For years one of the biggest knocks on mobile web apps was performance -- they just never felt as fast or as fluid as native apps. Yet proponents of web apps have begged to differ, and sought to show that web apps could, if not today then one day, prove good enough for general purpose use. Tired of the subjective arguments, Drew Crawford has tried to present an objective analysis on Sealed Abstract:
We spent some time debating this topic on Talk Mobile as well. As someone who distinctly remembers it taking nearly half a minute for the webOS calendar to open on my Palm Pre Plus, and the early generation of sweet iPhone apps it's almost inconceivable to me that anyone would argue web apps, cloud-based or local, could come anywhere near native performance. As to whether they could eventually become as fast, one day, that seems like an impossible argument for anyone to make primarily because native apps show no sign of slowing down any time soon.
What can be argued, I think, is that in some cases web apps make sense despite their performance. Speed and smoothness are important elements of an app, but they're only two elements, and in some cases other ones might take precedence. The idea of an informed trade-off makes a lot more sense to me than simply an HTML5 dream and wishing it so. An app you have to access from any device, any where, for example, makes availability more important than speed, and that's something web apps excel at.
Anyway, at nearly 10K words, you might want to block out some time from Crawford's piece, or InstaPockAbility it for when you do have time. If the subject of web apps vs. native code interests you, it's well worth it.