[b-hebrew] Karl's lexicon

JAMES CHRISTIAN READ JCR128 at student.anglia.ac.uk
Sat Aug 25 15:13:45 EDT 2007


KWR: Oy veh!! While that is doable, I guesstimate that the resultant file
would be from five to ten times as large. That would require multiple
entries for every root, noun, adjective and so forth. Further it would
be a bear to maintain, as any update or correction would require
tracking down every occurrence of a term and updating it. That's
already hard enough when dealing with synonyms. That's why I am
looking for a more elegant resolution through more intelligent
software.

JCR: Fair enough, I can see you're not buying the 
definition for each word thing. I've already explained
the data driven method of achieving the behaviour you 
require. I guess you aren't buying that because of the 
time it would take to link each word to its root.

I am guessing that by 'intelligent software' you mean 
something like a program which takes a word and works 
out which roots it could come from by using 
morphological clues. For words which follows regular 
'rules' (rules are programmable) this is possible. This 
is even possible for words which follow irregular 
'rules'. As long as their is a rule which can be 
programmed it is doable. The only cases which would 
provide real troubles in terms of logic are the cases 
which follow no rules other than the universal rule of 
'exception;. But even these could be hard coded into 
the software by the use of many 'if' statements.

If you are serious about this then the way to do it is 
to first develop the pseudocode (this requires no 
programming skills whatsoever). It basically entails 
you sitting down and writing down in logical order and 
fashion how you, the human, would go through this 
decision making process. If you can give me an 
unambiguous series of steps to follow I can throw the 
code together for you.

*However*, I would not advise using such code in any 
kind of application. The reason being that each time 
a user queries the meaning of a word the system would 
have to go through this decision making process to 
determine the word's root. This means consuming 
processor resources and time each and every occasion 
that a word is queried. I would rather use such a tool 
to automate the process of generating the data for a 
data driven approach. The philosophy of this approach 
is 'do the hard work once and reap the benefits for 
ever'.

If you can provide me with a suitably '|' separated 
plain text version of your dictionary I can give you 
a script which will populate a MySQL database with 
your data. I can also give you some sample PHP code 
which will query your data driven dictionary. Also, 
with your permission and if you like, I could host an 
online version of your data driven dictionary with a 
user friendly html front end interface on my server for 
you, making your work accessible online. I could also 
provide you with a secret interface making it routine 
for you to update your definitions.

James Christian Read - BSc Computer Science 
http://www.lamie.org/hebrew       -  thesis1: concept driven machine translation using the Aleppo codex 
http://www.lamie.org/lad-sim.doc  -  thesis2: language acquisition simulation
http://www.lamie.org/ngelta































































































rk accessible online. I could also 
provide you with a secret interface making it routine 
for you to update your definitions.

James Christian Read - BSc Computer Science 
http://www.lamie.org/hebrew       -  thesis1: concept driven machine translation using the Aleppo codex 
http://www.lamie.org/lad-sim.doc  -  thesis2: language acquisition simulation
http://www.lamie.org/ngelta
























































































More information about the b-hebrew mailing list