--- On Mon, 05 Jul 1999 12:23:11 -0400 Jeffrey Perry
<j.perry@xxxxxxxxxxxxxxxx> wrote:
>
>Yes, one comment. I think this is the right approach, I would suggest
>taking it one step farther though. If you build the parser into
>ethereal, then no recompiles would be needed for adding a new
>protocol. This is how Etherpeek on Windows works.
>
I like the idea of a parser -- it would surely help when adding protocols. It
might even make some existing decodes cleaner. I'm not sure about Jeffery's
suggestion of building it in, though. I have two concerns:
1) If there were a lot of protocols added dynamically, mightn't the startup
time be long? That is, if the program had to parse a few zillion lines of
protocol before doing any captures, this might take some time. This could be
alleviated by using the dynamic parsing primarily for development and then
encouraging folks to add compiled versions when the decoding is mostly stable.
This would probably necessitate two versions of the parser: one for the
built-in and one for generating some sort of code which could be compiled.
[BTW, I do not necessarily mean C-complied. Another option is to have the
parser be external to the program and generate a byte-code which could be
stored in an array or structure, or which could be read-in when the program is
started.]
2) I think we need to be careful about "feeping creatureism"[1]. I'm not
saying we're there, yet, but I can see it comming if we're not careful.
--john
[1] For those of you who haven't hear this it is a sort of "Spoonerism" for
"creaping featureism".
-------------------------------------
Name: John McDermott
VOICE: +1 505/377-6293 FAX +1 505/377-6313
E-mail: John McDermott <jjm@xxxxxxxxxx>
Writer and Computer Consultant
-------------------------------------
