You are viewing a plain text version of this content. The canonical link for it is here.
Posted to users@spamassassin.apache.org by Paul Aviles <pa...@nickelnetworks.com> on 2007/02/27 17:38:49 UTC

Locales and Languages

I am using http://www.yrex.com/spam/spamconfig.php to see how the
configuration of ok_locales and ok_languages changes and my brain is getting
stuck on it. 
 
Shouldn't they both match? If I select that I want to only exclude from
scoring say "en de es"  and run the configuration, then it generates the
following:
 
"ok_languages es"
"ok_locales all"
 
What happened to the other two options en and de?
 
If I select every locale from the top and remove Use Language Testing and I
only get the same. 
 
ok_languages            all
ok_locales              all
 
If I select the bolded locales I get:
 
ok_languages            zh ja ko ru th 
ok_locales              ja ko ru th zh 
 
If I select the bolded locales and remove the Language Testing I get:
 
ok_languages            all
ok_locales              en ja ko ru th zh 
 
So there are differences. Am I missing something or the script is wrong?
 
Regards,
 
Paul
 

RE: Locales and Languages

Posted by Paul Aviles <pa...@nickelnetworks.com>.
Never mind... "If this option is disabled".. I was reading it backwards...

Thanks
Paul

-----Original Message-----
From: Matt Kettler [mailto:mkettler_sa@verizon.net] 
Sent: Tuesday, February 27, 2007 7:31 PM
To: Paul Aviles
Cc: users@spamassassin.apache.org
Subject: Re: Locales and Languages

Paul Aviles wrote:
> I am using http://www.yrex.com/spam/spamconfig.php to see how the 
> configuration of ok_locales and ok_languages changes and my brain is 
> getting stuck on it.
>  
> Shouldn't they both match?
No. ok_locales doesn't support nearly as many different options as
ok_languages. ok_locales *only* supports the bold selections on that config
generator.

The fundamental problem is that ok_locales works by looking at what
character encoding you're using. But, for example, English, French, and
Spanish generally use the same character set, commonly ISO-8859-1.

Thus, there's no way for ok_locales to get into fine details like the
difference between French and Spanish. To it, they're the same, they're both
English. Only languages with comprehensively different character sets can be
detected this way, like English vs Japaneese. (note:
theoretically it should be able to detect Arabic too, but I don't think
there's support for that yet)

ok_languages on the other hand analyzes the content of the text, and can
detect things like the difference between German and English. There are
groupings of letters that aren't common in english, but common in german,
like "sch".

But of course, I've explained that much to you before.
> If I select that I want to only exclude from scoring say "en de es" 
> and run the configuration, then it generates the following:
>  
> "ok_languages es"
> "ok_locales all"
>  
> What happened to the other two options en and de?
I don't know. It works for me. I run the script and select "English"
"French" and "Spanish" and I get:

ok_languages            en de es

ok_locales              en

Perhaps there's a bug in your browser's javascript handling? I used firefox
2.0.0.2/win32.






RE: Locales and Languages

Posted by Paul Aviles <pa...@nickelnetworks.com>.
I got it now, it does makes sense. Hmmm, did you unchecked Use Language
Testing? If I do that, I only get 
ok_languages            all
ok_locales              en 

Paul

-----Original Message-----
From: Matt Kettler [mailto:mkettler_sa@verizon.net] 
Sent: Tuesday, February 27, 2007 7:31 PM
To: Paul Aviles
Cc: users@spamassassin.apache.org
Subject: Re: Locales and Languages

Paul Aviles wrote:
> I am using http://www.yrex.com/spam/spamconfig.php to see how the 
> configuration of ok_locales and ok_languages changes and my brain is 
> getting stuck on it.
>  
> Shouldn't they both match?
No. ok_locales doesn't support nearly as many different options as
ok_languages. ok_locales *only* supports the bold selections on that config
generator.

The fundamental problem is that ok_locales works by looking at what
character encoding you're using. But, for example, English, French, and
Spanish generally use the same character set, commonly ISO-8859-1.

Thus, there's no way for ok_locales to get into fine details like the
difference between French and Spanish. To it, they're the same, they're both
English. Only languages with comprehensively different character sets can be
detected this way, like English vs Japaneese. (note:
theoretically it should be able to detect Arabic too, but I don't think
there's support for that yet)

ok_languages on the other hand analyzes the content of the text, and can
detect things like the difference between German and English. There are
groupings of letters that aren't common in english, but common in german,
like "sch".

But of course, I've explained that much to you before.
> If I select that I want to only exclude from scoring say "en de es" 
> and run the configuration, then it generates the following:
>  
> "ok_languages es"
> "ok_locales all"
>  
> What happened to the other two options en and de?
I don't know. It works for me. I run the script and select "English"
"French" and "Spanish" and I get:

ok_languages            en de es

ok_locales              en

Perhaps there's a bug in your browser's javascript handling? I used firefox
2.0.0.2/win32.






Re: Locales and Languages

Posted by Matt Kettler <mk...@verizon.net>.
Paul Aviles wrote:
> I am using http://www.yrex.com/spam/spamconfig.php to see how the
> configuration of ok_locales and ok_languages changes and my brain
> is getting stuck on it.
>  
> Shouldn't they both match?
No. ok_locales doesn't support nearly as many different options as
ok_languages. ok_locales *only* supports the bold selections on that
config generator.

The fundamental problem is that ok_locales works by looking at what
character encoding you're using. But, for example, English, French, and
Spanish generally use the same character set, commonly ISO-8859-1.

Thus, there's no way for ok_locales to get into fine details like the
difference between French and Spanish. To it, they're the same, they're
both English. Only languages with comprehensively different character
sets can be detected this way, like English vs Japaneese. (note:
theoretically it should be able to detect Arabic too, but I don't think
there's support for that yet)

ok_languages on the other hand analyzes the content of the text, and can
detect things like the difference between German and English. There are
groupings of letters that aren't common in english, but common in
german, like "sch".

But of course, I've explained that much to you before.
> If I select that I want to only exclude from scoring say "en de es" 
> and run the configuration, then it generates the following:
>  
> "ok_languages es"
> "ok_locales all"
>  
> What happened to the other two options en and de?
I don't know. It works for me. I run the script and select "English"
"French" and "Spanish" and I get:

ok_languages            en de es

ok_locales              en

Perhaps there's a bug in your browser's javascript handling? I used
firefox 2.0.0.2/win32.