mailxplorer

January the 17th. Vaguely relevant though, even now.

TBL is worth over $50 million.

From what I can gather, I downloaded one tenth of Usenet - 11,000 groups. This means if I'd done all 110,000 groups it would have taken me about a year to download them, and an 8TB drive to store them (in 2013!). That wasn't really feasible...

Clucene was way too slow for body text, more than 1GB. I had my own header parser in C++ (though you can do that in Python easily).

I'm trying again on that 800GB with KISS DB (append-only hashtable), and Elasticsearch. Doesn't matter if GPL because it's a website.

It's my fault (not the IA), I thought I'd got all newsgroups but must have missed some. I just checked the main newsgroup list and it's incomplete, for some reason.

My plan though, is to dust off the old code, get a complete list of groups, get them, and then make it searchable.

Sorry about that, I didn't check. This was all done in 2013. Basically, I wanted to build a search engine but indexing the newsgroup posts (for header and text body search) would take too long. I abandoned it, then in 2016 I sent it to the IA.

So, I only just found out it's lacking a bunch of groups, 5 years on...