Skip to end of metadata
Go to start of metadata

Introduction

This document describes how to perform a search to get all the marks which sound similar to given name with IPAS version 2.7.0. It introduces pre-requisites in terms of system configuration parameters that are required and used for search of the marks. It covers the functionality as well as basic data structure of application types in IPAS.

Pre-requisites

Configuration 

Setting the phonetic rules in IPAS Designer

 IPAS Designer offers the following configuration options to set the phonetic rules in the ConfigParam perspective.

1.-Maximum phonetic similarities :

Maximum number of phonetic similarities to be stored in the similarities tables

2.-Phonetisation rules set to use for mark phonetisation:

 Rules available are the following:

  • 0: Portuguese rules (simplified set of phonemes).
  • 1: Portuguese rules (extended set of phonemes).
  • 3: Spanish rules.
  • 4: English rules.

Please take into account that the same version of file phoneticSearchRules.xml must be located in both the IPAS Manager Installation folder (where it is read by IPAS Manager to perform the massive phonetization) and in the IPAS Application Server “config” folder (where it is read by GlassFish to perform the phonetization of new trademarks and the phonetic search).

3.-Transliteration rules set to use for mark phonetisation:

These rules are applied to the mark names before applying the phonetization rules, as described below.

  • General structure of the rules XML files

A rules configuration file may have several rule sets, each of which is used for different purposes (prefixes, transliteration or phonetization).

A rule set consists in several rules, and each rule has two parts:

  • · The “if” part defines the triggering condition of the rule and specifies the text to be replaced.
  • · The “then” part specifies the replacement text to be used in the transformation.

For example, if the letter “Y” is desired to be phonetized as the phoneme “i” then the following simple rule would be required:

<rule>

<if>Y</if>

<then>i</then>

</rule>

The “if” part of the rule consists is a series of terms separated by whitespaces, and each term can be:

  • · A single character, indicating that the input data should contain this character in order for the rule to be fired. Single characters are replaced by the replacement text.
  • · A meta-character term, indicating any of a set of characters that the input data should contain in order for the rule to be fired. Meta-characters are replaced by the replacement text. The possibilities are as follows:

Meta-character

Matches …

VOCAL

… any vocal character

CONS

… any consonant character

CONS_NOT_<X>

… any consonant but not the value <X>. For Example CONST_NOT_W.

LETTER

… any letter

NOT_LETTER

… any character which is not a letter

DIGIT

… any digit character

  • · A context term, indicating additional conditions that must be fulfilled in order for the rule to be fired, but which are not to be replaced by the rule. The possibilities are as follows:

Context term

Matches …

^

… the beginning of the input text

$

… the end of the input text

/ following terms

… the specified “following terms” should exist AFTER the (meta)characters at the left of the /

preceding terms \

 

… the specified “preceding terms” should exist BEFORE the (meta)characters at the right of the \

IF_EXISTS terms

… the specified terms exists in the remaining input text to be processed

The “then” part of the rule defines the replacement text, consists is a series of terms separated by whitespaces, and each term can be:

  • · A single character, indicating the replacement text.
  • · A digit, indicating one of the characters in the matched input text based on its positions (1=first position, 2-second position, etc). This is needed in order to specify the exact value of a matched meta-character.
  • · An underscore character, representing a word break. This means that words are separated by this symbol.

Some examples follow in order to clarify the above concepts.

Beginning of text context

In the following example, if a text starts with Z followed by a consonant (not W) then it is

replaced by a text that starts with ES <consonant>

<rule>

<if>^ Z CONS_NOT_W</if>

<then>E S 3</then>

</rule>

End of text context

In the following example, if a text ends with L then it is replaced by a text ending with U

instead of L.

<rule>

<if>L $</if>

<then>U</then>

</rule>

Context expressed by a “followed by” term

In the following example the C when it is followed by E is replaced by S

<rule>

<if>C / E</if>

<then>S</then>

</rule>

Context expressed by a “preceded by” and “followed by” terms

In the following example the U preceded by a consonant and followed by a consonant

and a vowel, is replaced by IU.

IPAS JAVA - Installation Guide 164

<rule>

<if>CONS \ U / CONS VOCAL</if>

<then>I U</then>

</rule>

Context expressed by a “if_exists” term

In the following example the matched digit is eliminated if the sequence NOT_LETTER

F M NOT_LETTER exists in the rest of the text

<rule>

<if>DIGIT IFEXISTS NOT_LETTER F M NOT_LETTER</if>

<then></then>

</rule>

Prefix and suffix detection rules

The first set of rules which are executed by the module is the one configured for the detection of prefixes and suffixes, which are very noticeable parts of the name which almost “stand for themselves” and therefore deserve special consideration.

For example, the following rule would separate the “MC” prefix from the rest of the name as a separate word pronounced as “MAC”:

<rule>

<if>^ M C</if>

<then>M A C _</then>

</rule>

In this way, a name like MCFLURRY would be transformed in MAC FLURRY.

 

4.-Transliteration rules

 

The second set of rules which are executed by the module is the one configured for the transliteration, that is to say the transformation of characters into a simplified alphabet, normally the Latin one.

Depending on the language, the transliteration rules may include transformations to:

  • ·In the case of Spanish, all accented vowels (like Á, Â, Ä, etc) are replaced by the non-accented letter.
  • · In the case of Portuguese, some accented vowels are not replaced because the sound is different depending on the context and the phonetization rules need the accented letter (for example Ü sound differently depending on the context and cannot be blindly transliterated into U).
  • · In the case of Cyrillic languages, the transliteration rules are more complex but can be easily configured using the rules syntax discussed above.

 

5.-Default or minimum similarity percentage:

This is the default value for the minimum similarity percentage to be retrieved by the phonetic   search.  A value of 40 normally assures that the more similar trademarks are retrieved. A smaller   value would cause many more trademarks to be retrieved, with a maximum of 1000 similarities.

FuNCTIONALITY

Basic Flow

To find all the marks which sound like ‘Cola’, please follow the below steps:

  • Login into IPAS web interface.
  • Click on the navigation links “Trademarks” > “Phonetic Search”
  • Specify the word ‘Cola’ in the field “Mark Name Sounds Like”  
  • Click on “Search” link

The list of all trademarks which sound like ‘Cola’ will be displayed as shown below:

Additional Flow

1.- Print Report :

Reports can be generated by click on the Print Report link.

  • Click on the Print Report Link, Report file path will be available on the next page.
  • Select appropriate path and click on the proceed link, similarity search report will be generated.

2.- Show / Hide Logos:

These links will show or hide the logos of the listed file records respectively.

 

3.- Set Filter:

Set Filter link will allow to filter listed files based various criteria like pending applications, registered trademarks etc.

Alternative Flow

You can also use phonetic search also to perform the similar search to get all the mark sound similar.

Expected Result

Once above process is completed successfully, all the trademarks which sound like ‘Cola’ would be listed on the Search Results screen.

Further actions

Related Use Cases

TroubleSHOOTING

Affected Data Structure

When phonetic similarity search is done then database tables affected are:

  1. Search criteria like Similarity value and mark name will be inserted into the tables starting with WRK_BUSQ prefix which are internal work tables.
  2. Similar Files details searched will be inserted into tables starting with WK_PHONETIC_SEARCH prefix

 

 

  • No labels