--===============2115028334==
Content-Type: multipart/signed;
  boundary="nextPart1194051.oFUHiRlDCi";
  protocol="application/pgp-signature";
  micalg=pgp-sha1
Content-Transfer-Encoding: 7bit

--nextPart1194051.oFUHiRlDCi
Content-Type: text/plain;
  charset="iso-8859-1"
Content-Transfer-Encoding: quoted-printable
Content-Disposition: inline

On Saturday 20 August 2005 02:25, Tomasz Grobelny wrote:
> On Saturday 20 August 2005 01:15, Ingo Kl=F6cker wrote:
> > On Friday 19 August 2005 13:02, Tomasz Grobelny wrote:
> > > I thought of adding language detection to KMail. It could be used
> > > when you reply to somebody's mail to put "On %d you wrote:"
> > > phrase in appropriate language. And that works fine on my
> > > desktop. Additionally it should change signature and dictionary.
> > > And maybe other things which we together call identity. In that
> > > case usability questions come to mind:
> > > 1. Why are phrases not part of identities?
> >
> > In no particular order:
> > - for historical reasons
> > - because identities are already overloaded enough with settings
> > - there's no clear connection between identities and reply phrases
>
> The connection is at least as clear as between identity and signature
> (can be formal or informal, in English or Polish in my case).
>
> > A possible solution (although it's probably not so good
> > usability-wise): - the user can add multiple sets of phrases, each
> > set can consist of phrases in multiple languages
>
> Why in one set there should be phrases in multiple languages? What
> will phrases in one set have in common?

As I wrote a bit further down one set would include formal phrases ("On=20
15/08/2005 David Brown wrote:") in several languages and another set=20
would include informal phrases ("On Monday, 15 August 2005 my beloved=20
friend known as David has written the following:") in several=20
languages.

This would make it much easier for people using multiple languages. They=20
would just have to create two identities, a Work identity and a Friends=20
identity.

If we tied an identity to a specific language then the same user will=20
have to create two identities per language, i.e., Work (English), Work=20
(German), Friends (English), Friends (German).

> > - the user associates an identity with one set of phrases
> > - KMail automagically chooses the phrases in the correct language
> > from the set of phrases associated to the current identity
> >
> > This solution is flexible because it allows multiple sets of
> > phrases, e.g. one set with formal phrases and one set with informal
> > phrases. OTOH, this solution is convenient because it automatically
> > chooses the correct language for the phrases.
> >
> > Adding phrases directly to the identities won't happen.
>
> Originally I thought about doing it the same way it is done with
> transport: if I choose identity A then use set of phrases B (where by
> set of phrases I mean 4 phases in one language: reply to author,
> reply all, etc.).

I know, but I won't accept such a solution.

> > > 2. How should information on language be put into identity?
> > > Note that language detection is just one example that I can
> > > theoretically implement in a few days, but from usability
> > > standpoint it should work well even if one day we will be able to
> > > tell if the e-mail we are replying is formal or informal, kind or
> > > rude, or whatever.
> >
> > See above. There's no reason to associate identities with just one
> > particular language. In fact, I'd like to get rid of the dictionary
> > setting for identities. Just like the language for the reply
> > phrases the correct dictionary should be chosen automagically.
> > There would just
>
> So the idea is to break up identity into two parts: language specific
> and language independent?

Well, identities should be language independent. The language specific=20
options (language of phrases, dictionary, etc.) should be chosen=20
automatically based on the detected language. Of course, it might still=20
make sense to specify a default language for an identity so that it=20
also works if language detection isn't possible for some reason.

Regards,
Ingo

--nextPart1194051.oFUHiRlDCi
Content-Type: application/pgp-signature

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.0 (GNU/Linux)

iD8DBQBDByl9GnR+RTDgudgRAtmBAKCwpmZx07URZGwgA72C75k7ZCICyACdGuVN
JzOf01mTx4H1Z3Sq/BoybO0=
=gJk/
-----END PGP SIGNATURE-----

--nextPart1194051.oFUHiRlDCi--

--===============2115028334==
Content-Type: text/plain; charset="iso-8859-1"
MIME-Version: 1.0
Content-Transfer-Encoding: quoted-printable
Content-Disposition: inline

_______________________________________________
KMail developers mailing list
KMail-devel@kde.org
https://mail.kde.org/mailman/listinfo/kmail-devel

--===============2115028334==--