Where is the analysis processed?
Analyses of your documents are processed on Compilatio's servers, not on your device. When you request an analysis by clicking the [Analyse] button, your request is submitted. It is placed in a queue while we wait for the necessary Compilatio resources to become available to process the analysis. This queuing step can take just a few seconds or up to several hours from the moment you request your analysis to run on our platform.
Once the request has been submitted, you can sign off from your Compilatio account and log in later to view the analysis report.
The analysis process uses the text extracted from the document. This is the "copyable" text contained in the document. For the moment, images contained in documents are not processed by our software; they are ignored.
Occasionally, an alert in the form of an exclamation mark in a yellow triangle () will appear, asking you to verify the text extracted from the document. This happens when the automatic detection of language in the document has failed to recognise a significant portion of text. The text may therefore not be representative of the visible text in the original document. The cause may be intentional, where someone has attempted to trick the similarity-detection software, or it may be due to an incompatibility between how the text is displayed on the screen and the way it is encoded in the document. For example, some pdf generators use fonts that don't have standard encoding. As a result, the text displayed on the screen is legible, but the extracted text, based on character codes, is corrupted.
Searching for sources
Compilatio analysis begins with a search for source documents that have similarities with the text extracted from the document under analysis. This search for source documents is carried out among:
- hundreds of billions of web pages,
- +300 million content items, including: private scientific publications, open access scientific publications, content from internationally-renowned professional publishers, university and research center repositories, etc.
- all documents added and managed autonomously in the reference library, by users or managers of your institution, as well as those of other Compilatio client institutions.
Our software allows us to find not only the most significant sources of similarities, but also documents containing small similarities with as few as ten words, in addition to documents with similar text that has been slightly rephrased.
Find out more about the content of our databases on our dedicated page.
The next step is where these source documents are compared with the analysed document. This phase provides details about the similarities, which allows them to be quantified, and provides a rate of similarities per source.
The analysis report shows the list of source documents and ranks them based on their similarity rate, in the 'Sources' tab of the analysis report. A gauge corresponding to the rate of similarities found with respect to the analysed document appears beside each source.
Magister and Copyright software also allows you to see details for the detected similarities, in the side-by-side view and in the 'Summary' tab of the analysis report.