From the same article, although not related to this subject, I found this gem:

"Here are a few examples of things you can't really do well in a web application:

1. Create a fast drawing program

2. Build a real-time spell checker with wavy red underlines

3. Warn users that they are going to lose their work if they hit the close box of the browser

4. Update a small part of the display based on a change that the user makes without a full roundtrip to the server

5. Create a fast keyboard-driven interface that doesn't require the mouse

6. Let people continue working when they are not connected to the Internet"

It makes me smile that all of these are now incredibly trivial to implement in web apps :-)

Amazing how often I get asked to create #2 tho (about once every month or two). I usually ask, "So the fact that no site not owned by Microsoft or Google does this isn't scaring you at all?". And the answer is, "Well it would be nice to have". The next step is telling them it would take me over a month to build...and everything goes down hill from there with accusations about my competency.

> 6. Let people continue working when they are not connected to the Internet" > now incredibly trivial

Really? That was not my experience with one month off the grid using a Cr-48 in 2011.

Except there is no standard implementation for any of those things, every webapp treats them differently, offers a different subset, some offering none at all.

How do you do #4 and #6?

For #4, I assume he means something like "click-to-expand" which now can be done entirely client-side in Javascript.

For #4, assuming that the assets have been preloaded, you can display the change through your scripting, while sending the update to the server.