The Google Search Box, I found, did not play well with iWeb's code. Google is used to searching flat – i.e. single – directories, with all searchable content in a single top-level directory – unlike iWeb's subdirectory system mentioned above. The result was a Google search box that could only find the name of a file based on search parameters – but when you clicked on the search result, you got a “page not found” error. This was because Google was for some reason searching only the top level directory, even though the iWeb generated Web page from where the search results came list the correct subdirectory address. Working on this problem alone took more time than designing the whole site, but I was determined to have a Google search. In the end, I just copied the HTML files to the top level directory, for the sole purpose of allowing Google Search to open them. I'm still trying to figure out why no amount of tampering with the Google code can make the search engine look in subdirectories (I suspect that it has to do with Google's attempts to get businesses to buy search engines for their site, and not just use the free one they offer).
And then there was the mouseover story. On my old Newzgeek site, I had a feature where one could roll over an article title with a mouse and get an auto popup description of the story without having to click on the link. This, of course, is way beyond what iWeb can do, or what it was designed for. I was prepared to live without this feature (using the Acronym tag), until a reader wrote in to tell me he missed it – so add it I did, once again using manual markup techniques.
Ds@newzgeek.com