http://ayende.comAyende @ RahienCopyright (C) Ayende Rahien 2004 - 2021 (c) 202660@Harry McIntyre if you can't get rid of all of your usages of locks for that, and instead replace them all with interlocked calls instead. I would run for the hills.http://ayende.com/159233/multi-threaded-design-guidelines-for-libraries-part-ii#comment4http://ayende.com/159233/multi-threaded-design-guidelines-for-libraries-part-ii#comment4Mon, 22 Oct 2012 16:15:23 GMTInstead of "Concurrent Programming on Windows" I would first encourage to read the chapters on "Concurreny" in "Clean Code". Do not be afraid to follow the usual MS way: Static methods are thread-safe, instance methods are not thread-safe.http://ayende.com/159233/multi-threaded-design-guidelines-for-libraries-part-ii#comment2http://ayende.com/159233/multi-threaded-design-guidelines-for-libraries-part-ii#comment2Wed, 17 Oct 2012 12:41:17 GMTI'm working on a system based around a shared mutable in-memory datastore (persistence achieved via an eventstore which logs all method calls), as opposed to a DB. Do you have any thoughts on the best way to achieve thread safety on a system like this? I am using a lazy ReaderWriterLockSlim on begin/end request... it makes me feel slightly icky. http://ayende.com/159233/multi-threaded-design-guidelines-for-libraries-part-ii#comment1http://ayende.com/159233/multi-threaded-design-guidelines-for-libraries-part-ii#comment1Wed, 17 Oct 2012 11:14:15 GMT