--===============5071833795971740188==
Content-Type: multipart/signed; boundary="nextPart8999684.pKguNnYRRv"; micalg="pgp-sha1"; protocol="application/pgp-signature"
Content-Transfer-Encoding: quoted-printable


--nextPart8999684.pKguNnYRRv
Content-Transfer-Encoding: quoted-printable
Content-Type: text/plain; charset="utf-8"

On Saturday 10 March 2012 12:01:17 Thomas L=C3=BCbking wrote:
> Am 10.03.2012, 07:29 Uhr, schrieb Martin Gr=C3=A4=C3=9Flin <mgraessli=
n@kde.org>:
> > I wanted to send out a different RFC soon: What to do with tiling?
cc-ed Nikhil
>=20
> I know that you want to do it using JS ;-)
more to that underneath
>=20
> Mind that there's iirc somewhere a bug or wish to allow dynamic ratio=

> changes (ie, resizing a client is permitted and causes all other clie=
nts
> to resize as well) - i doubt this would be in cards if tiling is
> implemented in JS.
>=20
> A major problem of tiling is probably the nearly non-existent integra=
tion
> with quick tiling/maximization (legacy maximization floats the window=
, but
> there's no reasonable way back, you got an entry in the alt+f3 menu) =
and
> tabbing (completely ignored)
>=20
> Assuming none of the readers of this list can make much sense out of
> tiling, we should also probably call for a brainstorm on tiling first=
, ie.
> get some discussion and input from ppl. who really use tiling (and
> possibly have used other tiling WMs, i tried dwm - broke my entire de=
sktop
> ;-) so we can figure what's actually important.
I think you summarized the issues with the current tiling implementatio=
n quite=20
nicely:
* nobody of the core development team uses tiling
* as a result nobody knows how tiling should behave
* current tiling implementation has some showstopper bugs to be really =
useful,=20
like missing multi screen support
* tiling is currently unmaintained

None of these issues I see to be resolved soon which makes me question =
the=20
complete implementation. If it is not able to work with multiple screen=
s and=20
does not feel integrated who would use it? Furthermore are the users wh=
o use=20
tiling at all in the targeted user group of KDE? I know quite a lot of =
people=20
using tiling window managers - none of them uses KDE. Most would be sui=
ted=20
best by an xterm multiplexer :-)

So in the current state I only see two possible long term solutions, ei=
ther=20
making it really good or drop it. Given problem number one I don't see =
how we=20
should make it good.

Now what has it to do with scripting? Well scripting could be a nice so=
lution=20
to keep the current state of a tech demo. It would be more clear that i=
t is=20
only an addition and that KWin does not want to be a tiling window mana=
ger (no=20
competition to awesome). Furthermore it would be the proof of concept f=
or the=20
JavaScript bindings. But still: who's going to implement it?

So I am very uncertain about the future of tiling and would like to get=
 some=20
feedback.

Cheers
Martin
--nextPart8999684.pKguNnYRRv
Content-Type: application/pgp-signature; name="signature.asc"
Content-Description: This is a digitally signed message part.

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.11 (GNU/Linux)

iEYEABECAAYFAk9dBskACgkQqVXwidMiVrryaQCfQ52p6/vNfhPPFVA8OghYEUMp
EvoAnAwuC6dTy1z2IG+x+WD57VBkF6Mj
=mgzx
-----END PGP SIGNATURE-----

--nextPart8999684.pKguNnYRRv--


--===============5071833795971740188==
Content-Type: text/plain; charset="us-ascii"
MIME-Version: 1.0
Content-Transfer-Encoding: 7bit
Content-Disposition: inline

_______________________________________________
kwin mailing list
kwin@kde.org
https://mail.kde.org/mailman/listinfo/kwin

--===============5071833795971740188==--