dictionary programming?

Question

I want to write a dictionary software, web based or application based...

but since I am a newbie, I want to ask about what datastructure is normally used in dictionary software?

and what kind of database should i use? mysql? or just normal text file..

Thank You

Answer 1

I've looked into that myself. Too bad I lost all the bookmarks I had going while doing that research. But, from what I learned, or the thing that impressed me the most was a datastore of word roots as a text file with multiple java programs configured for composing the lexicon, the concordance, the dictionary and the theasarus and all "glued" together with Jython. I really don't think mySQL would be nimble enough after one structure. I really think because Linux automatically makes a list that it will lend itself with better parts to work with in the way of tools. Many dictionaries out there. Many trying to keep the datastore size down with web references to remote data.

The best example of a dictionary I have ever encountered is WordWeb. If his software cannot find the word you are looking for in his cross-referenced datastore, it will search the web. The author would like collaboration with anyone willing to convert the project as a Linux software.

Answer 2

You definitely want to use a mysql database. Most certainly will be a lot faster to search / sort records than a text file. Your focus should primarily lie in trying to anticipate what word your user is looking for as they type. That could give faster results. Another thing to think about is why recreate the wheel. If it is web based, maybe there is a company out there with a dictionary already and they have an API. You can create your site and connect it to theirs in the background and produce their results. Hope that helps.

Answer 3

How big is the dictionary. If it is a glossary with 50-100 entries that is one thing. If it is 30,000-100,000 words that is a lot messier.
If you are just looking up words occasionally, then a database will give suitable access speed with the possible option of using Soundex to get close words to misspellings. If your goal is to make a spell checker, then you have to hash the words and use pointers to a file with the rest of the information - hashing involves taking a word and converting it to a number to go into the middle of the file very quickly - the secret is how you construct the hash so similar words are different.

Answer 4

you can use text file, but when you end up in a million words, maybe it's better to use MS Access or largely SQL server