• L10N
• Romanian
• CommaTransition

Abstract

The correct diacritics for Romanian include (among others) the unicode codepoints 0x0218 (ț), 0x0219 (Ț), 0x021A (ș) and 0x021B (Ș) (t/T comma and s/S comma characters). Currently people are using some incorrect replacements t/T cedilla (ţŢ) and s/S cedilla (şŞ) instead.

Since the usage of the correct diacritics has been legislated in Romania and since we strive for correct Romanian we intend to switch the Debian distribution to use the correct diacritics. We are starting with the Debian Installer which is a good test bed for this since is an isolated environment (has no impact on interoperability and is entirely a Debian project).

This page focuses on the transition to use only correct t/T/s/S comma (țȚșȘ) characters in Romanian translations in the Debian distribution.

Status

Correct diacritics for Romanian include the problematic t/T comma and s/S comma characters (unicode points 0x0218, 0x0219, 0x021A, 0x021B).

Status:

[2007-08-10]

• a patch for the layout issue in gnome was submitted in the Debian BTS; a lengthy discussion about the default layout for Romanian installations happened in the sister ubuntu bug, but at this moment is not clear.

[2007-04-29]

• All the packages part of the Debian installer proper were converted to use the correct diacritics

[2006-11-29]

• postinstall issues:
  • it seems that there are some problems in the GTK widgets which prevent comma input; this affects GNOME installations
  • the GNOME desktop is not able to display all the variants of layout; as a consequence, both layout variants which use correct diacritics are not selectable via the GNOME keyboard applet

[2006-10-03]

• an experimental image with the correct diacritics was built by EddyPetrisor

  • results look promising for the graphical installer (see screenshots below)
  • the console small t comma looks ugly because the comma is too much to the left of the t - it looks more like a comma close but to the left of the t, instead of it being underneath it. The responsible font seems to be /usr/src/unifont.bdf in the package bf-utf-source
  • the mix of t/T/s/S cedilla and t/T/s/S comma comes from the fact that in most cases the messages containing sedilla charcter come from official packages, not from custom ones.
  • there are messages that come from packages which are not part of Debian Installer; these are more problematic given the fact that the upload rate of those is much more reduced than for Debian Installer packages, so it could be possible that they will not reach into Etch in time for the release. This is a particular concern for packages like apt, aptitude and dpkg.

Future plans

The transition will not be done before the release of Etch. Main reason for this is the big number of packages and current translations that still use the incorrect codepoints.

The next release, Lenny, will be using correct diacritics, while incorrect ones will be tolerated, although correct ones are preferred.

Screenshots

Notes:

• the missing lines are a problem of GTK/DirectFB and does not concern the transition

• there are some typos in the screenshots, they should all be fixed, but it doesn't hurt to report them on debian-l10n-romanian @ lists.debian.org