Ethereal-dev: Re: [Ethereal-dev] Re: Ethereal Gripe

Note: This archive is from the project's previous web site, ethereal.com. This list is no longer active.

From: "Ronnie Sahlberg" <ronnie_sahlberg@xxxxxxxxxxxxxx>
Date: Sat, 23 Aug 2003 15:02:40 +1000
One problem is that these languages might have to become more and more
specific to each individual protocol as they try to describe more and more
stateful features of the protocol.

Maybe one soon ends up with a specific language/compiler for each individual
protocol.
At which point the problem is translated from writing a dissector for
protocol X into
writing first a description of protocol X in a X specific PDL dialect
followed by
writing an X specific version of the PDL compiler.

Thus one may end up with it taking longer and more effort in tweaking the
compiler and developing the protocol description to feed to the tweasked
compiler than one would have needed to implement X directly as a dissector
in the first place.


The PDL language must be flexible enough to be able to express most common
types of protocols.
It must be easier to learn and master building PDL descriptions and building
the PDL compiler extensions that what would take for doing the dissector by
hand anyway.
It must be faster to do a description of X and whatever tweaks are nessecary
to the compiler than to do the
dissector by hand.

If it is more difficult or timeconsuming to develop and maintain the PDL
compiler and the descriptions than to develop and maintain the actual
dissector, then it is pointless.



----- Original Message -----
From: "Guy Harris"
Sent: Friday, August 22, 2003 5:17 PM
Subject: Re: [Ethereal-dev] Re: Ethereal Gripe


>
> On Thursday, August 21, 2003, at 11:52 PM, Tomas Kukosa wrote:
>
> > I would prefer using existing language to reinventing new one.
>
> I agree.  When reinventing the wheel, you run the risk of reinventing
> the flat tire. :-)
>
> I'll have to take a more detailed look at their language to see if, for
> example, our experience with reassembly (which is still a work in
> progress - there are probably more things that will need to be done to
> the reassembly code for protocols for which we don't yet do reassembly)
> will be helpful to them.
>
> _______________________________________________
> Ethereal-dev mailing list
> Ethereal-dev@xxxxxxxxxxxx
> http://www.ethereal.com/mailman/listinfo/ethereal-dev
>