Re: bytecode modules

--B0nZA57HJSoPbsHY
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
Content-Transfer-Encoding: quoted-printable

On Mon, Apr 04, 2005 at 09:58:55AM +0200, [email protected] wrote:
> * Bytecode or script for L&F modules.
>   - Performance.
>   - Requires changes to the server and lots of work.
>   + No network latency.
>   + Security, no theme plugins with spy-ware or other crap.
>  =20
> Here I'm arguing that designing a bytecode for theme/wm/widgets modules i=
s a
> good idea.

Certainly it's possible and has at least some advantages. However,
it's not necessary, and I can't think of a good reason to go to that
much effort.

As a statement of general principle, I can see no reason not to
include a reasonably implemented system along these lines, but you
should never be obliged to use it and the default theme(s) should be
native and not implemented in this manner. Also, I would discourage
people from writing themes which can only run in this manner, so part
of implementing it should be writing a translator that generates
native code (this should not be hard; translate it to C or something).

Lastly, implementing widgets in this manner is a bad idea. If you
can't do something using an existing widget, and you can't implement a
generic widget suitable for inclusion with the servers, then you
probably should be rendering directly and not using widgets at all. A
"Tetris widget" is always the wrong approach.

--=20
  .''`.  ** Debian GNU/Linux ** | Andrew Suffield
 : :' :  http://www.debian.org/ |
 `. `'                          |
   `-             -><-          |

--B0nZA57HJSoPbsHY
Content-Type: application/pgp-signature; name="signature.asc"
Content-Description: Digital signature
Content-Disposition: inline

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

iD8DBQFCU6lDlpK98RSteX8RAiOGAKCPlmJT8PE/u+uCA2/EdoEADmcCUQCfSAMj
MjGkBhfg/wz2Xjj5JlfkcXc=
=lFZJ
-----END PGP SIGNATURE-----

--B0nZA57HJSoPbsHY--