had a well-known bug that triggered the error), the big chunk of the blame rests with Site Meter developers, as they pushed a change to production without testing it thoroughly (AKA without testing in IE at all).

Doh! Nothing like alienating 70% of the total visitor’s to your site’s pages.

Thanks to Thomas for help with the list.

• Subscribe to the NETTUTS RSS Feed for more daily web development tutorials and articles.

Glen Stansberry is a web developer and blogger who’s struggled more times than he’d wish to admit with CSS. You can read more tips on web development at his blog Web Jackalope.

Liked this post? Vote for it on Digg below. Thanks!