I am referring to the title that is generated and displayed on HN for the link (I don't know if it's done through scraping). Twitter has proper title tags on status pages by the way (e.g. https://twitter.com/wikileaks/status/235818483588407296)
That title is supplied by the submitting user according to some informal recommendations. Every so often, there will be an argument over whether the submitted title is appropriate for the content. HN doesn't scrape anything for it.
