Monday, February 07, 2005

Looking Into A Reverse Engineering Alternative

Reading Andrew Tridgell's note about how he (and his team?) develop Samba is very amusing. He has four level of experiment, which were executed (repetitively and very patiently) over 12 years. Yes. That is more than 4000 days. I can only imagine his patience by comparing what our team have done (which only reach step 2) with his. But of course my project had only several weeks to spent.

But looking back to my "journey", we found a very crucial step in the protocol we need to emulate, hidden in the very tiny detail: just a brief initial conversation which must be done in the different form than the rest. Without that preliminary step, you can't talk at all. The machine will turn deaf, maybe laughing on your face, but refused to say anything.

This reminds me of my itch: what kind of application can we use to help us dissecting protocol? Should we take source code of Ethereal then write our own dissector? Can we do it more interactively by defining protocol structure (or format, or whatever you want to call it) in an editable window, then apply this to live data stream?

So, this could be a long-lived dream for me ...

By the way, I got that french cafe analogy from Myths About Samba by Andrew Tridgell in Groklaw, refered by Linux Today

No comments: