Il 28/09/2014 00:10, Bart ha scritto:
On 9/27/14, Giuliano Colla <giuliano.co...@fastwebnet.it> wrote:
If Bart, who is the author and rightful "owner" of Pochecker, doesn't
like the idea of making it too bloated, then creating a new tool becomes
an option to be considered.
First of all, this is a community driven project, so extending the
PoChecker tool to have editing/fiximg/sanitizing capabilties is an
option, if so desired by the community.
I just did not envision this sort of tasks for the tool at all when I
started out creating it.
I already stated why I don't think extending it this way is a good
idea earlier in this thread.
Please do not misunderstand me.
What I meant is that being you the author of the project you're also
rightfully the project leader.
It must be that way, because a leaderless project goes randomly nowhere.
That said, I think that it would be unwise attempting to twist your arm
to make the project take a different route, for two very good reasons.
Personal reasons: this is not a commercial activity, it's an activity
based on voluntary work. One must like what he does, be convinced that
it's going in the right direction, and enjoy doing it.
Technical reasons: being the author you're the one with the best
knowledge of how it was originally planned, which features can be easily
integrated in this frame, and which ones fall outside the planned frame.
The community may ask features, and contribute to some extent, but must
take very seriously your opinion.
About the scattering of tools/code for these tasks:
I created the SimplePoFiles, which basically is a rip-off of the
Translations unit (I cut out all parts I did not need) because:
[...]
Again don't misunderstand me.
I didn't mean Pochecker being scattered or whatever. Currently it's a po
checking tool and it does exactly what's it's intended to do.
What I meant is that the total of the tools required by a translator are
scattered, and very far from being integrated.
If we look how currently the translation process is carried on, we see
that it's very far from smooth.
1.
The IDE compares the po template and translations with resource
string, and updates all Msgid field where appropriate, marking them
as Fuzzy is changed. (btw. gettext and editing tools assumes a .po
template to have the .pot extension, and actual translations to have
.po extension, while in Lazarus both template and translations carry
.po extension. Why?)
2.
The translator locates the translations in need of editing. The tool
available is Pochecker
3.
The translator edits the translation by means of a po editing tool.
The tool available is poedit (or an equivalent), which doesn't
perform any check, but assists only in locating fuzzy and
unstranslated strings, and in adding some escape characters.
4.
If the translator wants to clean-up and sanitize the translation,
getting rid of useless remnants, edits again the file with a text
editor which provides none or minimal assistance for po file
specific format. Many tools available, from IDE editor to lot of others.
5.
The translator checks if the translation is good. The tool available
is Pochecker.
6. Back to point 2 to locate next translation.
Maxim may check this list, and see if I left out something.
They're many steps with different tools, many manual operations, when
computers are supposed to be devices providing automatic functions.
A single tool providing steps from 2 to 6 would be a great help to
translators, it would make the translation process much faster and less
error-prone. And most of the functions required are already available in
the Lazarus project units.
This result can be achieved either by evolving PoChecker into a
PoConditioner, or by creating a PoConditioner as a new tool.
[...]
Maybe extra the capabilities that SimplePoFiles has can be ported back
to the Traslations unit. This would cut down on maintemance.
This could be a preliminary step whatever is decided to do after. If you
need manpower for this, you may count on me. I also consider getting rid
of PChars a sort of moral duty :-)
Giuliano
--
Giuliano Colla
Project planning question: when it's 90% done, are we halfway or not yet?
--
_______________________________________________
Lazarus mailing list
Lazarus@lists.lazarus.freepascal.org
http://lists.lazarus.freepascal.org/mailman/listinfo/lazarus
