ubuntu-phone team mailing list archive

[Rodney Dawes <rodney.dawes@xxxxxxxxxxxxx>]

On Tue, 2014-04-15 at 17:52 -0400, Chris Wayne wrote:


> 1) Do we have a plan for the legacy deb-packaged apps we have on the
> image (or do we just plan to get those as clicks ASAP?)

I'm not sure what the packaging plan is there, but as far as the click
scope is concerned, the .desktop is the only place we have to get
information about the apps which are packaged as debs (such as
webbrowser-app). We're currently pulling the title, description, and
screenshot URL, from the .desktop file. The scope currently doesn't pull
any translations for it, but it's not terribly hard for it to start
doing so. There are some other things that are actual regressions from
the old scopes API version of the scope, which we're trying to get fixed
first.
  
> 2) Would inline translations have any effect on the image size
> compared to having the translations in a .mo? (More specifically for
> the description bits as the question re: Names was already answered in
> the other thread)

Virtually no. As plain text, the .desktop file will compress extremely
well, unlike the .mo files. There may be a slight increase of a few
bytes, for a few apps, but I don't see it causing any issue with image
size.

> 3) Would this method use Launchpad's translations service, or would it
> need to be separate?

That is dependent on the app authors. If apps are hosted on Launchpad,
then I'm sure they might use the LP translations service. For the apps
we upstream, we will continue using the LP translations service, yes. I
see no need to not do that.

> 4) Do we anticipate saving all apps' metadata translations on disk, or
> just those that are pre-installed?

Only those that are installed. Whether they are pre-installed or not
doesn't matter. Apps that are in the store, but haven't been installed,
won't have any local data on the device (except for minor caching of
network responses, which will be a per-user store, and not stored in /
anywhere). Likewise, if a device has multiple users, we will need to
support the possibility that those users might use different languages,
and so will need to be able to translate to those languages when
offline, when such translations have been provided by the app authors.

However, while we want great translations for core apps, I think it's
also safe to expect that the vast majority of apps that end up in the
store in the future, are likely to have little to no translation
support; particularly for "changes" info added to descriptions when new
versions of an app are uploaded. Many developers don't have the skills
or time to do full and proper translation support for their apps. They
just want to make an app and sell it.