[NNTP] Compressed LIST (and other commands) answers

[Ade Lovett]

I am still waiting to see evidence (anecdotal or otherwise) that there is a sufficient "problem" out in the wild that compression (in whatever form) will "solve".

NNTP's strengths (and weaknesses) are a result of its inherent simplicity.  One can implement a compliant server in a few thousand lines of code, over the course of a couple of days (yes, I have 3977-compliant proof-of-concept servers written in perl, python, ruby, and if you give me a week or so, I can probably knock one out in ksh).

The fact of the matter is that, as yet, no-one has provided a compelling reason as to why compression (or lack thereof) should be embedded in the protocol itself, as opposed to being layered over the top (in a similar manner that stunnel provides for "secure" transmission over <random protocol>)

We live in days of multi-megabit home connections.  Sure, not everyone has them, but I would hazard a guess that the majority do, further compounded by the fact that said majority are downloading non-compressable data.

All I ask, before spending time and effort on standardizing a compression methodology, is that the proponents offer up a real-world, real-case scenario, where such an implementation will provide _significant_ improvement above and beyond the current NNTP experience.

-aDe