당시  At that time    
   start quiz
  config wordflash
        
credits


design and concept by me
  • databases: postgres stores the dictionaries and lists in a UTF-8 encoded database. The discussion forums, user registration and data, login functionality, the osCommerce book side work off mysql.
  • php database interfaces
    • Pear::MDB2: ezcorean software uses MDB2 to connect to the database (the miniBB and gnuboard connections are also used in those places).

    • DB::Pager: all the pages which show a list of page numbers at the bottom, for the vocabulary lists.

  • old, discarded site pieces and parts
    • drupal was once used for the reading section, but it was totally incomprehensible to work with due to functions instantiating classes and calling their functions which call other functions, and it was buggy as hell .... In the end, the reading section is just a miniBB hack.

    • hanaTheReader background아깝다 I really should include this somehow ... ^_^

    • c++ cgi library: yes. believe it or not back in 2002 I tried to make my own library and make a c++ web site. It was fun. c++, mysql, linking libraries, uploading to sourceforge on an open source project. But, progress was slow because I had to compile to test, and NOBODY is programming in c++ for the web!@

    • then i switched to php, which is like 'vb meets the web' mentality.

    • TinyPHPForum: From April 05, 2005 - sept. 2005. chosen for its simplicity and multi-lingual code base. v2.0 RC5 © ... and some style sheet tgs from them still around. However, was flat-file, beginner coding, no options. Wanted something more.

  • javascript: this site has not been shy by forcing the user to download large javascript files!

    • listCollapse.js: for all the dynamic ul/ol lists in the boxes on the left which really enable navigation

    • gnuboard forum furnished lots of the javascript code which I cavalierly reused. I made especially good use of the sideview function, to show a dropdown list of meanings from the database.

  • forum software

    • gnu board discussion forum since looks like August 21 2005 to february 2006. database-driven, many features, excellent thoroughness. my introduction to php forum korea-style. however, lacked multi-lingual capabilities.

    • miniBB: from February 15 2006. This was the complete forum for the question forum and the examples for the list at the same time, until I changed back to gnuboard for the question forums March 22 2008.

    • gnu board again for the forum from March 22 2008, because the question forum is different from the resource side of the site.

At present, the site is a home-grown mixture of an interface to the lists in the postgres database, and a hack of minibb to pull up reply posts. Unfortunately for me, because it was easier for me to set up the postgres for all UTF-8 I put the the words themselves in there , but the follow-up posts are in mysql so it was a pain. The various study lists correspond to various "switch" variables which you can see in the URL. Then, the corresponding (postgres) table is brought up by adding the switch to a generic table prefix to make a table name up. In retrospect, all I did was write forum software, the switch could have been a field in a single table (like a forum_id), and put that in mysql and everything could be done with a single join. Instead, to associate the postgres words with the follow-up posts, the post_ids are given ranges with which posts are then associated by adding or subtracting a certain amount from the wordid in the postgres table. Unfortunately, when the site started, I had no idea it would end up as it is, however.

I also then ended up with a few interfaces for the site: the minibb interface (for grammar and the follow-up posts), a homegrown paging interface (which uses the DBPager class), and a hack of the minibb to make the reading section (of the site), and the gnuboard forum for the discussion forum. To integrate the same logins to work with these 3 different forums was a lot of time-consuming hacking. But I couldn't envision the purpose of the site until well into its development.

  • open-source dictionaries used on ezcorean
    • english-english called "wordnet". A master in the field. The database has 200,000 lines but I think the word count is around 87,000. My parsing of their text file initially had some problems, but I went through in another pass. Many words not in wordnet are available from the english wiktionary project which is a project of downright amazing scope.
    • english-korean called "engdic". The files required significant cleaning up before inserting into the database. This dictionary will be replaced with its unparsed version very slowly .

    • korean-english called "my dictionary". The english-korean dictionary above was added with 6,000 post popular words, and all hanja in the dictionary were added here at ezcorean.com, as well as thousands of words. stats as of January 25, 2009 are as follows:
      • 4-english words per korean words: 1492 korean words like that
      • 3-english words per korean words: 3283 korean words like that
      • 2-english words per korean words: 1879 korean words like that
      • 19302 .... hanja and the word with no definition (these words all need definitions)
      • 165,714 .... total entries in the database , including the above to-be-worked items

      In addition, there were more than perhaps 30,000 so-called 'words' previously converted into example sentences.
    • french-english called "?". it's the small 34,700 some word one from the net. Did not require any significant amount of processing, but severly lacking in words. I can speak french so am not very interested in developing this dictionary.

    • french-french from fr.wiktionary.org ... i parse the french wiktionary with this script here to wiki.pl which I wrote myself.

    • japanese: called "edict". easy to parse, the professional.

    • german: en-de.txt. There was a logical problem with some of the formatting I couldn't get into the parser. multiple german words on left, multiple english words on right, using semicolons where | was called for. The first line of the file is an example.
      also, these line was improperly formatted because of word;NOTHING:
      durchtrennen; abtrennen; trennen; teilen | durchtrennend; abtrennend; trennend; teilend; | durchgetrennt; abgetrennt; getrennt; geteilt | trennt durch; trennt ab; trennt; teilt | trennte durch; trennte ab; trennte; teilte :: to sever | severing | severed | severs | severed

      berwältigen; | überwältigend | überwältigt | überwältigt | überwältigte :: to overcome {overcame; overcome} | ov

      I only point this out because with those exceptions, it was an excellently maintained and simple file, unlike engdic and wordnet.

  • overlib (mouseovers generate definitions)
    • hannaTheReader: php classes for parsing rss feeds. Had to be changed to parse HTML code embedded in the RSS files, no easy task in php (unlike Perl, there are no libraries for those routine tasks).

    • ezcorean html parsing functions: these allow UTF-8 html (in KOrean) to be parsed according to split_variables (which depend on the source language -- korean, japanese, german,etc. ). Then the words become tokens and are searched for in the databse to create easy-to-read texts. Yes, php was a bad choice for any multibyte programming, and it has hit a wall as far as speed and file sizes which it can parse. I'm also afraid of changing servers or php versions as they change multibyte stuff on a whim.

    • overlib.js javascript file: for the dynamic pop-up enabling non-registered surfers to change the skin language
      I don't use this as the site has been reduced to just a korean site due to my present involvment in Korean.

 


 공연히  Futile or vain or useless    
   start quiz
  config wordflash
 그런  Such a    
   start quiz
  config wordflash
 그런  Such a    
   start quiz
  config wordflash
New members in last 24 hours: blackinwhite2010 ,   triplen unique visits today: 1076
전체 unique: 1,232,331
dictionary searches today 114
page views today: 16510
webmaster contact 연락처
All the information (which is not under another copyright already) contained in these pages is Copyright (C) me 2006-2099.

This site, ezcorean.com, was developed as a completely charitable endeavor. Ezcorean.com seeks to further everybody enjoying learning Korean (I love languages) and enjoy the joy of speaking a foriegn language. If you do not quickly find what you want, please send an email to the webmaster or post in the forums and perhaps somebody will help you in your language learning journey. Peace to all!

Hosted at johncompanies.com