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

Re: Bug#682543: Package description issues

Martin Eberhard Schauer wrote:
>    Description: Gettext for Ruby
>     Ruby GetText Package is Native Language Support Library and Tools
>     which modeled after GNU gettext package.
>     .
> In the first paragraph there seems to be quite a lot of unnessary
> capitalisation as well as missing articles und verbs. Actually article and
> verb appear at rubydoc.info (4):

>       Ruby-GetText-Package is a Localization(L10n) library and tool which is
>       modeled after the GNU gettext package.

Rubyforge has this Ruby package filed under the name "gettext", and
Debian has the (source and binary) package as "ruby-gettext", so it's
a bit irritating that the upstream maintainer chooses to label it
"Ruby-GetText-Package", especially with the BiCaps.  We could avoid
the name and just say "This Ruby package..."

> Description: gettext for Ruby
>  Ruby-GetText-Package is native language support (NLS) library and tools
>  which are modeled after the GNU gettext package.

I'd prefer to start from the version that talks about
"Localization(L10n)" - "Native Language Support" is usually a sign of
dated docs.  Of course that version still has unnecessary caps on
"Localization" and "L10n", and since the abbreviation isn't used again
there's no real point introducing it in the first place.

I'd also suggest ending with "modeled after GNU gettext"; it's a clone
of the program, not the GNU packaging style.

>  .
>  Features:
>    * simple APIs (similar to GNU gettext)
>    * rgettext creates po-files from Ruby scripts.
>      The po-files are compatible with GNU gettext.
>    * rmsgfmt creates a mo-file from a po-file.
Surplus indent throughout.

Lists are easier to read if they're parallel; this is one noun phrase
and then two sentences.  In fact, the first item boils down to "the
library API is just like the real gettext" (which has already been
mentioned).  Maybe we should promote that out of the list and have one
bulletpoint per executable in the toolset. 

> CCing l10n-english as I don't know what is right: po-file, po file, PO file.

Or maybe even ".po file" or "GNU gettext Portable Object file".  The
GNU gettext docs say "PO file", so I'll go for that.

So how about:

 Description: gettext for Ruby
  This Ruby package is a localization library and toolset modeled after
  GNU gettext, and with a similar simple API.
  It provides:
   * rgettext - creates gettext-compatible PO files from Ruby scripts;
   * rmsgfmt - creates a MO file from a PO file;
   * rmsgmerge - extracts translatable strings from Ruby scripts.

JBR	with qualifications in linguistics, experience as a Debian
	sysadmin, and probably no clue about this particular package

Reply to: