Chinese Yellow Pages |Classifieds | Knowledge |Tax |IMEToggle navigationwww.comrite.comWhat’s the difference between BNF, EBNF, ABNF, ASN.1 ?No ratings yet.August 25, 2016  |By Min Wang |Post commentAbstract Syntax Notation One (ASN.1) is mainly for binary protocol, it is a standard and notation that describes rules and structures for representing, encoding, transmitting, and decoding data intelecommunications and computer networking.. A good tools like asn1c can convert ASN.1 to C structure, encoder/decoder. BNF, EBNF, ABNF are for text format parsing. In terms of power, •   Read More »C++11 std::unique_ptr as semantic value in Bison C++ modeNo ratings yet.August 24, 2016  |By Min Wang |Post commentRecently I spent some time to write a DSL ( domain specific Language) lexer and parser. Without double flex, bison (C++) will be used.  Bison ( version 3.0.2/3.04) generally support C++ object as the semantics value.  That is a good thing  since we do not need  %destructor  to enable memory deallocation during error recovery; the •   Read More »fp-tree and fp-growth in plain englishNo ratings yet.August 10, 2016  |By Min Wang |Post commentThe normal use case is to discover interesting relations between variables in large databases, e.g: i(t) 1 ABDE 2 BCE 3 ABDE 4 ABCE 5 ABCDE 6 BCD the above are some transactions, let’s A, B, C … are some products people bought. we want to find which set are frequent set .e.g: BC  is •   Read More »Geocoding 40 millions US addresses totally free with PostGIS 2.1 on debian 8No ratings yet.July 28, 2016  |By Min Wang |Post commentPostgis 2.1 ( on debian 8) make geocoding US addresses (  about 40 Million ) quite easy, and it is totally free! Thanks for the PostGIS and PostGRESQL open source communities. To import all the US tiger data from ftp://ftp2.census.gov/geo/tiger/, you need at least 110 G disk partition for postgresql, I would recommend 200G to be •   Read More »K-means clustering and Bisecting K-means in Plain EnglishNo ratings yet.July 13, 2016  |By Min Wang |Post commentK-means clustering is unsupervised machine learning algorithm. Wikipedia has a great demo as below on how it works: Demonstration of the standard algorithm 1. k initial “means” (in this case k=3) are randomly generated within the data domain (shown in color). 2. k clusters are created by associating every observation with the nearest mean. The •   Read More »When/where C++ five constructors (copy/assignment/mmove) are invokedNo ratings yet.July 1, 2016  |By Min Wang |Post commentC++ default , copy, move, assignment constructors Sometimes it is confusing where/when  the default , copy, move, assignment constructors  in C++ are invoked,  here some are notes: e.g.: class A { public: A(); // default constructor A(int i);  // my constructor A&  A( const A&) ; // copy constructor A&  A( A&) ; //  another copy •   Read More »Bison/Yacc cheat