OWL is a totally community-driven project unlike KnowledgeTree which used OWL as a base. In general, OWL looks much less sophisticated than its PHP peer with a somewhat rough user interface and less functionality. The user interface appears to be designed for more technical users who are comfortable with a relatively steep learning curve of becoming familiar with icons and terminology.

ARCHITECTURE

OWL runs on PHP (4.1+) and has options of MySQL, PostgreSQL, and Oracle for a database. Documents are either stored in the file system or in a MySQL database while metadata is stored in the database (except if in the case OWL-Lite which only uses the file system). There is a backup feature which ensures that a backup contains an archive of the database and the file system in a matching state.

CREATE AND STORE

When uploading documents through the web client, users have the option of entering basic meta including title, keywords, and description. There is an option to integrate virus checking for all uploaded documents. By clicking a checkbox, entered keywords are automatically added to a remembered list to be used in later document uploads. Users can also set major and minor revision numbers. After setting the initial values the revision numbers are auto-incremented according to the editor’s selection of minor or major revision. The metadata that is entered is extremely important because, as mentioned later, the search engine depends on it.

Users can upload a directory of files at a time in a Zip archive. Metadata values (including the title) will be the same for all the documents uploaded in this way. However, it is simple to go back and change these values after upload. The administrative interface has an initial load feature which allows an empty repository to be propagated with files placed on the server’s file system. Files uploaded to the OWL directory structure will be added to the database when browsing the archive.

ORGANIZING AND COLLABORATION

Documents are stored in a folder hierarchy that is easy to reorganize using the move option that is available on both folders and documents. There is basic check-in/check-out functionality which locks the file from editing by others. OWL supports a comments feature that allows users to add messages about a document. Users can also “monitor” folders and documents so that they are notified whenever there is a change. A review feature allows for documents not being published before a set of defined users have approved on the documents content.

Prior versions of documents are stored and can be accessed through the user interface. The versioning system has a simple design of renaming a prior version by adding a version suffix (before to the file type extension) to the file name. For this reason, whenever a file is over-written, it must have the same file name as the original file. Files and folders deleted by users are put into a “Trash Can” that is visible to the administrator who can delete them permanently or restore them.

There is a simple but nice persistent dashboard on the user interface that presents useful statistics including how many documents have been added recently, how many documents the current user has checked out, and how many have been updated. Drill down links display the documents that are behind these numbers.

Users can email documents individually or in bulk. There is an option to attach the file or just send a reference to the document. The attachment option is useful when emailing to individuals who do not have an account on the system.

SEARCH & ACCESS

OWL has a simplistic permission system where a document owner can select from several policy options including: "€œEveryone can read/download,"€ "€œOnly you can read/download/write," and"The selected group can read/write/download (NO DELETE)."€ Users can also associate a password with a file or folder to further limit access. The next version of OWL might allow for more granular security as it introduces Access Control Lists (ACL).

Users can download or preview individual documents and can also download a folder at a time as a zip archive. However, files should be updated individually.

OWL's search engine relies heavily on document metadata rather than full text which is only done for a select set of file types (PDF and MS Word). Search result ranking ascribes different weights to keyword matches in the title, file path and name, keywords, description and comments. There is no advanced search, but the basic search has options of whether or not to consider the full text of the document and whether to restrict the search to the current folder. There is also a "€œfind related documents"€œ feature from that finds documents with similar metadata entries to the current document.

USABILITY

Non-technical users may feel intimidated by the user interface. Most of the operations are represented by small cryptic icons that take a while to get familiar with. A quick reference card that explains what the icons mean should be kept screen-side as the user gets acclimated. The limited search could be critical limitation.

COMMUNITY AND RESOURCES

Documentation is sparse but the forum is generally active and responsive. Being a totally community driven project, development on OWL goes in fits and starts. The community is dynamic. For example, recently an external developer created an alternative user interface for OWL and made it freely available (after attempting to charge $10 per download and then being informed of OWL’s GPL licensing restrictions).