Re: libquran, a release?

[Mohammed Yousif <mhdyousif at gmx dot net>]

 I'm moving this to the mailing-list, as it gets too deep.

On Monday 30 June 2003 10:35, Maher Ben Moussa wrote:
> Sorry too here. I was too busy too with school.
> I like the XML format. It's what we need. Only I don't know if it's
> possible by the XML standard to use a tagname as an attribute.
>
> I think about something like this:
>
> <quran>
>   <language name="Indonesian" fullname="Bahasa Indonesia" lang="id">
>     <audio>
>       <author>Ali Hofaizy</author>
>       <copyright>(c) xxxx</copyright>
>       <format>Speex</format>
>       <filename>sCCCaVVV.spx</filename>
>     </audio>
>     <translation>
>       <author>Departemen Agama Republik Indonesia</author>
>       <copyright>(c) xxxx</copyright>
>       <filename>quran.id.xml</filename>
>     </translation>
>
>     <tafsir>
>       <author>xxx</author>
>       <copyright>(c) xxxx</copyright>
>       <filename>tafser.id.xml</filename>
>     </tafsir>
>   </language>
> </quran>
>

 I like it very much, and here are how I think of implementing it,
  + There should be something like /etc/quran or ~/.quranrc
     containing one line indicating the path of the data files
     (i.e. /mnt/cdrom) instead of the current libquran.xml.
  + The clients should link to the library statically and, thus it
      is not supposed to be installed for the clients to work and
      it doesn't make sense to _still_ need it for the data and the
      libquran.xml file and, as Damt noted, data should be separated
      and the conf file should be read from ~/.quranrc first then
      from /etc/quran to avoid confusion.
  + The path noted by /etc/quran should contain packages for
      each language:
          ar/
          ar/quran.xml
          ar/tafsir.ibnkathir.xml
          ar/tafsir.jalalain.xml
          ar/hozaify/s???a???.spx
          ar/khayaat/s???a???.spx
          ar/conf.xml
      and conf.xml maybe:
        <quran>
         <language name="اللغة العربية">
          <audio>
           <author>Ali Hofaizy</author>
           <copyright>(c) xxxx</copyright>
           <filename>hozaify/sCCCaVVV.spx</filename>
          </audio>
          <audio>
           <author>Mohammed Saeed Khayaat</author>
           <copyright>(c) xxxx</copyright>
           <filename>khayaat/sCCCaVVV.spx</filename>
          </audio>
          <text>
           <author></author>
           <copyright>Original Text</copyright>
           <filename>quran.xml</filename>
          </text>
          <tafsir>
           <author>Al-Hafez Ibn Kathir</author>
           <copyright>(c) xxxx</copyright>
           <filename>tafser.ibnkathir.xml</filename>
          </tafsir>
          <tafsir>
           <author>Al-Jalalain</author>
           <copyright>(c) xxxx</copyright>
           <filename>tafser.jalalain.xml</filename>
          </tafsir>
         </language>
        </quran>

      This allows for greater flexibility, but about the langauge name,
      there is no need for more than the name option and it should
      be in the target language, the code should be the name of the
      language package directory.
      We can implement this with some changes in the API (little, I hope)
      quran_init should use the same routine as the already one for
      detecting audio packages to detect language packages
      quran_open should not only take the language code, but also the
      filename from the conf.xml file, the language code is used to get
      in the directory for that language and the filename (which should
      be gotten using quran_init from conf.xml for that specific language)
      for opening a specific translation (this way, it can support multiple
      translations and interpretaions which is used the same way).
      quran_open_audio almost will stay the same.

  +  libquran should not depend on data provided by ogg tags
      (author, copyright...etc) as there will be more than 6000
      file and it will be a killer to check all these files and you canl
      not be sure if they all have the same data, but rather I go
      for setting all these in {LANGUAGE_CODE}/conf.xml, and
      getting rid of the 'name' file (That was a dirty hack by me
      which is by no means the best way to go).

> And about freezing the API. Can you please wait one week. I want to
> check the API first, because I noticed some little things months ago
> that have to be changed.
>

 No problem, that was why I sent this email in the first place.

> PS: I will commit the bug of libquran this evening or tomorrow

 Please update your copy, I already sent it with some
 improvements.

>> Salam,
>>
>> I'm sorry I've been so busy lately as we have terrible economical 
>> situation in my country.
>>
  I'm so sorry to hear that, I hope Allah will help you Damt.

-- 
Mohammed Yousif
"Ich liebe ihr, weil ohne sie kann ich nicht bin!"
We _will_ restore OUR Jerusalem.