The search box lets you specify one or more words, to be searched for on this web site. Multiple words must be separated by whitespace.
The search is case insensitive: upper and lower case are considered equal, but see below for an exception. Singular and plural are different. Do not use any punctuation marks, hyphens, quotes etc., but split your query up into separate words as much as possible.
Don't use accented characters in words to be searched. Map them as follows: áäà→a, ćç→c, éëè→e, óöò→o, ü→u
Words containing (other) characters outside the ASCII range (e.g. accented characters not mentioned in the previous paragraph, or £, € ...) are only supported in a strictly case sensitive way. This is work in progress. In combination with the wish to operate case insensitive, proper handling of accented characters is very complicated.
Words that appear in an excessive amount of pages (over 500) of this site were eliminated from the search database. The reason for this is that such a word would produce so many hits that your answer would be diluted to the point that it would become useless. They will be ignored in your search query, and an appropriate notification will appear.
If a page has been selected for display (without the help of these words), these words will be highlighted if they happen to be present in that page.
The search database does not contain any patterns that are non-words, such as plain numbers (e.g. years and page numbers). Care has been taken to preserve car registration numbers, publication numbers, VIN-numbers, etc.
If you precede a word in your query with a hyphen (e.g. -Benz) all pages containing that word will be suppressed.
Grouping a range of words within double-quotes (") will search for pages containing these words adjacent to each other in exactly the given sequence. But this facility has a catch in performance; as we will explain below.
Hyphens inside a group are ignored. It is allowed to precede an entire group with a hyphen.
The search database contains all words from the web site, but has each word in isolation. Relationships between words, such as adjacency, can only be determined by (re-)scanning the page files themselves. Scanning a page just before it has to be displayed is no problem. But imagine how much capacity is wasted if the software first has to find out that, say, 1218 pages fit the search profile (by scanning many more), and then the reader quits after he has seen the first result list of 20. These large scale scans could require dozens of seconds of the full capacity of the server computer.
Therefore the following compromise has been taken: the "total results" count that is displayed above each result page, is calculated as if each word of your search was given in isolation. So that count will probably be too high, and the text "(... or less)" will be displayed with it. When the result list pages are being built, a closer look at the web pages will eliminate the ones that do not have the proper adjacency relationship. Then the value of the counter will shrink on the fly, and may very well shrink to zero. In summary: when your search does contain double quotes (") for word grouping, the initial "total number of pages found" count will be too high. It will shrink on the fly. Have you ever noticed that e.g. Google shows this same "shrink on the fly" effect?