So normally people who design for print adapt the size of content to fit the page, not the contrary.
I think there must be some misunderstanding about terminology here...
If you submit a scientific paper to a peer-reviewed, international journal, then typically you submit your text in one file (typically Word), and all figures as separate files, one figure one file, vectorized or in as high resultion as possible, and cropped as closely as possible.
A layout person at the journal then generates the final layout for paper and online publication.
If you don't follow the requirements exactly, the journal will typically incur additional charges, or ask you to re-submit.
Sorry but that's a fact. I don't want to argue about whether PDF is the final format or how PDF was "meant" to be used; fact is, it's as I described above for this type of scientific journals. So cropping makes absolutely sense.
As for defining the page boundaries, this is of course possible but tedious and constraining. There is a reason why many other programs offer an automatic crop option.
There are also very useful programs like pdfcrop on linux. However, pdfcrop does not work in all situations, it chokes on some PDFs and this is the situation I am in now. So I wanted to crop right in Scribus.