 |
Where is the online documentation located? |
|
Version 1.2 BETA 
Code Samples
Online documentation
Download compiled HTML
help
Portable
Application Description (PAD) |
|
|
|
|
|
Is the tokenization process fast? What kind of
system overhead is required? |
|
Since we recommend pre-tokenizing your existing data, you will only be
tokenizing the search criteria entered by your users and taking that integer
value and matching it with the integers stored in your database. Integer
matching in database queries is much faster that character data matching.
So yes, it's very fast. |
|
|
|
|
|
Does it only work on names or can I use this for any
word? |
|
It will work on any word, even medical and legal terms. |
|
|
|
|
|
Does it provide me with a list of names? |
|
No, this is a tool for comparing data. You provide the names, etc. and it
will tell you if they match because they sound alike or if they match because
they are in the same nickname group. |
|
|
|
|
|
Can I just use it for the Nickname matching? |
|
Yes, several have just had a need for the nickname matching feature to obtain
the Diminutive Group. This enables you to determine name matches like
"Dick" = "Richard". |
|
|
|
|
|
How big is your demo database used on this site? |
|
We have more than 43,000 unique first names and 170,000 unique last names in
our demo database. They have been collected from several disparate
sources over the years. We wanted kind of a worst-case scenario to
demonstrate the searching capabilities you will have available to your
applications if you integrate the tokens produced by the SoundsLike component. |
|
|
|
|
|
What are the steps required to add phonetic and nickname
matching to my applications? |
|
Here are some recommendations:
1
|
Decide what columns/fields you want to add intelligent matching to
(first name, last name, product name, etc) |
2
|
Add necessary columns to your database. You need at least
the "Narrow Token" [Int32] for each column you need to phonetic search
on. The "Wide Token" is optional and can be used for more relaxed
matching. For first names be sure to also add a column for Diminutive
Group (nickname matching) |
3
|
Add logic to your application to tokenize the text data each time
it changes or when new records are added |
4
|
Write a quick program to spin through your database and tokenize
the text data for the first time |
5
|
Add logic to tokenize the search data entered in your app and use
those integers to match against the database |
6
|
Reap the benefits of offering your users and clients a more
robust, intelligent searching application |
|
|
|
|
|
|
Can I include the SoundsLike DLL in an application that I
am developing to sell? |
|
Yes, licensed copies of the DLL are royalty free. You may embed this
Software in any software product that you sell or redistribute outside your
organization. We would like to know about it, however. |
|
|
|
|
|
Are the narrow and wide token integer values the
equivalent of the primary key and alternate key of the Double Metaphone
algorithm? |
|
No, the narrow and wide tokens are close and loose matches of the same
word. SoundsLike is not an implementation of the Double Metaphone
algorithm. |
|
|
|
|
|
Does your component modify or use the registry? |
|
No. |
