Hi Abaris

I wasn't offended by your post, but you made it seem like you weren't considering all the possibilities. But I thank you for the heated debate. It was constructive. Please forgive my sarcasm - I usually try to give a little anecdote at the end of a sarcasm rampage.....

:-)

Cable users download a file, delete it, and download it again next month to show their friend, then delete it again.

No modem user would do this. (OK, but, not often!)

And as the other guy said, they download DivX tying up download slots for a day. You think by giveing them the movies quicker you will satisfy them? They will just download three DivX a day instead of one. Build more highways if you want more cars.

I am a modem user, I share 20G of mp3, most you won't find elsewhere.

I also download two files at a time, and upload without noticing.

A 56 k modem has 5.3k of download and 3.4 k of upload dedicated. You can't change these figures. That's the way the modem protocol works.

A modem user maxing out their download bandwidth still has 2.4 k of upload bandwidth. My uploads typically go through at 2.5k whatever else I am doing (because I rarely ever am uploading anything by anything except gnutella - gnutella gets 99% of my upstream bandwidth).

Gnutella does have a fairly inefficient protocol at the mo, but the traffic does not seem to be too much that I have any trouble getting 4.5 k download. And still surf web pages and write junk like this. And upload at 2.5k. And search for about 8 files. All at once.

And it is because modems provide 'enough' bandwidth that most users will be on modems for the forseeable future. As bandwidth becomes more available in developed countries, countries which had no ISPs are getting ISPs and modem users.

Only when the whole world has been converted to broadband will modem users go away.

It would be nice if people would spend more time on increasing the search capability so it is easier to find a reliable source for a rare file and tell before downloading the file if it is what you want (ie bitrate etc). Figure out how to increase the horizon and make grep searches so we can search more servers for our files. Then if you have multisegment downloading you already have more sources to download, no need to implement this troublesome automated swarming. ;-)