The popular wisdom states that to forget the past is to be condemned to repeat it. In software development, it seems we never forget the past; therefore, we continually try to support it. Too much time and too much effort are expended making static, ugly Web sites that can be viewed by 2.0 and 3.0 browser versions. Im going to make the argument that we abandon support for older browsers.
Of course, it helps to define the terms. In Internet time, an old browser is one that has been in general release for over a year. There was a time in software development cycles when we would maintain backward compatibility with software that was less than three years old. But like the Titanic, that ship has sailed. The Internets gravitational vortex has caused a rift in the software development continuum. We must accelerate our adoption of new technologies lest we be sucked into a black hole of mediocrity.
Even if youre not trying to sell a product, active content should rule. Think how much easier it would be to explain how the company 401(k) plan works if you could use animation, sound, voice-overs, and other interactive candy. Active content is the key to traffic, and traffic is the gauge we use to measure the success of a site.
A Little Background
Let me explain how this article came about. My company decided to create a Web site so that people could purchase and navigate the data we create directly over the Internet. My company tracks who advertises in newspapers, and we have a lot of data on our AS/400 that people would like to be able to browse. We got together with marketing and designed a plan of attack so that we could let more people know about what we do, give away information as a loss leader in order to generate traffic, and, ultimately, expand our customer base.
During early design, we were concerned about browser compatibility. We wanted a site that could publish our statistics, but it had to be intuitive and easy to use; Internet users are not going to call for support. If they dont know how to get your product, theyll just leave your siteand youll lose a sale.
Marketing wanted to make sure that every possible client could purchase and view our data, so they put up a persuasive argument that we needed to support all Web browsers to reach the maximum potential customer base. With that in mind, we decided to create a site that would be viewable no matter which level of browser the client was using. This was the beginning of a torturous period of my life.
Why did we start with grand ambitions and end on the HTML garbage heap? We sacrificed to the god of compatibility. Just how many people are using older browsers anyway?
Now, We Surf for the Truth...
I did some digging on the Web and found a wealth of data from the University of Illinois, Urbana-Champaign (www.cen.uiuc.edu/bstats/ latest.html), which has been recording which hosts have hit its servers since March 1996. Because they are counting hosts, not hits, the numbers represented are browsers, not people or hits. There is some chance of underreporting of browsers in terminal server or large UNIX environments, but its good enough to get a general feel for browser utilization. The numbers that I quote in the following paragraphs (which are represented graphically in Figures 1 through 3) compare the month of May for years 1996 through 1999.
In May 1996, 83 percent of the browsers that hit the servers were Netscape Navigator, but fledgling browser Microsoft Internet Explorer (IE) garnered 9.5 percent. The rest were NCSA Mosaic, Spyglass, and a host of others as the browser market was extremely fragmented. It is interesting to note that Navigator Version 2 was released in March 1996 and that, by May, 59.9 percent of Navigator users had upgraded to Version 2 and 8.3 percent of Navigator users were on the 3.0 beta.
By May 1997, a new pattern began to emerge. The browser field at this point was beginning to solidify, but Navigator still led, accounting for 69.8 percent of all hosts, and
71.2 percent of the people using Navigator had upgraded to the Version 3 beta, and 4.3 percent were using the Version 4 beta. IE accounted for only 27.6 percent of the hosts, but almost 85 percent of IE users had upgraded to the Version 3 beta. Both companies waited to release their final Version 3 browsers until August 1997.
Fast-forward to May 1999. IE took the lead with 57 percent versus Navigators
40.3 percent. In the IE camp, 64.3 percent were still on Version 4, but 23.9 percent had upgraded to Version 5, and only 11.3 percent were left with Version 3. On the other side of the front, 36.5 percent of Navigator users had moved to Version 4.5, 44.7 percent were using Version 4, and only 17.1 percent were still on Version 3.
Analyzing these numbers uncovers the following tidbits:
As of just a few months ago, 2.7 percent of all browser users were using a browser other than Navigator or IE.
IE has a larger share of users than Navigator.
The share of users using Version 3 browsers is rapidly declining.
Most browser users upgrade to new versions quickly.
To Boldly Go Where No Site Has Gone Before
Where do all of these numbers lead? With the facts I gave in hand, the following questions are easier to answer:
Q: Should we take advantage of DHTML in our Web design? A: Yes. Taking into account the version upgrade rate that is evident, it is reasonable to assume that, by the end of 1999, most serious Web users will be using Version 4 and above browsers. Although Navigator does not completely implement the DHTML standard, you can use the cross-platform DHTML features to create exciting, dynamic content.
Q: What about using Macromedia Flash to include animation, interactive menus, and such on our Web site? A: Most definitely! Netscape now ships the Shockwave plug-in with its browser. Microsofts support for ActiveX component downloading and automatic installation take the headache out of third party plug-ins. Shockwave is now the standard for streaming animation on the Web.
Q: If we use the latest technologies, wont we alienate customers who have not upgraded to a newer browser? A: The trend is that Web users quickly update their browsers to the latest version. What probably drives this trend is that users go to sites and discover that they cannot experience the content because of browser limitations. By developing new content using the latest browser features, you will be driving the upgrade cycle. Wouldnt you rather be driving the upgrade cycle than riding it?
Final Frontier Thoughts
DHTML is still not fully supported on Navigator, but it probably will be in early 2000, so why not jump ahead of the competition? As more sites take advantage of DHTML, either Netscape will be forced to support it, or users will continue to migrate to IE in order to experience the content.
Shockwave is another way to generate interactive content. With 147 million browsers capable of showing Flash and Shockwave content, Id say this one is a safe bet
also. You can generate some fantastic stuff, and I urge you to surf some of the incredible Shocked sites at www.macromedia.com.
Figure 1: Between 1996 and 1999, browser use shifted from Netscape Navigator to Microsoft Internet Explorer.
Internet Explorer Other
Internet Explorer Versions
Figure 2: Internet Explorer users quickly upgrade to new versions.
1996 1998 1999
1996 1998 1999