Before we get too deep into the technical side of things, let me take a step back:
Ajax, for those who don't know, is a technology where parts of the page can be modified, based on new data from the server, without needing to refresh the whole page. As far as I know, there are two advantages to having query results use Ajax:
2) Results can be refreshed, manually or automatically, without reloading the page.
3) It eliminates the problem of MediaWiki caching of query results (the old "why isn't that new page showing in the list?" problem), because the query results aren't cached along with the rest of the page.
Until now, I think I've mostly heard #1 as a reason to use Ajax in SMW query results - though, Yury, you seem to be talking mostly about reason #2. (Correct me if I'm wrong.) That seems strange to me, because I think on the vast majority of wikis, changes don't happen nearly fast enough to require any type of real-time refreshing; and certainly not for a refresh rate measured in seconds.
Personally, I think reason #3 might be the strongest one, by the way.
In any case, I have a question: Yury - if Ajax is possible for every result format, why both with an "ajax=[true/false]" parameter at all? In other words, is there ever an advantage to *not* using Ajax, given the option?