[zope3-i18n] Report after first complete french translation

Philipp von Weitershausen philipp at weitershausen.de
Mon Oct 10 23:15:00 EDT 2005

Sebastien Douche wrote:
> Hi there :)
> I have finished first french translation of Zope 3.1. The work is
> buggy and I will enhance it with iterations. But, for now I report a
> couple of thoughts :
> # Why msgstr include many \n and <tab> ? It's ugly.

Yes, whitespace in messages is a bit ugly, but it can be safely ignored.  It is usually
there due to source code formatting (we limit ourselves to 80 character rows in Zope 3).

> # Must to define quote : ' ' or ` ` or `` `` but not all.
> # Sometimes it's hard to determine if the sentence is for developer or
> user (the context). Example : "If set to True ...". If it's for UI, ok
> you can translate "True" ("Vrai" in french) but if it's for developer
> (for explain a line of code), you must leave "True" because the code
> is only in english.

Well, Boolean values (true and false) are a concept. So, I would always translate it, even
if the message is for developers. They will understand when you write "Vrai". After all,
there's only one choice they can make...

> # It should be good to have a description of these technical terms :
> pagelet
> pagetet chooser
> registry
> registration
> generation (src/zope/app/generations)
> mapping (src/zope/app/workflow/stateful/browser)
> # List of msg I wish an explanation :
> 34 : Global Component Architecture
> 38 : Local Component Architecture
> "(global component) + architecture" or "global (componant architecture)" ?

global/local (compoent architecture)

> 252 : Realm

For words like this I advise to consult a technical dictionary. "Realm" is usually the

> 269 : Browser Form Challenger

This is a component that challenges users into providing login credentials by presenting
an HTML form to a browser.

> 518 : ${num} robot unit

Like Adam said, no real translation is needed here, it's just a test message. In fact, it
really should be in a different translation domain, though the extraction tool right now
doesn't know how to extract different domains from Python code correctly.

> 1161 : Timeout resolution (in seconds)
> 1162 : Defines what the 'resolution' of item timeout is. Setting this
> higher allows the transience machinery to do fewer 'writes' at the
> expense of  causing items to time out later than the 'Data object
> timeout value' by  a factor of (at most) this many seconds.

Where's your problem with those?

> 1224 : Checked out
> 1225 : Checked in
> 1226 : id of the effective user at the time the bookkeeping
> information was created
> 1230 : Checkout
> 1231 : Checkin
> 1232 : Uncheckout
> 1239 : Checkin Message

Like Adam said, "to check in" and "to check out" are terms borrowed from  source control
and versioning systems.

> # This is a list of msg (id is in Kbabel) that shoud be rewritten :
> 153 : Specific Required Interface Adapters
> 154 : Show specific required interface adapters
> 155 : Extended Required Interface Adapters
> 156 : Show extended required interface adapters
> 157 : Generic Required Interface Adapters
> 158 : Show generic required interface adapters
> What is specific/extended/generic ? What is required ? Maybe stuff like this :
> specific adapter with required interface
> 259 : A Pluggable Authentication Persistent Authentication Plugin

You are right, these names are unnecessarily confusing. I'll think of better ones and
discuss it with Stephan.

> 436 : Make the adapter a trusted adapter\n
> \n
>         Trusted adapters have unfettered access to the objects they\n
>         adapt.  If asked to adapt security-proxied objects, then,\n
>         rather than getting an unproxied adapter of security-proxied\n
>         objects, you get a security-proxied adapter of unproxied\n
>         objects.\n
> 445 : Make the subscriber a trusted subscriber\n
> \n
>         Trusted subscribers have unfettered access to the objects they\n
>         adapt.  If asked to adapt security-proxied objects, then,\n
>         rather than getting an unproxied subscriber of security-proxied\n
>         objects, you get a security-proxied subscriber of unproxied\n
>         objects.\n

Where's the problem with those?

> # List of msgid with typo error :
> 939 : The id uniquly identifies this menu.

This has been fixed in Zope 3.1.0-final.

> 949 : When a user selects a browser menu item, the URLgiven in the
> action is displayed. The action is usually given as a relative URL,
> relative to the object the menu item is for.
> 1221 : \n
>         Id of the version history related to the version controlled resource.\n
> \n
>         If this isn't set (is None), \n

I will fix this in the Zope 3.1 branch and the trunk.

>  #List of msgid that leave a error in kbabel :
> 85 : <subscription>
> 373 : <no name>
> 1107 : <permission not activated>
> 1142 : <role not activated>

Must be a problem with KBabel. I see nothing wrong with them.

> PS : I hate Launchpad.

What do you hate about it?

Thanks for your input


This message was sent using IMP, the Internet Messaging Program.

More information about the zope3-i18n mailing list