Merge pull request #5 from JulianLeviston/JulianLeviston-patch-1

Improve Readme

README

@@ -1,19 +1,23 @@
-TCache is a transactional cache written in Haskell, with configurable persistence. It allows conventional STM transactions for objects that syncronize with their user defined storages. Default persistence in files is provided for testing purposes State in memory and into permanent storage is transactionally coherent. 0.9.0.4 : solved a bug in the management of weak pointers that evaporated registers from the cache 0.9.0.3 : Solved a lost registers bug. 0.9.0.1 : Solves a bug when object keys generate invalid filenames. changes in defaultPersistence to further separate serialization from input-output
+TCache is a transactional cache written in Haskell with configurable persistence. It allows conventional STM transactions for objects that synchronize with a user-defined storage. State is kept synchroized between memory and permanent storage via transactions. Default persistence is via files, and is provided for testing purposes.
 
-0.9: This version now has full text search. Serialization is now trough byteStrings
+0.9.0.4 : Solves a bug in the management of weak pointers that evaporated registers from the cache
+0.9.0.3 : Solves a lost registers bug.
+0.9.0.1 : Solves a bug when object keys generate invalid filenames, and includes changes in defaultPersistence to further separate serialization from input-output.
 
-This version support the backward compatible stuff, that permits transparent retrievals of objects and transcactions between objects without directly using STM references ('with*Resource(s)' calls), Now it goes in the oposite direction by providing explicit STM persistent references (called 'DBRefś') that leverage the nice and traditional haskell reference syntax for performing database transactions.
+0.9: Introduces full text search. Also adjusts serialization so it is achieved through byteStrings.
 
-DBRefs are in essence, persistent TVars indexed in the cache, with a traditional readDBRef, writeDBRef Haskell interface in the STM monad. Additionally, because DBRefs are serializable, they can be embeded in serializable registers. Because they are references,they point to other serializable registers. This permits persistent mutable and efficient Inter-object relations.
+This version supports backward compatibility and also permits transparent retrival of and transactions between objects without directly using STM references. It now provides explicit STM persistent references ('DBRefs') that leverage clean, traditional haskell reference syntax for performing database transactions.
 
-Triggers are also included in this release. They are user defined hooks that are called back on register updates. That can be used for easing the actualization of inter-object relations and also permit more higuer level and customizable accesses. The internal indexes used for the query language uses triggers.
+DBRefs are essentially persistent TVars indexed in the cache using a traditional readDBRef/writeDBRef Haskell interface like that of the STM Monad.  Additionally, because DBRefs are serializable, they can be embeded in serializable registers. As they are references this means they point to other serializable registers. This enables persistent, mutable and efficient inter-object relations.
 
-It also implements an straighforwards non-intrusive pure-haskell type safe query language based on register field relations. This module must be imported separately. See Data.TCache.IndexQuery for further information
+Triggers are also included in this release. These are user defined hooks that are get called on register updates. They can be used for making the actualization of inter-object relations easier, and also permit higher-level customizable accesses. The query language internally uses triggers for its indexing.
 
-Now the file persistence is more reliable.The IO reads are safe inside STM transactions.
+This version also implements a straight-forward, non-intrusive, type-safe pure Haskell query language that is based on register field relations. This module can be imported separately. See Data.TCache.IndexQuery for further information.
 
-To ease the implementation of other user-defined persistence, Data.TCache.DefaultPersistence needed to be imported explcitly for deriving file persistence instances.
+The file persistence implementation is more reliable as IO reads are now in STM transactions.
 
-The 0.9 version adds full text indexation and search, incorporated in the experimental query language
+To ease the implementation of other user-defined persistence, Data.TCache.DefaultPersistence needs to be imported explcitly for deriving file persistence instances.
 
-It also changes the default Persistence mechanism. Now ByteStrings are used for serialization and deserialization . a Serializable class and a Persist structure decouple serialization to/from bytestring and write-read to file. Both can be redefined separately, so default persistence can be changed with setPersist to write to blobs in databases, for example. Default persistence has'nt to be in files.
+The 0.9 version adds full text indexing and search, which is incorporated into the experimental query language.
+
+It also changes the default Persistence mechanism. Now ByteStrings are used for serialization and deserialization. a Serializable class and a Persist structure decouples serialization from bytestring and read/write to files. Both can be redefined separately, so the default persistence could be changed with setPersist to write to blobs in a databases, for example. Default persistence now no longer has to be in files.