(While I think that your question is usually not regarded as a good question for this site, I think that it should be preserved because almost everyone new to the field is similarly confused at first and it's hard to find straight, sober, and balanced information about it.)
Is the Semantic Web dead? Are ontologies dead?
Many people moved away from the Semantic Web. On the other hand, there are still many using it. There's always been a lot of confusion about where its value lies. There are use cases that genuinely benefit from the semantic web technologies but they are quite few and far between.
Linked Open Data
That's partially why the subfield of Linked Open Data was created. You can see Linked Data as a more pragmatic stripped down version of the Semantic Web (or as a necessary enabler of the grander semantic web vision).
Linked Data, while more pragmatic, still uses ontologies. It is just not so uptight about using OWL and designing your ontologies first and in a very formal way.
A knowledge base such as Freebase (now being replaced by Wikidata) doesn't even build on Semantic Web technologies (it's always been available also in the form of RDF dumps though). On the other hand, even Freebase builds on concepts similar to those in the Semantic Web and it does have a form of ontologies.
Software to generate ontologies
There is no software that would create high-quality ontologies automatically for you. At least not in the sense of OWL/first-order logic ontologies. On the other hand, many kinds of software, including Cratilo mentioned by you, can help you to build a lower step in the semantic spectrum such as a glossary or maybe even a folksonomy which can later be transformed into more of a taxonomy or ontology by other algorithms. There might be even software to create fullblown OWL ontologies but I think that's still rather an active area of research.
Ontologies and methodologies
As always, it depends on what exactly your needs and goals are. In fields such as biomedicine and life sciences in general, they create complex formal ontologies because they have data with highly varied structure and ontologies help them keep things organized and usable. In contrast, an eshop might be better off sticking with normal relational database modelling and adding only global identifiers (URIs) in the spirit of Linked Data if they want to build a proper knowledge graph later.
Even if you need ontologies you might want to skip the methodologies first. Especially if you are really new to the field and have little idea what your needs are.
Jena, Sesame, ...
Jena, Sesame, Virtuoso, etc. are triplestores - they are used to store and query RDF. Most ontologies can be represented in RDF. Even those written in OWL. RDF has the semantic part, RDF/S, which allows you to formulate some basic ontologies. That might well be enough for start. Even RDF/S can get confusing and convoluted when you start to think about blank nodes, named graphs (are they fixed? are they dynamic?), etc.
While I think that your question is usually not regarded as a good question for this site, I think that it should be preserved because almost everyone new to the field is similarly confused at first and it's hard to find straight, sober, and balanced information about it.
