3 ms·
I parsed through the Library of Congress digital records to find an out-of-copyright subset, and developed a search engine that is unique in that it is technolo
by westcort 4y ago
I parsed through the Library of Congress digital records to find an out-of-copyright subset, and developed a search engine that is unique in that it is technologically impossible for me to track searches, as the index loads in the page. It has about a million entries.
https://www.locserendipity.com/ https://www.locserendipity.com/
- westcort 4y agoAlso this Project Gutenberg randomizer: https://www.locserendipity.com/Gutenberg.html https://www.locserendipity.com/Gutenberg.html
- NelsonMinar 4y agoNeat trick! I took a look and the page is about 4MB or maybe 8MB, less than 1MB compressed. Smaller than the typical animated ad. Is there an actual index data structure? What I see on the page is just a bunch of simple <a> links that are linearly scanned. Nothing wrong with that mind you, it seems to work well.
- westcort 4y agoIt’s just simple keyword matching in the order of appearance on the list. I did something similar with the DMOZ directory: https://www.locserendipity.com/DMOZ.html https://www.locserendipity.com/DMOZ.html Also, as an experiment, I developed an auto-summarizer, scraped and summarized all of the edu domains on DMOZ, pared the non-active ones, and the ordered them by the amount of content (more content ranks higher). The result is this search engine: https://www.locserendipity.com/edu.html https://www.locserendipity.com/edu.html