Techniques for multi-lingual GUI design

[TheDcoder]

Hello everyone!

It has been a very long time since I made a post in this forum

I am working on a freelance project which requires me to create a multi-lingual GUI, and I could think of a few ways to do that. But I want to know what you guys have in store for this, please share all of your awesome tips and hack for creating multi-lingual GUIs.

If this thread gets enough submissions, maybe we can make this into a meta-thread listing all of the techniques.

Thank you for the submissions in advance!

Regards,
TheDcoder.

[mikell]

On 7/11/2020 at 11:03 AM, TheDcoder said:

I could think of a few ways to do that

Wow. So much ?
Given the requirements in such a situation (i.e. mainly allow further additions) , IMHO there are not 50 ways to do that... maybe you will get slight differences in the flavour but the concept will remain the same 

[TheDcoder]

Maybe, I did not give it much thought

I am working on my own sloppy system right now, which looks like this:

Global Enum $L_DUMMY, _
$L_TITLE, _
$L_FOO, _
$L_BAR, _
$L_DUMMY2

Global Const $L_EN[5] = ["", _
    "Hello World", _
    "Foo", _
    "Bar", _
    "" _
]

Global Const $L_HI[5] = ["", _
    "नमस्कार दुनिआ", _
    "गु", _
    "गाली भरो मादर***", _
    "" _
]

Global $l = $L_HI

 

Edited July 13, 2020 by TheDcoder

[jchd]

Don't make your coding difficult!

expandcollapse popup

; supported languages
Global Enum $LANG_EN, $LANG_RU, $LANG_DE, $LANG_FR

; give strings a readable codename instead of an error-prone index
Global Enum $MSG_Close, $MSG_StdChkBox, $MSG_Quit, $MSG_ChkedBox, MSG_UnChkedBox

; array of string by language, string
Global $Dictionary = [ _
    [ _                         ; EN
        "Close", _
        "Standard Checkbox", _
        "Quit", _
        "The checkbox is checked.", _
        "The checkbox is not checked." _
    ], _
    [ _                         ; RU
        "близко", _
        "Стандартный флажок", _
        "Уволиться", _
        "Флажок отмечен.", _
        "Флажок не отмечен." _
    ], _
    [ _                         ; DE
        "Schließen", _
        "Standard-Kontrollkästchen", _
        "Verlassen", _
        "Das Kontrollkästchen ist aktiviert.", _
        "Das Kontrollkästchen ist nicht aktiviert." _
    ], _
    [ _                         ; FR
        "Fermer", _
        "Case à cocher standard", _
        "Quitter", _
        "La case est cochée.", _
        "La case n'est pas cochée." _
    ] _
]

; language in force
Global $Lang

;===========================================================================
; example use
;---------------------------------------------------------------------------
; randomly select a language and display some strings
For $i = 1 To 3
    $Lang = Random($LANG_EN, $LANG_FR, 1)
    Use_Example()
Next

Func Use_Example()
    For $i = 1 To 3
        MsgBox(0, "", $Dictionary[$Lang][Random(0, UBound($Dictionary, 2) - 1, 1)])
    Next
EndFunc

 

[water]

Before coding I think you need to make some decisions:

• Do you just want to support left-to-right languages or right-to-left languages as well?
• How do you handle languages that need more space in a GUI than others (e.g. "food intolerance" needs more space in German: "Nahrungsmittelunverträglichkeit").
  Do you want to use smaller fonts, line breaks ... to not have this literals truncated?
• How do you handle untranslated literals? What's the default language to use in this case?
• ...

Many years ago (~ 2008) I have written an UDF to allow multi language error messages.
it reads language dependant literals from a file, replaces placeholders with variable data and returns the string to the script.
Unfortunately it doesn't handle any of the mentioned problems.

[jchd]

3 minutes ago, water said:

Do you just want to support left-to-right languages or right-to-left languages as well?

Not an issue at all, thanks to Unicode!

4 minutes ago, water said:

How do you handle languages that need more space in a GUI than others (e.g. "food intolerance" needs more space in German: "Nahrungsmittelunverträglichkeit").

Can be handled with the UDF @Melba created for this (StringSize or something).  It's also relatively easy to create a crude and simple string splitter, even if very far from LaTeX.

More difficult is to design GUI elements elegantly fitted to a variey of languages.

[water]

44 minutes ago, jchd said:

Do you just want to support left-to-right languages or right-to-left languages as well?

I was thinking of the RTL extended styles like $WS_EX_LAYOUTRTL, $TVS_RTLREADING etc. But I have never used them myself.

[argumentum]

...why not an external file ( .ini or .lng ) where the user can provide a translation, saving you from guessing a language you are not very familiar with. ( my 2 cents )

[TheDcoder]

@jchd Thanks for your alternate method of using a 2D array instead of an 1D approach like myself. Some people might prefer a 1D approach as it makes the expression shorter ($l[$L_EN][$L_FOO] vs $l[$L_FOO])

@water You have raised some very good points, especially the point about GUI layout.

@argumentum Absolutely doable, that is what I originally wanted to do, but decided against that to make things simpler (in the short term). There is an excellent UDF called AU3Text by @dany which does exactly this, and much more. It also contains a version of Ini functions which are capable of properly handling unicode input I think. Sadly the demo gave me a bunch of syntax error when I attempted to run it in the latest AutoIt

 

[jchd]

An external source is open to misuse but maybe a good solution. For instance Notepad++ uses .xml files for its localization, implements in-band escapes to insert variables inside strings and even allows multiline text, e.g.:

<find-in-files-filter-tip value="Find in cpp, cxx, h, hxx &amp;&amp; hpp:
*.cpp *.cxx *.h *.hxx *.hpp

Find in all files except exe, obj &amp;&amp; log:
*.* !*.exe !*.obj !*.log"/>
...
            <find-status-replaceall-1-replaced value="Replace All: 1 occurrence was replaced"/>
            <find-status-replaceall-nb-replaced value="Replace All: $INT_REPLACE$ occurrences were replaced"/>
...
            <find-status-replaceinfiles-1-replaced value="Replace in Files: 1 occurrence was replaced"/>
            <find-status-replaceinfiles-nb-replaced value="Replace in Files: $INT_REPLACE$ occurrences were replaced"/>

Of course $INT_REPLACE$ would ideally take care of the script (= written language) used for a given translation, as latin and many others say 123456 when Devanagari would use १२३४५६

[TheDcoder]

2 hours ago, jchd said:

An external source is open to misuse

Indeed, no way to get around that. But really, if we are concerned about security there isn't much that can be done as the binary itself can be modified to change the strings.

Another advantage of external sources is that it makes it possible to just download the languages which are needed, as opposed to all supported languages, which would essentially be bloat.

[mikell]

On 7/14/2020 at 2:24 AM, argumentum said:

..why not an external file

It's obviously the most handy and versatile way. Most multilingual softwares have English language hardcoded and provide external language files. This way allows easy corrections, further language additions etc

[Nine]

My personal preference is to go with a database.  It allows me to quickly get all labels, messages, errors, lists, etc.  What I usually do is to assign an ID to each pages of a GUI so I query in an array all the texts with a single call.

It is also extremely easy to add or remove a language.  I never hardcode a language, because if the client wants to modify a certain term, you do not need to provide a new set of exe, he can make his own modification without involving you.

Edited July 15, 2020 by Nine

[TheDcoder]

@Nine Sounds interesting, by "database" I assume you mean an SQLite database?

[Nine]

@TheDcoder Not necessarily SQLite.  Any DB would fit.  Oracle was my first love, but of course SQLite is now a good choice as it is free, uses low resources and performs very well.

[Earthshine]

20 or so years ago I  used an ini file With all the different languages for all the text boxes and controls and anything that had text on the form and message box text so when you load the form you have to iterate through all the controls on the form and look up what their text should be

Edited July 16, 2020 by Earthshine

[MrCreatoR]

Check out my ATT - Application Translate Tool.

[TheDcoder]

@Nine I see, a fully-fledged DB seems to be an over-kill, but it is an option nonetheless

@MrCreatoR Nice, I came across it once but I wasn't able to find it after that.

[TheDcoder]

On 7/14/2020 at 2:27 PM, TheDcoder said:

There is an excellent UDF called AU3Text by @dany (...) Sadly the demo gave me a bunch of syntax error when I attempted to run it in the latest AutoIt

I have posted a fixed version in that thread, I am going to try it with a new project, wish me luck guys!

[TheDcoder]

Due to several reasons, AU3Text was just not working out for me... so instead, I decided to create my own version:

It uses a forked version of EasyCodeIt to extract the translation strings and internally uses Scripting.Dictionary to keep track of the strings.