Re: Was: [xsl] mode and moved to Namespaces

[ac <ac@xxxxxxxxxxxxx>]

Hi Jirka,

One of the first performance consideration issue is memory usage.  There 
is a lot happening in the application and the number of supported 
languages could be high.  How to keep everything loaded is not an issue.

If our main requirement was elegance, we would probably have lots of 
elements and attributes and metadata attached to the dictionary, and the 
dictionary would not be designed only for quick translation to a high 
number of languages in tight space, but for vocabulary, definitions, 
synonyms, antonyms, syntactic analysis, semantics management, and 
translation.  In fact, we do that in other sections of the application.  
That is not the problem we are trying to solve here.

The current translation dictionary is setup somewhat like:
...
<word en:title="Mr" f-en:title="Mrs" fr:title="M." f-fr:title="Mme" ... />
<word en:noun="chair" fr:noun="chaise" ... />
...

all feminine variants can be returned with:
/dic/word/@*[starts-with(name(.), 'f-')]
all French feminine can be returned with
/dic/word/@f-fr:*
all French feminine adjectives can be returned with
/dic/word/@f-fr:adjective
all translated English words return form
/dic/word/@en:*
all English nouns, whatever gender, can be obtained with something like
/dic/word/@*:nouns[contains(name(), 'en:')]

The reality is that this dictionary is used for translation and
the translated term is effectively returned by a key-based query that is 
more like
key('translator', $target, $dic)/@*[starts-with(name(), $lang)]

It must be good to know what is right from what is wrong,
especially with an absolute perspective.
I have to admit that I have always had some disbelief about absolute 
beliefs,
but I will keep an open mind, at least just in case.

Regards,
Andre



> -----BEGIN PGP SIGNED MESSAGE-----
> Hash: SHA1
>
> ac wrote:
>
> > Maybe you did not get the streaming content requirements which
> > precludes, at least in principle, doing dynamic loading of each
> > localization file.  Any unpredictable number of languages need to be
> > available, in a single transformation.  Constant re-loading will not
> > solve anything.
> Subsequent calls to document()/doc() function are returning previously
> loaded and parsed document tree so there is no speed/memory penalty
> attached to it. Your transformation will use as much memory as is needed
> for languages used during this transformation.
>
> > As for defining additional namespaces and changing the schema, what is
> > wrong with doing it once, at design time, for all languages that will
> > need to be supported?
> I have never taken part in a project where everything was known at the
> design time. Maybe be this is rare exception and you will be very lucky.
>
> > Simply wrong is easy to say, but it is probably a relative thing like
> > all others.  Could you better explain to me why using namespaces would
> > be "simply wrong", especially when they better solve the problem at stake?
> With pleasure. For example given your source:
>
> <word en:instance="Mr" en-f:instance="Mrs" fr:instance="M."
> fr-f:instance="Mme" ... />
>
> it would be very unnatural to write queries against such model. Imagine
> you need to retrieve for example:
>
> - -- all English words
> - -- all female variant of word
>
> In your example you are using namespaces to encode information which
> would be more natural to express as attribute/element values or their
> respective nesting.
>
> > You refer to my dictionary example, but not to my locator examples, with
> > GML, for example, or the others.  I get the feeling that the overall
> > picture could mean something too.
> Honestly, thread is too long to digest what you are referring too now.
> I'm just saying that way you are using namespaces for storing
> localization data is wrong. I'm not saying anything about other use of
> namespaces in your solution.
>
> 				Jirka
>
> - --
> - ------------------------------------------------------------------
>    Jirka Kosek      e-mail: jirka@xxxxxxxx      http://xmlguru.cz
> - ------------------------------------------------------------------
>         Professional XML consulting and training services
>    DocBook customization, custom XSLT/XSL-FO document processing
> - ------------------------------------------------------------------
>   OASIS DocBook TC member, W3C Invited Expert, ISO JTC1/SC34 member
> - ------------------------------------------------------------------
> -----BEGIN PGP SIGNATURE-----
> Version: GnuPG v1.4.9 (MingW32)
> Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org/
>
> iEYEARECAAYFAk2tqbgACgkQzwmSw7n0dR72cwCeJAXsTMm9xiKhxOYXDWvEX/3/
> TJcAoJN/DVjcZdEhM+mHIuViYqodFtXb
> =54ju
> -----END PGP SIGNATURE-----