How much is involved in DB publishing?

Abstract

XML has been intensive investigated lately, with the sentence, that "XML is (has been) the standard form for data publishing", especially in data base area.

That is, there are assumptions, that the newly published data take mostly the form of XML documents, particularly when databases are involved. This presumption seems to be the reason of the heavy investment applied for researching the topics of handling, querying and comprising XML documents.

We check these assumptions by investigating the documents accessible on the Internet, possible going under the surface, into the "deep Web". The investigation involves analyzing large scientific databases, but the commercial data stored in the "deep Web" will be handled also.

We used the technique of randomly generated IP addresses for investigating the "deep Web", i.e. the part of the Internet not indexed by the search engines. For the part of the Web that is accessed (indexed) by the large search engines we used the random walk technique to collect uniformly distributed samplings. We found, that XML has not(yet) been the standard of Web publishing, but it is strongly represented on the Web. We add a simple new evaluation method to the known uniformly sampling processes.

These investigations can be repeated in the future in order to get a dynamic picture of the growing rate of the number of the XML documents present on the Web.