Talk:Direct Connect (file sharing)

From Wikipedia, the free encyclopedia

Contents

[edit] Protocol?

It seems that a lot of things have been put into the separate NMDC/DC++ articles, and that the parts specific to those should be moved off this page. What should go here, then? Maybe a basic description of the protocol? —Preceding unsigned comment added by 24.131.177.150 (talk • contribs) 15:39, December 7, 2004

Also, should we use DC_plus_plus or DCPlusPlus for the DC++ article name? the executable is named the latter, but C++ uses the former naming scheme —Preceding unsigned comment added by PseudonympH (talk • contribs) 16:21, December 7, 2004

67.171.11.171 removed "unsupported, unecessary sentences regarding the "record" for number of users on a DC hub" from the article. I was just wondering, that in which way are such user records "unecessary"? There are articles about all kinds of stuff, even weird fish species on Wikipedia. I don't think it is NECESSARY for anyone to know the average size of a damn fish, but some people might find this information interesting. The same could be applied to this article. Even though it isnt NECESSARY for anyone to know the highest amount of users ever fitted in a dc hub, many will find it interesting. Unsupported cannot really be said, because as said, there were around 14 000 people in the hub, and as such, many of them must have seen how many users there were in the hub, and if found, could probably support this record. Also, this hub was discussed on the hublist.org chat hub, and many of the people from there went to the record hub. If we thought that all of these people could be lying, we could as well remove the World War II article from Wikipedia, giving "100% of the witnesses were lying". Therefore, I propose adding the user records, and other information people could find interestring, to this article. -meh —Preceding unsigned comment added by HJV (talk • contribs) 20:20, April 23, 2005

[edit] More Info Needed

What's so great about the DC protocol? Is it secure? Does it do automatic CRC checking of files? How do download/upload queues work? Anything special? —Preceding unsigned comment added by 68.40.169.172 (talk • contribs) 23:44, June 16, 2005

Hi, I added some information on the weaknesses of the direct connect protocol. Also, removed some unneccesary NPOV attempts relating to the featuresets of the alternatives to the official client. No one questions that the alternatives are generally more featured. The OSS versions were made _because_ the official client was slow to add features and fix bugs.

I'm not badmouthing the protocol unfairly. I've programmed for it, and it sucks. I didn't say that, but did point out things that are objectively wrong with it.

In my opinion, this should not be merged with neo modus direct connect entry, as there is too much general information on alternative clients, protocol issues and the DC "scene" to lump all that with NMDC. The official client is basically on the way out, and the eventual replacement of the protocol with the DC++ protocol is going to wipe it out if Hess doesn't redo the official client from the ground up to support these new features.

Laters. —Preceding unsigned comment added by 69.81.22.227 (talk • contribs) 16:05, July 30, 2005

[edit] Merge

I'm against the merge. Direct connect file-sharing application is pretty much used as a blanket for DC incarnations... whereas NeoModus is the original... but now, not nearly the most popular. I think mergins ia bad. gren グレン 22:29, 29 August 2005 (UTC)

Well I merged Direct Connect network into this article. I reworded a lot of it and added headings, so I hope nothing got lost on the way. The title Direct connect file-sharing application doesn't seem right to me. Maybe its too long or too specific. Perhaps it should be something like Direct Connect (file sharing). —Vadmium 11:40, 5 December 2005 (UTC)

[edit] Belonging

85.179.26.170 wrote " Due to the centralized nature of Direct Connect, some doubt that the protocol belongs to the P2P family.". Who are these "some"? Source for that claim? —The preceding unsigned comment was added by Ullner (talk • contribs) 04:25, September 3, 2006 (UTC)

The Peer-to-Peer article defines P2P to be the same as what I'd call decentralized file sharing algorithms. I was around when Napster rose to popularity, and I made no centralized/decentralized judgement to consider a given network P2P. Even so, the P2P article itself mentions "hybrid P2P", which DC is (since the independent servers are akin to those used in Napster or eDonkey 2000). I don't think that sentence fits with how Wikipedia defines P2P, so I'm removing it. --GargoyleMT 12:12, 8 September 2006 (UTC)

That had bothered me for a while. The fact that it continues to run, long after the demise of neo-modus.com attest to the fact that it is definitely 'hybrid-p2p'. Thanks GargoyleMT. Jon H ;). —Preceding unsigned comment added by 24.23.189.131 (talk • contribs) 05:43, November 30, 2006

[edit] Protocol Section

The section on the protocol doesn't mention anything specific about the protocol. Instead, it describes some of it's inefficiencies, and basically recommends an alternative. I feel that the section should be removed, renamed, or substatially altered. Hotdogger 23:10, 20 February 2007 (UTC)

Agreed. --GargoyleMT 21:19, 21 February 2007 (UTC)
i have found backup link to protocol description: http://www.thenighthawk.biz/smf/index.php#16
Eh, well, it's not really a "description" of the protocol. It's 'just' the protocol commands, no explanation. We already got that linked to in the article... --Ullner 16:56, 11 May 2007 (UTC)
I've updated the protocol section. I hope it's better now. --Ullner 14:54, 2 August 2007 (UTC)

[edit] ADC

ADC protocol isnt really "better". originally it was supposed to use less bandwith and less resources than DC++ protocol wich is not the case, it removes the "myINFO" line wich was almost always unnessecary, but instead it intorduced something much longer INF line and some ID's that are in most cases unnessecary, wich makes it even worse than the original protocol, the only good thing is the id's ability to ban a user even if he changes nick/ip (or use the id for authentication) Gunnaraztek 16:36, 9 April 2007 (UTC)

What are you basing your judgement on? The initial INF is larger, yes, however; as people send updates as they're in the hub (share, slot, hubs updates etc), ADC will use less bandwidth, as less is sent. (Only the actual changed info.) I've attempted to document the differences at the DC++ blog, and so far I've only come up with instances where ADC will out-perform NMDC. Unfortunately, there have yet to be a real test with a hub and the consumed bandwidth over time. The different IDs serve a purpose, although they may seem redudant at first glance. In what way are they "unnecessary"? In addition, I'd like to say that this article doesn't explicitly say that ADC is better, or worse, just that its "designed to address the former's [NMDC's] weaknesses." --Ullner 17:13, 9 April 2007 (UTC)

Perhaps ADC should be merged (yes, there's no content on that page) here? Considering they both use the same hub-client concept? --Ullner 13:02, 3 August 2007 (UTC)

[edit] Dates

I'm finding the dates section a bit pointless. The only date that should really remain is the Open Direct Connect project reference. The dates of when Neomodus was created and when the project died might be interesting. Also, the date of the first third-party hub, and the first version of DC++. But I'm not sure any date at all is even interesting here... I'll remove that section, if there's no objection. --Ullner 13:07, 3 August 2007 (UTC)

[edit] Spam

Why isn't spamming problem mentioned at all? I mean those unsolicited bulk private messages sent to p2p-clients. For example, advertising (especially promoting rival DC hubs) and slot-begging (fully automatic!). It looks like this problem doesn't bother any developer, and absolutely no client-side countermeasure is available, while a plently of spamming tools apparently exist. Even such a simple feature like banning a user is not implemented, so you have to rely on firewall (which only helps if offender's IP address is static). Isn't is strange not to account for spam in the modern world? 217.172.21.161 (talk) 08:32, 23 May 2008 (UTC)