Smalltalk › Squeak › Squeak - Dev

still looking for marking string for translation

_‹ Previous Topic Next Topic _›

Classic

List

Threaded

11 messages Options

Hilaire Fernandes-5

still looking for marking string for translation

Hello people,

I am still looking for a way to mark, in image 3.8, string for translation?

Does the procedure described at http://minnow.cc.gatech.edu/squeak/402
became obselete?

At worst I will have to use directly local language string.

Hilaire Fernandes

Yoshiki Ohshima

Re: still looking for marking string for translation

Hilaire,

> I am still looking for a way to mark, in image 3.8, string for translation?
>
> Does the procedure described at http://minnow.cc.gatech.edu/squeak/402
> became obselete?
>
> At worst I will have to use directly local language string.

I'm not exactly sure by "marking", but the closest thing would be:

NaturalLanguageTranslator registeredPhraseFor: phrase.

and then use the LanguageEditor.

-- Yoshiki

Hilaire Fernandes-5

Re: still looking for marking string for translation

Yoshiki Ohshima a écrit :

> Hilaire,
>
>
>>I am still looking for a way to mark, in image 3.8, string for translation?
>>
>>Does the procedure described at http://minnow.cc.gatech.edu/squeak/402
>>became obselete?
>>
>>At worst I will have to use directly local language string.
>
>
> I'm not exactly sure by "marking", but the closest thing would be:

The only localisation system I know is gettext, in this system the
string to translate are marked directly in the source code (with a
special function as _()). Next, such marked strings can be grouped in
the translation file for the translator.

I was thinking about something to get the same effect (i.e. get the
string ready to be translated)

>
> NaturalLanguageTranslator registeredPhraseFor: phrase.
>
> and then use the LanguageEditor.

Ok.

Will it be possible to group strings in one independant package (I don't
know in which form), so it can be moved allong the application classes.
Otherwise I am afraid the marked string for translation could be very
esily loose as I guess they may be dependant on the image.

Hilaire Fernandes

>
> -- Yoshiki
>
>

Bert Freudenberg-3

Re: still looking for marking string for translation

Am 19.02.2006 um 20:50 schrieb Hilaire Fernandes:

>>> http://minnow.cc.gatech.edu/squeak/402

> The only localisation system I know is gettext, in this system the
> string to translate are marked directly in the source code (with a
> special function as _()). Next, such marked strings can be grouped in
> the translation file for the translator.
>
> I was thinking about something to get the same effect (i.e. get the
> string ready to be translated)

Just send #translated, as mentioned on the wiki page above.

- Bert -

Hilaire Fernandes-5

Re: still looking for marking string for translation

Bert Freudenberg a écrit :

> Am 19.02.2006 um 20:50 schrieb Hilaire Fernandes:
>
>>>> http://minnow.cc.gatech.edu/squeak/402
>
>
>> The only localisation system I know is gettext, in this system the
>> string to translate are marked directly in the source code (with a
>> special function as _()). Next, such marked strings can be grouped in
>> the translation file for the translator.
>>
>> I was thinking about something to get the same effect (i.e. get the
>> string ready to be translated)
>
>
> Just send #translated, as mentioned on the wiki page above.

Yes, this I already know but it is not sufficient. As Yoshiki suggested
I also need to add each string in the translation dictionnary.

What would be helpfull is to have a separate translation dictionnary for
each package.

As far as I can understand, there is for now one big dictionnary to
record the string for translation (NaturalLanguageTranslator
allKnownPhrases).

Hilaire

>
> - Bert -
>
>

--
ADD R0,R1,R2,LSL #2

stéphane ducasse-2

Re: still looking for marking string for translation

hilaire

you are right. A dictionary per package is needed.
This is also an indication that having first class package would help
to attach to them resources.
(argh I do not have the time to look at that). I will try to dig into
what alex did to introduce
first class packages and see how far he was.

Stef

On 20 févr. 06, at 09:52, Hilaire Fernandes wrote:

> Bert Freudenberg a écrit :
>> Am 19.02.2006 um 20:50 schrieb Hilaire Fernandes:
>>>>> http://minnow.cc.gatech.edu/squeak/402
>>> The only localisation system I know is gettext, in this system the
>>> string to translate are marked directly in the source code (with a
>>> special function as _()). Next, such marked strings can be
>>> grouped in
>>> the translation file for the translator.
>>>
>>> I was thinking about something to get the same effect (i.e. get the
>>> string ready to be translated)
>> Just send #translated, as mentioned on the wiki page above.
>
> Yes, this I already know but it is not sufficient. As Yoshiki
> suggested I also need to add each string in the translation
> dictionnary.
>
> What would be helpfull is to have a separate translation
> dictionnary for each package.
>
> As far as I can understand, there is for now one big dictionnary to
> record the string for translation (NaturalLanguageTranslator
> allKnownPhrases).
>
> Hilaire
>
>
>> - Bert -
>
>
> --
> ADD R0,R1,R2,LSL #2
>

Hilaire Fernandes-5

Re: still looking for marking string for translation

What is first class package?

Hilaire

stéphane ducasse a écrit :

> hilaire
>
> you are right. A dictionary per package is needed.
> This is also an indication that having first class package would help
> to attach to them resources.
> (argh I do not have the time to look at that). I will try to dig into
> what alex did to introduce
> first class packages and see how far he was.
>
> Stef
>
>
> On 20 févr. 06, at 09:52, Hilaire Fernandes wrote:
>
>> Bert Freudenberg a écrit :
>>
>>> Am 19.02.2006 um 20:50 schrieb Hilaire Fernandes:
>>>
>>>>>> http://minnow.cc.gatech.edu/squeak/402
>>>>
>>>> The only localisation system I know is gettext, in this system the
>>>> string to translate are marked directly in the source code (with a
>>>> special function as _()). Next, such marked strings can be grouped in
>>>> the translation file for the translator.
>>>>
>>>> I was thinking about something to get the same effect (i.e. get the
>>>> string ready to be translated)
>>>
>>> Just send #translated, as mentioned on the wiki page above.
>>
>>
>> Yes, this I already know but it is not sufficient. As Yoshiki
>> suggested I also need to add each string in the translation dictionnary.
>>
>> What would be helpfull is to have a separate translation dictionnary
>> for each package.
>>
>> As far as I can understand, there is for now one big dictionnary to
>> record the string for translation (NaturalLanguageTranslator
>> allKnownPhrases).
>>
>> Hilaire
>>
>>
>>> - Bert -
>>
>>
>>
>> --
>> ADD R0,R1,R2,LSL #2
>>
>
>

--
ADD R0,R1,R2,LSL #2

stéphane ducasse-2

Re: still looking for marking string for translation

If you look how package are right now implemented they are a facade
on the class category (the way
Smalltalk sort classes). But they are not real object.

For example, we cannot have package level comment to attach description.
I do not know how postload actions are specified but this would make
sense to me to
edit a package method and get sure that once all the classes are
loaded the package method is executed.

Having package as real entity would help. We could put resources
there too. RIght now you can do the same
but in a none organized manner everybody can decide to have one class
in the package having an initialize
method or having class methods to handle resources but this is not
consistent.

Stef

On 20 févr. 06, at 17:38, Hilaire Fernandes wrote:

> What is first class package?
>
> Hilaire
>
>
> stéphane ducasse a écrit :
>> hilaire
>> you are right. A dictionary per package is needed.
>> This is also an indication that having first class package would
>> help to attach to them resources.
>> (argh I do not have the time to look at that). I will try to dig
>> into what alex did to introduce
>> first class packages and see how far he was.
>> Stef
>> On 20 févr. 06, at 09:52, Hilaire Fernandes wrote:
>>> Bert Freudenberg a écrit :
>>>
>>>> Am 19.02.2006 um 20:50 schrieb Hilaire Fernandes:
>>>>
>>>>>>> http://minnow.cc.gatech.edu/squeak/402
>>>>>
>>>>> The only localisation system I know is gettext, in this system the
>>>>> string to translate are marked directly in the source code (with a
>>>>> special function as _()). Next, such marked strings can be
>>>>> grouped in
>>>>> the translation file for the translator.
>>>>>
>>>>> I was thinking about something to get the same effect (i.e. get
>>>>> the
>>>>> string ready to be translated)
>>>>
>>>> Just send #translated, as mentioned on the wiki page above.
>>>
>>>
>>> Yes, this I already know but it is not sufficient. As Yoshiki
>>> suggested I also need to add each string in the translation
>>> dictionnary.
>>>
>>> What would be helpfull is to have a separate translation
>>> dictionnary for each package.
>>>
>>> As far as I can understand, there is for now one big dictionnary
>>> to record the string for translation (NaturalLanguageTranslator
>>> allKnownPhrases).
>>>
>>> Hilaire
>>>
>>>
>>>> - Bert -
>>>
>>>
>>>
>>> --
>>> ADD R0,R1,R2,LSL #2
>>>
>
>
> --
> ADD R0,R1,R2,LSL #2
>

Yoshiki Ohshima

Re: still looking for marking string for translation

In reply to this post by Hilaire Fernandes-5

Hilaire,

There is an extension of LanguageEditor that lets you save the
selected entries in the LanguageEditor into a do-it expression. If
you copy the do-it to a postscript of a change set or such, you can
make a single package with the code and translations.

(Again, the extension is only in Squeakland... Sorry.)

-- Yoshiki

Hilaire Fernandes-5

Re: still looking for marking string for translation

I am afraid I don't understand the feature you described for the
LanguageEditor.

So now, I start to realize that Squeak != Squeakland, is it that?
At least, is there any common code repository or not?

Also if Squeak != Squeakland, I think I understand more or less how is
developped Squeak, but what about Squeakland.
After all, most email exchange in this list is related to Squeak, and in
the Squeakland list it is more end-user, but I think I never see email
exchange about how Squeakland is developped.
Looks like there are many things I don't understand in development model
of Squeak and Squeakland.
Worst I am confused, I am developping a dynamic geometry framework. To
do so I just use a basic Squeak 3.8 image, but I also want it to be
useable in Squeakland.

Hilaire Fernandes

Yoshiki Ohshima a écrit :

> Hilaire,
>
> There is an extension of LanguageEditor that lets you save the
> selected entries in the LanguageEditor into a do-it expression. If
> you copy the do-it to a postscript of a change set or such, you can
> make a single package with the code and translations.
>
> (Again, the extension is only in Squeakland... Sorry.)
>
> -- Yoshiki
>
>

--
ADD R0,R1,R2,LSL #2

Giovanni Corriga

Re: still looking for marking string for translation

In reply to this post by stéphane ducasse-2

Il giorno lun, 20/02/2006 alle 10.18 +0100, stéphane ducasse ha scritto:
> hilaire
>
> you are right. A dictionary per package is needed.
> This is also an indication that having first class package would help
> to attach to them resources.
> (argh I do not have the time to look at that). I will try to dig into
> what alex did to introduce
> first class packages and see how far he was.
>

An Application abstraction would be good, too.

Giovanni