Given the recent news on LinkedIn moving away from HTML5, I thought this post would be worthwhile to write, to counter those who are once again knocking HTML5 for serious mobile development. First off I want to reiterate the fact that both LinkedIn and Facebook still think HTML5 in general, and HTML5 on the mobile web both have a bright future, and neither company has completely written it off.

As a student as the University of Texas, I am very fortunate to be able to work closely with Bob Metcalfe, co-inventor of Ethernet, and founder of 3Com. Ethernet celebrates its 40th anniversary this year with a history of obviously great success. Last week Metcalfe was speaking on why Ethernet 'won', and he attributes that to the fact that it was an open standard. As he was speaking, I started realizing there are a lot of similarities between the history of both Ethernet and HTML5.

First, a bit of a background on Ethernet.

Ethernet vs. Token Ring

The main competitor to Ethernet in 80's was IBM's Token ring. Like Ethernet, Token Ring was a local area network technology.

Standards

IBM's Token Ring was also a standard, but a bad one because IBM was not committed to interoperability (where many different systems can work together). The fact that Token Ring was developed by a PC manufacturer, subsequent manufacturers were less likely to choose a standard controlled by a competitor who wasn't focused on making it work in as many environments as possible.

Similar to IBM and Token Ring, Apple controls Objective-C thus Android and others will clearly never support it, and each operating system has their fairly controlling, closed-off App Store. These factors do not bode well for 'native' apps. The only current ubiquity can be found in HTML5.

Interoperability led to Ethernet's overall success, and it makes sense that HTML5's open standard equivalent will do the same.

Performance Improvements

The Ethernet-HTML5 analogy, though Metcalfe is probably a lot less likely to acknowledge, holds true on the performance end as well - at least at first. Without question native code performs much better than Javascript, just as Token Ring was technically more ideal than Ethernet in many areas. Ethernet however improved more rapidly by offering more and more bandwidth, while keeping it's price much lower than Token Ring. In the mid-80s Ethernet adopted twisted pair ether wiring (like Token Ring) which made Ethernet cheaper, faster, and available in more places.

Javascript engines, especially on mobile, are improving at an incredible rate to the point that they will bring near-native performance. In fact, with LinkedIn's move away from HTML5, speed wasn't the reason, "It’s not performance issues, like speed or rendering", but rather memory issues (Source).

With the performance aspect no longer as much of an issue, the other benefits (lower cost of not having to fully port apps) will contribute to HTML5's success.

History Repeating Itself

Ethernet and HTML5 have a surprising amount of similarities, and if history repeats itself as it often does, HTML5 will succeed over native apps.

The fact of the matter is Microsoft, Apple, Google, Mozilla, Intel, and just about any major tech company strongly supports the idea of HTML5, and that idea is rapidly becoming reality in today's native-app driven mobile world.