Localization of TurtleArt

[Walter Bender]

As I recall (it was a while ago) ImageMagic had lots of issues with
non-Latin scripts. But it would be an easier route than the Gimp.

-walter

On Fri, Feb 29, 2008 at 8:52 AM, Alexander Todorov
<alexx.todorov at gmail.com> wrote:
> -----BEGIN PGP SIGNED MESSAGE-----
>  Hash: SHA512
>
>
> Walter Bender wrote:
>  > I'm a bit rusty, but you can use the Gimp to do this, using Scheme
>  > scripts. I did have a bit of trouble with positioning on some RTL
>  > scripts as the Gimp is using fairly antiquated text rendering
>  > internally, but it is generally OK.
>  >
>
>  I will give GIMP a try although I was thinking of ImageMagic and the like.
>
>
>
>  > -walter
>  >
>  > On Fri, Feb 29, 2008 at 6:08 AM, Alexander Todorov
>  > <alexx.todorov at gmail.com> wrote:
>  >> -----BEGIN PGP SIGNED MESSAGE-----
>  >>  Hash: SHA512
>  >>
>  >>  Hello everybody,
>  >>  I've spent some time thinking about this ticket:
>  >>  http://dev.laptop.org/ticket/3585
>  >>
>  >>  TurtleArt developer(s) have decided to use gif images that represent different
>  >>  shapes with text inside. What we need is a helper script that will take the
>  >>  translations from the PO file and apply them on no-text image to produce the
>  >>  localized image.
>  >>
>  >>  <pseudo_code>
>  >>
>  >>  helper(base image, text from PO) = localized image
>  >>
>  >>  </pseudo_code>
>  >>
>  >>  The idea is simple but I still hit some issues. Any help or ideas will be
>  >>  appreciated.
>  >>
>  >>  1) Most of the images have a single word or couple of words next to each other.
>  >>  Some of them like if-then or if-then-else have text which is not on a single
>  >>  line. e.g.
>  >>
>  >>  if
>  >>   ----- then
>  >>
>  >>
>  >>  if
>  >>   ----- then
>  >>
>  >>   ----- else
>  >>
>  >>  I'm not sure how we can automate that easily without specifying coordinates
>  >>  where the text should appear. A possible solution is to divide all these strings
>  >>  into different layers and make the helper/PO aware of them.
>  >>
>  Looks like the images already contain the text in separate layers so it will be
>  easy.
>
>
>
>  >>  2) There are some "block" images that show all available blocks for the chosen
>  >>  category. I guess they fall in 1) if we talk about automating their localization.
>  >>
>  >>  3) The source files are Photoshop PSD ones. Is that an open format? I don't
>  >>  really know but still haven't found a tool that can work with them properly
>  >>  except GIMP. And isn't against OLPC vision to use commercial software to produce
>  >>   an OLPC activity? I'm willing to implement the helper mentioned above but I'd
>  >>  prefer some graphics format that I can manipulate easily in code. We can also
>  >>  upload the base images to git.
>  >>
>  >>  4) What will happen with all the localized images? How they will be distributed.
>  >>  We certainly don't want all other languages hanging around and occupying disk
>  >>  space when they are not necessary. At present the English ones are 728KB.
>  >>  Multiply that by 10/20 languages and we're talking about MBs here.
>  >>
>  >>  Thanks,
>  >>  Alexander.
>  >>  -----BEGIN PGP SIGNATURE-----
>  >>  Version: GnuPG v1.4.5 (GNU/Linux)
>  >>  Comment: Using GnuPG with Red Hat - http://enigmail.mozdev.org
>  >>
>  >>  iD8DBQFHx+eyhmd3WOiFct4RCgmGAJ4iWC/clQZBTyPezgsqMkPk4Lc5swCguJLz
>  >>  SryyS4/fBw1wt4U1xI3ObtY=
>  >>  =2vdk
>  >>  -----END PGP SIGNATURE-----
>  >>  _______________________________________________
>  >>  Devel mailing list
>  >>  Devel at lists.laptop.org
>  >>  http://lists.laptop.org/listinfo/devel
>  >>
>  >
>  >
>  >
>
>  -----BEGIN PGP SIGNATURE-----
>  Version: GnuPG v1.4.5 (GNU/Linux)
>  Comment: Using GnuPG with Red Hat - http://enigmail.mozdev.org
>
>  iD8DBQFHyA4Ehmd3WOiFct4RCpgGAJ9v5w02msE/74Av1B7/IWouv7bgjwCgp8K0
>  v4txJxT9AQpO4YLuSO2vZxo=
>  =Besv
>  -----END PGP SIGNATURE-----
>



-- 
Walter Bender
One Laptop per Child
http://laptop.org