YaCy 1.6 is out since yesterday. This release contains mostly enhancements, bugfixes, a special ‚greedy learning mode‘ and most important, a feature that many of you requested using anonymous messages to http://sayat.me/YaCy, a link that is shown on the goodby-screen of YaCy after a shut-down. Many people used this link to report, that they would run YaCy permanently if YaCy tries to keep it’s CPU load and IO low. And thats therefore the main feature of version 1.6: if YaCy is running in the background and used for searching, it will try to keep it’s IO and CPU load low.
Here is what we did in detail:
- We examined the IO problem and found out that Solr needs regular optimization processes. Without this, the IO is very high during DHT transmission (the peer-to-peer sharing of search indexes). With an optimized Solr, this process is done much more efficient.
- We integrated a CPU load sensor which causes that no DHT transmission is made if the CPU load is too high (affects sending and receiving of indexes).
The new ‚greedy learning‘ mode will cause that YaCy loads linked documents from the first search results, until a total of 1000 documents are in the local search index. This is mostly reached at least after the first three searches and after that, YaCy can benefit from these documents in it’s search.
We urgently need your help, please become a YaCy developer! If you have any ideas or suggestions how to enhance YaCy, the please clone our git repository, watch a video how to start developing and send us a merge request!