Hi list,

I first tried to introduce directly the <i18n:text> in the document like 
this
      <p>  <i18n:text>Seach</i18n:text></p>

The result is 

Internal Server Error
Message: null
Description: No details available.
Sender: org.apache.cocoon.servlet.CocoonServlet
Source: Cocoon Servlet
Request URI
evolution/indexi.html
cause
The prefix "i18n" for element "i18n:text" is not bound.
request-uri
/evolution/indexi.html
Apache Cocoon 2.2.0-dev

Then I realised that the modification of the browser language tranlate the 
tabs and the menu.
So I try to go on with this method.

I like the idea to separate text from non text. 
I did'nt use svg files. May be it interresting too.
In my  exemple  we only modify indexi.xml to replace an image.
If we had one file for each language (index_en.xml,index_de.xml 
,index_se.xml) we should modify each file.
In general we can keep the layout, and just change the text.

I know  there are two other possibilities than the browser lang to 
indicate the language.
index.en.xml, index.de.xml , index.se.xml 
index.xml?lang=en, index.xml?lang=de, index.xml?lang=se,

In those cases the sitemap.xmap should analyse the parameters and the 
suffixs to dectect the language.
The sitemap.xmap should delclare a priority the resolve conflicts.
Example langformatpriority=langparameter, browserlang,filesuffixlang

My first idea was to have a separate first "treatment" to translate the 
source text everywhere:
 -in "boxes" organized in the general layout given by the skin (header, 
tabs, menu, footer)
- in "boxes" organized in layout inside the page.
But then I follow the sitemap.xmap and I modify as little as possible.

Why did you switch the order and nested the map:select?
I switch the order so that the pfd suits the browser lang value. But 
eventhought, 
I had a still a problem with the pdf display. At the beginning,  it suits 
the browser lang value. 
Then it sticks to one lang.
I don't know yet, if it a server cache problem or a mistake from my 
implementation.

May be the following version with the "dispatcher" will simplifiy the way 
to have the same display. 


Paul






Thorsten Scherler <thorsten@apache.org>
28/02/2006 19:49
Please respond to user
 
        To:     forrest dev <dev@forrest.apache.org>
        cc:     forrest user <user@forrest.apache.org>
        Subject:        Re: forrest 0.7 multi language site contribution 
and question    (second version)


El lun, 27-02-2006 a las 08:34 +0100, Paul.Divine@alcatel.fr escribió:
> 
> Hi list, 
> May be this mail should be send to a dev list. 

:) 

This discussion will go on on dev. ;)

This is a really nice mail, thank you very much for this infos.

...the only thing that I wonder is whether we can simplify the
process. ;)

> 
> 
> In order to display with dynamic forrest (started with forrest run) 
> a page with a text corresponding to the language of the preference 
> of the browser of the user, I did the following modifications. 
> This only concerns the page, not the menu or the tabs. 
> The basic idea is to 
> - created a new DTD for document to introduce an element tt which 
> surround the text to be translated. 

Why did you not decided to use xml:lang? 
http://www.w3.org/TR/REC-xml/#sec-lang-tag


> - to translate this tt element to an i18:text element with an xslt
> stylesheet 
> - apply the i18n transformation to translate. 

Why did you not directly used
http://cocoon.apache.org/2.1/userdocs/i18nTransformer.html
<para title="first" name="article"  i18n:attr="title name">
  <i18n:text>This text will be translated.</i18n:text>
</para>

> Note 
> In the examples 
> forrest is      in D:\DATA\apache\apache-forrest-0.7 
> forrest site is in D:\DATA\apache\apache-forrest-0.7-paul-site 
> 
> A) I create a new DTD document-v21.dtd with a new module 
> document-v21.mod 
> in directory 
> D:\DATA\apache\apache-forrest-0.7\main\webapp\resources\schema\dtd\ 
> 
> The document-v21.dtd is 
> 
> <!ENTITY % common-charents PUBLIC 
>     "-//APACHE//ENTITIES Common Character Entity Sets V1.0//EN" 
>     "common-charents-v10.mod"> 
> %common-charents; 
> 
> <!-- ===============================================================
> --> 
> <!-- Document --> 
> <!-- ===============================================================
> --> 
> 
> <!ENTITY % document PUBLIC 
>     "-//APACHE//ENTITIES Documentation V2.1//EN" 
>     "document-v21.mod"> 
> %document; 
> 
> 
> Note 
> Only the two lines are different from document-v20.dtd 
>     "-//APACHE//ENTITIES Documentation V2.1//EN" 
>     "document-v21.mod"> 
> 
> 
> in directory 
> D:\DATA\apache\apache-forrest-0.7\main\webapp\resources\schema\dtd\ 
> The document document-v21.mod is a copy of document-v20.mod 
> with differences given hereafter 
> 
> 1) The line 
> <!ENTITY % text "#PCDATA"> 
> becomes 
> <!-- <!ENTITY % text "#PCDATA"> --> 
> This line is put in comment 
> 
> 2) The folowing lines are added after 
> <!ENTITY % simpletext "#PCDATA"> 
> <!ENTITY % text "simpletext|tt"> 
> 
> 3) before the lines 
> <!-- ==================================================== --> 
> <!-- Body --> 
> <!-- ==================================================== --> 
> 
> add 
> 
> <!--   tt text to translate --> 
> <!ELEMENT tt (%simpletext;)> 
> 
> 
> 
> 
> The basic idea is to surround text to be translated with the tt
> element. 
> 
> 
> B) A new stylesheet i18ntext.xsl to replace <tt> by <i18n:text> 
> in directory D:\DATA\apache\apache-forrest-0.7\main\webapp\resources
> \stylesheets 
> 
> <?xml version="1.0"?> 
> <xsl:stylesheet version="1.0" 
>                 xmlns:xsl="http://www.w3.org/1999/XSL/Transform" 
>                 xmlns:i18n="http://apache.org/cocoon/i18n/2.1" > 
> 
>   <xsl:import href="copyover.xsl"/> 
>   <xsl:template match="tt"> 
>      <i18n:text> 
>      <xsl:apply-templates select="@*"/> 
>      <xsl:apply-templates /> 
>      </i18n:text> 
>   </xsl:template> 
> </xsl:stylesheet> 
> 

see above, why did you not directly declared i18n:text in the dtd and
used tt?

> C) Modify the sitemap.xmap in order to do the two modifications 
>    (markup modification and translation) for html page and pdf page 
> in directory 
> D:\DATA\apache\apache-forrest-0.7\main\webapp 
> 
> After the lines 
>       <map:match pattern="**body-*.html"> 
>         <map:generate src="cocoon:/{1}{2}.xml"/> 
>         <map:transform type="idgen"/> 
>         <map:transform type="xinclude"/> 
> 
> Add the lines 
>  <!-- paul begin --> 
>         <map:select type="config"> 
>           <map:parameter name="value" value="{defaults:i18n}"/> 
>           <map:when test="true"> 
>             <map:transform src="{forrest:stylesheets}/i18ntext.xsl"/>
> 
>             <map:transform type="i18n"> 
>               <map:parameter name="locale" value="{request:locale}"/> 
>             </map:transform> 
>             </map:when> 
>         </map:select> 
> <!-- paul end --> 
> 
> and after the lines 
>       <map:match type="regexp" pattern="^(.*?)([^/]*).fo$"> 
>         <map:select type="exists"> 
>           <map:when test="{project:content.xdocs}{1}{2}.fo"> 
>             <map:generate src="{project:content.xdocs}{1}{2}.fo"/> 
>           </map:when> 
>           <map:otherwise> 
>             <map:aggregate element="site"> 
>               <map:part src="cocoon:/skinconf.xml"/> 
>               <map:part src="cocoon:/{1}{2}.xml"/> 
>             </map:aggregate> 
>             <!-- <map:transform type="idgen"/> --> 
>             <map:transform type="xinclude"/> 
> 
> add the lines 
> <!-- paul begin --> 
>         <map:select type="config"> 
>           <map:parameter name="value" value="{defaults:i18n}"/> 
>           <map:when test="true"> 
>             <map:transform src="{forrest:stylesheets}/i18ntext.xsl"/>
> 
>             <map:transform type="i18n"> 
>               <map:parameter name="locale" value="{request:locale}"/> 
>             </map:transform> 
>                 </map:when> 
>         </map:select> 
> <!-- paul end --> 

Why did you decide to use inline translations over document specific.
Like:
index_en.xml; index_de.xml; index_es.xml; ...

> D) in the pdf plugin  output.xmap we inverse the presence test 
> to solve a problem of cache (this is a workaround) 
> (the cache does not recognise that the language has changed, so it 
> displays the same page) 
> in the directory 
> D:\DATA\apache\apache-forrest-0.7\build\plugins
> \org.apache.forrest.plugin.output.pdf\ 
> 
> After the lines 
>   <map:pipelines> 
>     <map:pipeline> 
>       <!-- generate .pdf files from .fo --> 
>       <map:match type="regexp" pattern="^(.*?)([^/]*).pdf$"> 
>         <map:select type="exists"> 
> put in comment the lines 
> <!-- 
>           <map:when test="{project:content.xdocs}/{1}{2}.pdf"> 
>             <map:read src="{project:content.xdocs}/{1}{2}.pdf"/> 
>           </map:when> 
>           <map:otherwise> 
>             <map:generate src="cocoon://{1}{2}.fo"/> 
>             <map:serialize type="fo2pdf"/> 
>           </map:otherwise> 
> --> 
> and after add the lines 
> <!-- paul begin --> 
>           <map:when test="{project:content.xdocs}/{1}{2}.xml"> 
>             <map:generate src="cocoon://{1}{2}.fo"/> 
>             <map:serialize type="fo2pdf"/> 
>           </map:when> 
>           <map:otherwise> 
>             <map:select type="exists"> 
>               <map:when test="{project:content.xdocs}/{1}{2}.pdf"> 
>                 <map:read src="{project:content.xdocs}/{1}{2}.pdf"/> 
>               </map:when> 
>             </map:select> 
>           </map:otherwise> 
> <!-- paul end   --> 


Why did you switch the order and nested the map:select?

> E) In forrest.xmap in D:\DATA\apache\apache-forrest-0.7\main\webapp 
> In order to catch the new DTD. 
> After the lines 
>         <sourcetype name="document-v20"> 
>           <document-declaration public-id="-//APACHE//DTD
> Documentation V2.0//EN" /> 
>         </sourcetype> 
> Add the lines 
> <!-- add paul begin --> 
>         <sourcetype name="document-v21"> 
>           <document-declaration public-id="-//APACHE//DTD
> Documentation V2.1//EN" /> 
>         </sourcetype> 
> <!-- add paul end --> 
> 
> Afer the lines 
>           <map:when test="document-v20"> 
>             <map:transform
> src="{forrest:stylesheets}/docv20todocv13.xsl" /> 
>           </map:when> 
> Add the lines 
> <!-- add paul begin --> 
>           <map:when test="document-v21"> 
>             <map:transform
> src="{forrest:stylesheets}/docv20todocv13.xsl" /> 
>           </map:when> 
> <!-- add paul end --> 
> 
> F) in D:\DATA\apache\apache-forrest-0.7\main\webapp\skins\common
> \translations 
> fill catalogs CommonMessages_xx.xml with in the languages you want. 
> For the example 
> I use CommonMessages_en_US.xml and CommonMessages_fr.xml 
> with already translated "Search". 
>   <message key="Search">Rechercher</message> 
> 
> G) Update Catalog of forrest 
> In D:\DATA\apache\apache-forrest-0.7\main\webapp\resources\schema
> \catalog.forrest 
> After the line 
> -- Document Type Definitions -- 
> 
> add the line 
> PUBLIC "-//APACHE//DTD Documentation V2.1//EN" 
>        "dtd/document-v21.dtd" 
> 
> 
> H)in catalog.forrest.xcat in the directory 
> D:\DATA\apache\apache-forrest-0.7\main\webapp\resources\schema 
> After the line 
> <!-- Document Type Definitions (current) --> 
> 
> add the lines 
> <!-- paul begin --> 
> <public publicId="-//APACHE//DTD Documentation V2.1//EN" 
>         uri="dtd/document-v21.dtd"/> 
> <!-- paul end --> 
> 
> After the line 
> <!-- Reusable modules --> 
> add the lines 
> 
> <!-- paul begin --> 
> <public publicId="-//APACHE//ENTITIES Documentation V2.1//EN" 
>         uri="dtd/document-v21.mod"/> 
> <!-- paul  end  --> 
> 
> 
> I) Put in your page the new DOCTYPE and the text to translate in <tt>
> markup. 
> Exemple indexi.xml 
> in the directory 
> D:\DATA\apache\apache-forrest-0.7-paul-site\src\documentation\content
> \xdocs 
> 
> <?xml version="1.0" encoding="UTF-8"?> 
> <!DOCTYPE document PUBLIC "-//APACHE//DTD Documentation V2.1//EN"
> "http://forrest.apache.org/dtd/document-v21.dtd"> 
> <document> 
>   <header> 
>     <title><tt>Search</tt> untranslate</title> 
>   </header> 
>   <body> 
>     <section id="overview"> 
>       <title><tt>Search</tt>  untranslate</title> 
>       <p><tt>Search</tt>  untranslate 
>      </p> 
>     </section> 
>   </body> 
> </document> 
> 
> 
> 
> J) In forrest.properties of the site put the indicator to true 
> Example 
> D:\DATA\apache\apache-forrest-0.7-paul-site 
> After the line 
> #project.i18n=true 
> add the line 
> project.i18n=true 
> 
> K) In the directory of your site start the site with command "forrest
> run" 
> 
> Example 
> cd D:\DATA\apache\apache-forrest-0.7-paul-site 
> SET JAVA_HOME=C:\java\j2sdk1.4.2_03 
> set FOREST_HOME=C:\apache\apache-forrest-0.7 
> set PATH=%PATH%;%FOREST_HOME%\bin 
> forrest run 
> 
> L) To look at the results 
> 
> - With your favorit browser put your language preferences to 
> english  en_US 
> (tool/internet option/language ..) 
> 
> http://localhost:8888/indexi.html 
> You can see "Search" 
> and if you clic on the PDF icon you should see the same in pdf
> format. 
> 
> - With your favorit browser put your language preferences to german
> de 
> 
> http://localhost:8888/indexi.html 
> You can see "Suche" 
> and if you clic on the PDF icon you should see the same in pdf
> format. 
> 
> - With your favorit browser put your language preferences to french
> fr 
> 
> http://localhost:8888/indexi.html 
> You can see "Rechercher" 
> and if you clic on the PDF icon you should see the same in pdf
> format. 
> 
> 
> Note 
> For the tab and the menu 
> Catalogues of the language should be define 
> D:\DATA\apache\apache-forrest-0.7-paul-site\src\documentation
> \translations 
> langcode.xml 
> 
> menu_de.xml menu_fr.xml the menu_lang file should exist 
> 
> Restrictions 
>  - The research function in the site is looking at the text in the
> page 
>    Only the basic langague is seen, not the translations. Some work
> is 
>    needed to add to the index the text of each language. 

Yeah, that would be easier if we use index_en.xml,...

>  - The other tags of i18n like "if" are not taken into account. 

yeah, why not using i18n:text right away?

>  - To inform the user of the multi language possibilities 
>    a new skin should be done. 

yeah, right, but I think we should rather extend the dispatcher
(implementing/testing i18n is next on my list). It is very easy with jx
in the structurer to responding/requesting index_en.xml; index_de.xml;
index_es.xml; ... 

>  - Only document-v20.dtd can be translated all the other formats
> cannot. 

Hmm.

> 
> The question is: Why the Pdf function has no translation? 

I do not understand this question.

salu2
-- 
thorsten

"Together we stand, divided we fall!" 
Hey you (Pink Floyd)