[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: RRD graphs of l10n stats over time moved to i18n.debian.org


Dňa Sun, 12 Aug 2012 10:18:22 +0200 Christian PERRIER <bubulle@debian.org>

> As the last bit of move to i18n/l10n services to a new machine, I
> completed the move of RRD graphs of translation completion over time.

thanks for this ;-)

> These graphs, originally prepared by Nicolas François, haven't
> received a big publicity up to now and maybe are ignored by many
> people.

there was the the bytes statistics too
(http://i18n.debian.net/debian-l10n-stats/web-rrd/year/byte-*.png) i am
not able to find them now. Please, there are not generated now?

> Please notive that these graphs are mostly made of quite "raw" data
> and are not nice and fancy web pages with tons of statistics and
> colors. Indeed, they are meant to build such pages but the work is
> still to be done.
> The top URL is http://i18n.debian.org/l10n-rrd
> There is a lot that can probably be done in the future with this data.

One idea: is problem to generate the images with number values, please?

One another question, about Compendium...

First, in the http://i18n.debian.org/compendia/ is recursive link to
compendia dir, i am not sure, it is needed?

Next, what is purpose of this compendium (i know what generally
the gettext's compendium is for). I ask, because (exactly compendium-sk)
contains a lot of fuzzy messages, which are fuzzy because the string have
different translation in sources.

If the purpose of the compendium is point to the differences in
translation, then the purpose is fulfilled. But if it is provided with
purpose to help with new (or updated) translation, it is mostly unusable,
by these fuzzy messages.

My opinion is to provide the compendium generated with "--use-first" flag,
which will drop other variants of the translation and provides compendium
usable to use for translation. And provide another, the differential
compendium, generated with "--more-than 2" and pipe it to the "msgattrib
--only-fuzzy --no-obsolete", or something similar.

What is your opinion, please?



Attachment: signature.asc
Description: PGP signature

Reply to: