|Data[how to use]
The grab started with one tracker item per user (plus another for themes). Because the long tail made these look implausible to save in time, and because single transient errors (which the site was full of) would lose all the work of one of these, we switched to item-per-URL. Very late in the grab, we discovered that what looked like a capacity issue of the targeted was actually poor international peering on the part of its ISP - a South Korean warrior ran much faster than the rest. This came too late, though, and we saved (pure guessing) ~30-65% of the sites. Media and RSS feeds stayed up, so we got those afterwards, where we knew about them.
Users can 2023-12-18.• • • up until
- • • •
- • • •