Download next hublist..

Archived discussion about features (predating the use of Bugzilla as a bug and feature tracker)

Moderator: Moderators

Locked
OLDoMiNiON
Posts: 202
Joined: 2003-01-06 06:22
Location: Salford, England.
Contact:

Download next hublist..

Post by OLDoMiNiON » 2003-06-06 09:13

If you have multiple lists in your config, 'd like to be able to skip through them of my own accord, not only if one list isn't working.. since, some may be working, but are just crap.. heh.
Hope u understand me, it's too early for computing :-S

HaArD
Posts: 147
Joined: 2003-01-04 02:20
Location: Canada http://hub-link.sf.net
Contact:

Post by HaArD » 2003-06-06 17:02

Me too! I usually end up going in to settings changing the first hublist c&P then refresh....

D/L NEXT PH LIST would be very cool!

HaArD

Nazo
Posts: 68
Joined: 2003-04-03 14:35

Post by Nazo » 2003-06-08 22:12

Hrm, while on that matter, it might be pretty cool to get the lists from them all and merge them together removing any obvious duplicates. (Same address or exact same description provided the description is long enough to actually be likely to be unique.)

OLDoMiNiON
Posts: 202
Joined: 2003-01-06 06:22
Location: Salford, England.
Contact:

Post by OLDoMiNiON » 2003-06-09 07:47

..thus defeating the whole point in splitting the massive bandwidth between multiple machines ;)

Wisp
Posts: 218
Joined: 2003-04-01 10:58

Post by Wisp » 2003-06-09 07:53

Nazo wrote:Hrm, while on that matter, it might be pretty cool to get the lists from them all and merge them together removing any obvious duplicates. (Same address or exact same description provided the description is long enough to actually be likely to be unique.)
yep, and optional filter all the bad hubs out (timed out, banned, full, etc)

OLDoMiNiON
Posts: 202
Joined: 2003-01-06 06:22
Location: Salford, England.
Contact:

Post by OLDoMiNiON » 2003-06-09 08:12

did you see what i just said?

The whole point in multiple lists is to split the bandwidth load between hosts! If every DC++ client is modified to dld every list, just think how much bigger the bandwidth loads on the lists would then be! ..the eventual outcome would be that most of them would shut down, and you'd be left with only the few fastest hosts.. and possibly the worst lists ;) ..you just don't know. So - BAD feature! :)

Wisp
Posts: 218
Joined: 2003-04-01 10:58

Post by Wisp » 2003-06-09 08:34

OLDoMiNiON wrote:did you see what i just said?

The whole point in multiple lists is to split the bandwidth load between hosts! If every DC++ client is modified to dld every list, just think how much bigger the bandwidth loads on the lists would then be! ..the eventual outcome would be that most of them would shut down, and you'd be left with only the few fastest hosts.. and possibly the worst lists ;) ..you just don't know. So - BAD feature! :)
If DC++ just could download all the hublists once and then filter all the bad hubs out and refresh the number of users, etc, you don't have to download the hublist every time

and btw, a hublist is only a few kb, so i guess a server won't be overloaded soon if a lot of people use it, but i could be wrong

sarf
Posts: 382
Joined: 2003-01-24 05:43
Location: Sweden
Contact:

Post by sarf » 2003-06-09 09:24

Unfortunately, the current model of hublists are that they take care of validation / updating the number of users, not the client. The hub-owners would also be quite upset if all DC++ users occasionally tried to ping them to see if they were valid and how many users they had.

The size of the hub list is somewhat irrelevant, Wisp (though they are larger than "a few kilobytes"), since the usual way to see the public hub list is to download it over and over again (the NeoModus hub list is actually based on this behaviour). When arnetheduck checked the number of DC++ clients a while back he estimated them to somewhere around 20-200k (can't remember how many zeroes there were after that 2 :) ), so there are quite a lot of clients that are willing, able and ruthless enough to whack most hublist servers if they were accessed by all (or even many) of the DC++ users.

Sarf
---
The value of a program is proportional to the weight of its output.

Locked