So with the advancing of HTML5, Web 2.0 and everything, more and more websites are using dynamic loading. One of the more advanced technique is dynamic loading portion by portion, that is, only load the amount of content that fits the consuming context such as depending on where you are looking at/reading and how fast you are reading, only load one screen of content. This further reduces unnecessary data transmission because if you load a whole page of content for the user, the user might find it boring halfway through it and leave and essentially you wasted half of the data that has been transferred.
However, with the dynamic loading by portion technique I find another problem: it’s quite hard to expect how much content there is left. It’s like reading a newspaper or book page by page without knowing how thick it is. Sometimes it gets frustrating if you don’t know how much more time you need to finish it, doesn’t it?
What I think would help this experience, is to load the structure of the content without the data inside. Take a long article as an example, it would be loading a meta data describing how many paragraphs there are in the article and how long each paragraph is. With this data, the whole page could be constructed without the actual content (maybe filled with spaces). And the data will be loaded in when the user actually is going to read it. Now we have the structure of the article built, the scrollbar effectively tells the user how long this article is, and can plan how to consume it accordingly.