Site-Wide Search

Searches can only 'see' published content. (They also cannot see into secure pages, or pages that are completely hidden from navigation.)

For a site that is already live, once you have a search results page and that page is activated, our search engine will begin to index your site. For most sites, this process can take up to 48 hours (up to a day to enter the next indexing queue, and then up to a day for indexing and return of search results).

For a brand new site, to ensure that your search results are not empty when your site first launches (before it is indexed by our search engine), consider launching your site without activating the search results page. Without the search results page, the search box in the header will also not show. Then a day or two later, activate the search results page. Your search will then function normally.

Changes to existing pages and new pages may also take time to appear in search results, but you should expect to see them by the next day. This process will occur faster if your site is known to our search engine to change frequently (i.e. the more often you change your site content, the sooner changes will be reflected in search results).

Deleted pages will similarly continue to display in search results for some time after the changes are made, although visitors who attempt to click through to the deleted page will land on a 404 error ('no such page'). Again you should expect these changes by the next day, and this process will occur faster if your site is known to our search engine to change frequently.

UB's search engine will not be able to find your site if it is not linked from other UB pages it indexes. This situation is very unusual, as most sites are linked to other UBCMS sites even if the path to them is quite lengthy, but if your site is not being found it may need to be added by special request to the list of start pages for our search crawler. For assistance, Contact the DCT Help Team.

Just like the commercial Google search engine, our local search appliance creates an index of our UBCMS system by periodically crawling through all live UBCMS sites. In essence, it crawls through your site by following the links on your pages, including links embedded in Flash content, and even in associated files like PDF and Word documents. But it cannot log into secure areas or see completely hidden content.

For a page to be indexed it must be accessible by a direct path of links from your home page. The search appliance will not follow links embedded in Javascript code, it cannot submit HTML forms (so for example, your 'thanks' page may be hidden), nor can it log into secure areas. The page should also be linked from a site within its own sub-domain -- our search engine only indexes pages within the same sub-domain (e.g. nursing.buffalo.edu or www.buffalo.edu).

The crawler will find content in external embeds as long as the external embed is just displaying HTML, but if the external embed uses AJAX or JavaScript to pull in that content, it will not be indexed. (Your site must also be linked from other previously published UBCMS sites, otherwise the search engine will not find it to begin with, but only rarely is this not the case.)

Beware if you use any of the following custom features:

• Secure content and even links to secure pages are normally hidden from being indexed. If you want them visible in a search you will need to take extra steps (e.g. see below). Similarly, you cannot set up a local search from within a secure area -- instead, consider using the Dynamic List search.
  
• Page Properties 'hide in navigation' or 'hide in lists' -- Be careful with these. Depending how your site is constructed, they may hide it from being indexed.
  
• Shared Content may be missed if it is only accessible through a carousel, slide deck, or list that only reveals some of the content (a random display of several slides from a large collection, for example).
• Content that is only accessible through special scripting may also be missed (e.g. external embeds, Javascript, form submissions).
• Page Properties 'No Index' -- This setting intentionally blocks a page from being indexed (see Advanced settings under Site Visibility).

If your site utilizes some of these custom features, you may benefit from setting up a master archive page that reveals all of your site's content. You will need to make this archive page clickable off another page on your site, perhaps in the footer. For example, the News Center provides a "browsable archive" of old news releases in its Media Kit. However, using this as a workaround for secure content would thus make that content public, including to public search engines like Google.

If you intentionally wish to block a page from being indexed by our search engine or Google/Yahoo/Bing, etc., adjust your page's Properties' Advanced tab under Site Visibility. ('No index' should be 'site will not be indexed by search engines'.)

When someone tries to load a page that does not exist on your site, perhaps after clicking an old bookmark, a 404 page appears in its place. (404 is the name of a particular server error, in this case, indicating that the page was 'not found').  

We are able to install a custom 404 page (or other error pages) for an entire virtualhost (e.g. all of mgt.buffalo.edu) by special request. CIT will then place your version of the 404 page where it will be generated automatically whenever one of your visitors tries to load a non-existing page.  

You’ll need to provide us with the complete HTML code for your new page, such as by making a page in the UBCMS and then copying and adjusting the HTML source until you are satisfied with what people will see. (These error pages are served by the apache proxy server, rather than the actual UBCMS servers and there is not an automatic connection for the error pages like there is for normal pages.)

In addition to your main search in the header, a second search can be implemented for your forms or policies in a customer service area. To add an additional search, use the Search component.