First of all, I started my research with examining the different types of publication file formats. which were created for the same goal. The oldest and most common file format is CSV which does not provide publication specific attributes but it is easy to use it. The EndNote, Refman and RIS files operate with better semantics and these file formats are more similar to each other in point of format, because it is possible to keep key-value pairs in them and we can describe everything with them from articles and conference notes to books. My chosen file format was BibTex based on LaTex which is supported by many publication sites and it uses more readable attributes instead of short keys so it is managable manuallz in an easy way.
Secondly the same can be stated about publication websites as file formats even though their purposes are the same, they are quite different. Google Scholar has the biggest database which contains a lot of entries from every field of science but sometimes its data is not relevant enough. Other popular sites are Scopus and Web of Science which are not that open sites such as Google Scholar is, and they only contain publications on engineering. Their data is precise and they use many important journals’ references so there is about a hundred years’ data available on them. One of the newest sites is Research Gate which is different in a way of looking at the publications from the scholar’s point of view. So it not only provides the possibility for maintenance, looking for a job, and scientific arguments. In the end I used the BME PA database for my application because it has a very good export possiblity feature and relevant content.