>From Hubert Chan on Thursday, 21 June, 2001:
>>>>>> "Joseph" == Joseph Pingenot <[EMAIL PROTECTED]> writes:
>below). Although if you volunteer to make it happen... :-)
>Hubert> Changing all the packages to work properly wouldn't be a simple
>Hubert> task. (Not saying that it's a bad idea, though.)
>Joseph> Aside from the issues of creating a unified command line
>Joseph> interface, I don't see how.
>Well, the problem isn't in creating the wrappers (IMHO). It's in making
>the front-ends work with the wrappers: it wouldn't be as simple as
>doing a search-and-replace for pgp with dpgpw. Some programs (I'm sure
>that x-pgp-sig-el isn't the only one) think that they know something
>about the pgp implementation, and so making these programs work with the
>wrapper might require some extra effort.
*Now* I see what you're referring to. Too many trees.... Can't see
forest.... :)
Although it *sounds* complicated (and probably is), an interim
solution could be a 3-tier (3-animal? :) approach:
0) Application calls, say, pgp with pgp syntax.
1) A pgp-xlat package (?), maintained by the PGP person, is used
to translate the pgp commandline to the generic commandline.
It would create a PGP->dpgpw translation, invoking dpgpw at
the end. PGP wouldn't necessarily need to be installed, only
pgp-xlat, which would have a pgp->dpgpw translation wrapper in,
say, /usr/bin/pgp. If PGP happens to be installed, it could be
called, say, /usr/bin/pgp-real.
2) dpgpw then uses the dpgpw-<implementation> wrapper to translate
the generic syntax to the implementation-specific syntax
3) the implementation is called and all goes on as if pgp were
actually called.
The only prolems are:
a) This is pretty complex
b) This involves even *more* packages to be installed. This point could
be minimized if all pgp implementers worked together to create a
*single* pgp-xlat package containing their specific translations
to the generic commandline. Hrm. Or all translations could be
bound up into the ever-less-virtual generic pgp package.
c) This *still* doesn't address the problem brought up before of the
different implementations' different, well, implementations. :)
d) This requires work and lots of coordination.
This would *might* go a *ways* to making *most* front-ends be able
to use *most* pgp implementations, but the best solution still remains
getting frontend developers and pgp implementers to sit down and
unify on these things.
Sure is a fun puzzle to problem-solve, though. ;)
-Joseph
--
[EMAIL PROTECTED]
"IBM were providing source code in the 1960's under similar terms.
VMS source code was available under limited licenses to customers
from the beginning. Microsoft are catching up with 1960."
--Alan Cox, http://www2.usermagnet.com/cox/index.html
