GraphicBrain.com, the specialized search engine and personal assistant that the Flemish Innovation Centre for Graphic Communication (VIGC) launched a few months ago; keeps working on the improvement of its services. Now they’ve added a feature to retrieve the text of lost web pages from its cached memory. By doing this, GraphicBrain.com makes lost information available again. "We’re certainly not the first to do this, but it is really a major improvement of our services. This will certainly strengthen our position as the point of reference for people in the graphic arts and related industries who are looking for information." says Eddy Hagen, manager GraphicBrain.com. Google was the first search engine to offer this feature. "That’s the reason why so many people like Google: you’re sure that you can read the documents that were retrieved. Which isn’t always the case with websites." adds Hagen. The Web is very dynamic. Although GraphicBrain.com started indexing websites only a few months ago, several sites have already changed their whole site, with a lot of dead links as a result. "And then you have to make a choice: either you try to delete all the dead links from the result lists, or you start looking for another solution. We’ve chosen the second option, without hesitation. From our daily experience within VIGC - we’re using GraphicBrain.com ourselves on a daily basis - we already noticed that some really interesting information wasn’t available anymore on the new websites. Getting the information out of our engine was a logical solution." reports Eddy Hagen. GraphicBrain.com stores the whole text of every document. The main reason for this is the fact that this text is used as the basis for the summary a user sees in the result list. This summary is generated on the fly for each search and is linked to the query. A maximum of 10 sentences - with the highest relevance to the search - are taken from the text to form quite a good summary of the document. Now GraphicBrain.com is also using this ‘cache’ to show the text of lost documents. Another nice way to use the cache is to read the text of pdf documents. Sometimes it can take a while to load a pdf. Now a user can first look at the text - which will load very quickly - and decide whether he wants to see the whole document, which might be several Mb’s. On a daily basis GraphicBrain.com has an average of about 150 visitors.15% to 20% of them return within the same month. Monthly the service attracts nearly 5 000 visitors, from all over the world. Northern America (54.7%) and Western Europe (32.4%) account for the largest number of visitors.