Re: Yoruba, Hausa and Igbo (yo, ha, ig) po files
- From: Marcel Telka <marcel telka sk>
- To: Thomas Thurman <tthurman gnome org>
- Cc: legal-list gnome org, gnome-i18n gnome org, kde-i18n-doc kde org
- Subject: Re: Yoruba, Hausa and Igbo (yo, ha, ig) po files
- Date: Wed, 14 Jan 2009 13:29:25 +0100
Hi Thomas,
On Tue, Jan 13, 2009 at 10:57:14AM -0500, Thomas Thurman wrote:
> Ysgrifennodd Marcel Telka:
>> On Mon, Jan 12, 2009 at 09:43:51PM -0500, Thomas Thurman wrote:
>>> Maybe in future we need to make the gettext tools enforce a rule that
>>> the header block of a .po file contains licence information, instead
>>> of assuming it's okay to leave it in the comments.
>>
>> Then we should also enforce compiler (gcc) to add license information to
>> binaries...
>>
>> Do we really want that?
>
> Yes, I'd like that too. Actually, it would be really useful to warn
> people linking non-GPL binaries against GPL libraries.
>
> But I don't think the cases are parallel. You can't point some tool at
> a compiled binary and get the source code back. You can throw .mo files
> at msgunfmt and get *almost* the original .po back, but lacking the
You could do this with assembler source code too :-).
Lack of a tool to do it with C/C++/whatever is not a reason :-).
I think both cases ARE parallel.
> comments; if the comments are just comments, that's fine because they're
> just hints to humans, but if they contain actual useful metadata,
Are we considering copyright info as an "useful metadata"?
If I want I could add my own comments to a po file with another kind of
"useful" metadata. In general, all comments could be considered "useful
metadata".
> there's no reason that metadata shouldn't live in the header block along
> with the metadata we already carry, like contact email address and name
> of the last translator.
Good point. Do we need such data in the header? For what purposes? Do we
have something similar in C/C++ (compiled, sources)? The answer is no
:-).
>
> Honestly, the only metadata that *needs* to be in the header block is
> the content encoding; everything else is for human convenience (author
> name and contact details could live in comments as licence data already
> does, date can be figured out from source control, project can be
> figured out from the place the .po or .mo file is found, language team
> can be figured out from the pathname and the project name). There's no
> reason not to add another field to the mix to stop cases like this from
> happening again.
Exactly. There is no reason to keep such data in po file. What about get
rid off of that data?
To clarify: I am not voting against adding copyright data into the
header, nor voting for removing some data from there. I am just showing
that there could be also other possible view...
... and, I am voting for no change. :-)
Have a nice day.
--
+-------------------------------------------+
| Marcel Telka e-mail: marcel telka sk |
| homepage: http://telka.sk/ |
| jabber: marcel jabber sk |
+-------------------------------------------+
- References:
- Re: Yoruba, Hausa and Igbo (yo, ha, ig) po files
- Re: Yoruba, Hausa and Igbo (yo, ha, ig) po files
- Re: Yoruba, Hausa and Igbo (yo, ha, ig) po files
- Re: Yoruba, Hausa and Igbo (yo, ha, ig) po files
- Re: Yoruba, Hausa and Igbo (yo, ha, ig) po files
- Re: Yoruba, Hausa and Igbo (yo, ha, ig) po files
- Re: Yoruba, Hausa and Igbo (yo, ha, ig) po files
- Re: Yoruba, Hausa and Igbo (yo, ha, ig) po files
- Re: Yoruba, Hausa and Igbo (yo, ha, ig) po files
[
Date Prev][
Date Next] [
Thread Prev][
Thread Next]
[
Thread Index]
[
Date Index]
[
Author Index]