Working internal tools; continued reliability work.
A polished research summary is hard to trust when the underlying page is missing. I built super-web-scrape and related research tools to keep the source material attached to the work, including an honest record of what could not be obtained.
The difficult cases are ordinary ones: a page that will not load, an incomplete transcript, or a collection interrupted halfway through. A useful workflow needs to resume without pretending the missing material was read. Otherwise the final summary can sound much more complete than the evidence warrants.
The printmaking research made that concrete. Source collection, attribution checks and comparison with books exposed claims that needed correction. More collected text was not the same as better understanding. I separated acquiring material from accepting it as knowledge.
These are tools I use for my own research and learning. Their value is in making an answer inspectable and unfinished work recoverable. The public story is what that changes for the researcher; the collection machinery stays behind the scenes. Read the printmaking story.