[haiku-development] Re: Fwd: [haiku-doc] Organizing translations of User Guide/Welcome Package
- From: "Jonas Sundström" <jonas@xxxxxxxxxxx>
- To: haiku-development@xxxxxxxxxxxxx
- Date: Sat, 03 Jan 2009 16:37:55 +0100 CET
"Niels Reedijk" <niels.reedijk@xxxxxxxxx> wrote:
...
> Now the intermediate format I want to propose
> is the gettext format[6]. While gettext is geared
> towards app translations, its tools are very good
> for what we want. The source template is a POT file,
> we generate this file from the source docbook.
I'm possibly not seeing the entire picture here, but
I think that doing piecemeal translation of sentences
(or terms for that matter!) and then stringing that
together would result in a language soup that no one
should be exposed to.
Lack of context is a problem. Blind translation
seldom results in anything good. (Swedish Ubuntu.
Fansubbing. Ehrm.)
Not all languages have the same structure, order,
and flow. The smallest unit of translation should be
one paragraph, IMO, so that there can be a natural
flow from sentence to sentence.
(Ideally a document should have the same
author/translator, so that the same tone/pulse/
uninterrupted storytelling permeates the work.)
I suggest looking for a content management system
which could understand parallel documents
(one master and a set of slaves, translations)
each with -versioned- paragraphs which should ideally be
in sync between all set of documents. There should be a
way to find and edit unsynced paragraphs.
A GUI something like this:
(view with a monospaced font)
**************************************************
English Swedish Spanish (...)
document document document
**************************************************
some
previous ... ... ...
paragraphs
**************************************************
A paragraph En paragraf Un párrafo ...
in English. på svenska. en español.
... ... ...
(version 8) (version 5) (version 7)
**************************************************
some
following ... ... ...
paragraphs
**************************************************
(Post-translation, these three paragraphs should have
the same version number: 8 (or 8+ something)
If the source format was easy enough to parse it
should not be overly hard to make an editor like this.
I admit to having oversimplifed things by not seeing
the texts as more than a series of paragraphs. The
other document parts (headers, links, etc) would have
to be represented too.
Illustrations would have to be managed somehow.
Maybe as image thumbnails. Double-click to edit.
/Jonas.
Other related posts:
