search code

Ok, first of all you need to understand how search engines work.

There are basically two ways to accomplish what you want. One is to use a script that will spyder the pages you define right after you enter the search query. This would tpically be done with a java applet but if you want to search all of Able2Know there are several million links and it would take a long long time.

And it'd crash your browser.

So you'd need to search a database that has spidered the pages prior to the search.

For that you need to either make a spider, pay for a spider, use a free spider (usually limited to a few thousand pages), or simply use an existing search database.

The last option is easy. Just use Google's search box. Of course if the site is not indexed in Google it won't help you.



Replace Able2Know with the domain you want to search.

Here's some short code I wrote for me own site that does the same thing.

Again, replace able2know with whatever domain you wanna search.

Husker, do you mean to
a) embed a search done with google, with results shown at google,
b) create your own search engine database, with results shown by your server,
c) perform a search using google, with XML results parsed and displayed by your server,
d) or just jumping to another sites search results?
(For instance, click here to search A2K for everything Monger has said about Linux.)


Without XML, if you want your server to collect and format the search results from another website's search engine, you might use something like WinHTTP to fetch the page, then parse it into whatever data you need. A little harder than XML but not much.

PS -- Craven, does phpBB support XML search results? Don't see any embedded, but who knows.

I really don't recommend spidering other people's pages with a tool you are unfamiliar with though. To give an example, just Google's spiders themselves use over 30 gigabytes of bandwidth a month. If this wasn;t google but some Huskerbot I'd not really want that server load and bandwith.

So just be careful what you use. A well written spider takes server resources into mind and will not down a server.

The easiest solution would be to use a remotely hosted "site search" feature and it will spider a medium sized site for you.

If you pay you can spider more pages but it might be against their TOS to spider sites you don't own.

To put together a site search of a site you own you can use a script that will index your pages. Typically it will run on a cron job and index all your files.

Note that this is a solution for flat files and database data might need a custom solution to search.

Another option if you are dealing with few pages is to manually index them. It's not worth doing for more than a few pages though.

Cb,

Nah, it would need a modification to do so. Probably exists. I might look into it.

the google one seems to be working out ok

Well, if you want search that advertises hoagiesgifted with a massive banner that's the code to use.

For anyone who wants to customize google search for their own site just fill out the form here & it'll generate the code for you: Customizable Google Free Search