Subject: | Re: A more efficient Usenet
| Date: | Tue, 16 Sep 2003 19:46:58 +0000 (UTC)
| From: | davidsen@tmr.com (bill davidsen)
| Newsgroups: | news.software.nntp
|
In article <FkdPa.163566$lK4.4588694@twister1.libero.it>,
Mirco Romanato <painlord2k@yahoo.it> wrote:
| Yenc have CRC for the message and the full file (I trust Road Warrior
| about this, because I don't know Yenc well enough).
| But Yenc messages can not be checked against the full file before
| having the full file DLed.
| For this is needed a Merkle Hash Tree type of hash.
I'm not sure what this buys you, if all the parts check out good and the
assembley checks out bad, either the assembly of the parts was faulty or
the creation in the first place was. Using a tree to check larger
subparts doesn't buy much in terms of getting a correct whole, if the
parts check out as good you don't know what to redownload other than the
CRC of the whole file, perhaps.
My main complaint with yEnc is that at least half the encoders produce
broken images which can't be assembled by a decode written to the spec.
And non-UNIX software only does slightly better while accepting clearly
non-standard files. I'm not sure that having a Mac, UNIX and Windows
version of a standard is desirable :-(
--
Bill Davidsen <davidsen@tmr.com> CTO, TMR Associates
As we enjoy great advantages from inventions of others, we should be
glad of an opportunity to serve others by any invention of ours; and
this we should do freely and generously.
-Benjamin Franklin (who would have liked open source)
|