contains 42 publications, 9 e-mails, 19 Word documents, and 7 text files. These resources have been indexed using the Beagle++ system obtaining a total of 11914 RDF triples which have been stored in the RDF repository 3.2 Real dat For evaluating the retrieval effectiveness of a personal information retrieval sys- tem, a test collection that accurately represents the desktop characteristics is needed. However, given highly personal data that users usually have on their desktops curre ntly there are no desktop data collections publicly available Therefore, we created for experimental purposes our internal desktop data col- lection. More detail can be found in 11] The collection that we created- and which are currently using for evaluation experiments-is composed of data gathered from the PCs of 14 different users. The participant pool consists of PhD students, Post Docs and Professors in our research group. The data has been collected from the desktop contents present on the users Privacy Preservation In order to face the privacy issues related to providing our personal data to other people, a written agreement has been signed by each of the 14 providers of data, metadata and activities. The document is written with implication that every data contributor is also a possible experimenter. The text is reported in the following L3S Desktop Data Collection I will not redistribute the data you provided me to people outside L3S. Anybody from L3S whom I give access to the data will quired to sign this privacy statement The data you provided me will be automatically processed. I will not ok at it manually(e.g. reading the e-mails from a specific person During the experiment, if i want to look at one specific data item or a group of files/ data items, I will ask permission to the owner of he data to look at it. In this context, if i discover possibly sensitive data items. I will remove them from the collection Permissions of all files and directories will be set such that only the 13s-erperiments-group and the super-user has access to these files d that all those will be required to sign this privacy statement. Currently Available Data The desktop items that we gathered from our 14 colleagues, include e-mails(sent and received), publications(saved from e-mai attachments, saved from the Web, authored/co-authored), address books and calendar appointments. A distribution of the desktop items collected from each Iser can be seen in table 1contains 42 publications, 9 e-mails, 19 Word documents, and 7 text files. These resources have been indexed using the Beagle++ system obtaining a total of 11914 RDF triples which have been stored in the RDF repository. 3.2 Real Data For evaluating the retrieval effectiveness of a personal information retrieval sys￾tem, a test collection that accurately represents the desktop characteristics is needed. However, given highly personal data that users usually have on their desktops, currently there are no desktop data collections publicly available. Therefore, we created for experimental purposes our internal desktop data col￾lection. More detail can be found in [11]. The collection that we created - and which are currently using for evaluation experiments - is composed of data gathered from the PCs of 14 different users. The participant pool consists of PhD students, PostDocs and Professors in our research group. The data has been collected from the desktop contents present on the users’ PCs in November 2006. Privacy Preservation In order to face the privacy issues related to providing our personal data to other people, a written agreement has been signed by each of the 14 providers of data, metadata and activities. The document is written with implication that every data contributor is also a possible experimenter. The text is reported in the following: L3S Desktop Data Collection Privacy Guarantees – I will not redistribute the data you provided me to people outside L3S. Anybody from L3S whom I give access to the data will be required to sign this privacy statement. – The data you provided me will be automatically processed. I will not look at it manually (e.g. reading the e-mails from a specific person). During the experiment, if I want to look at one specific data item or a group of files/data items, I will ask permission to the owner of the data to look at it. In this context, if I discover possibly sensitive data items, I will remove them from the collection. – Permissions of all files and directories will be set such that only the l3s-experiments-group and the super-user has access to these files, and that all those will be required to sign this privacy statement. Currently Available Data The desktop items that we gathered from our 14 colleagues, include e-mails (sent and received), publications (saved from e-mail attachments, saved from the Web, authored / co-authored), address books and calendar appointments. A distribution of the desktop items collected from each user can be seen in Table 1: