More IRC spies

Mary · SearchIRC Admin Joined: 03 May 2003 Posts: 686

I don't know if this has anything to do with Google, but its certainly ominous for IRC.

http://cosco.hiit.fi/irchiver/FAQ/

It is a FAQ for a project called Irchiver.

The company, CoSCo of Finland, researches and develops information retrieval systems. They feel discussions in IRC channels have scientific value.

Irchiver does not ask permission to join networks or log channel discussions, and they do not give channel owners notice of their presence. They feel that since IRC is public, not encrypted, and logging is a standard feature on many IRC clients, that it is "fair" for them to join channels and log discussions without the ops knowledge or permission.

Logs will be published, conversations will be studied, and data made available on the web.

The FAQ says, the success of this project is ultimately up to IRC administrators, and "Users should be informed e.g. in MOTD about general logging policy."

I found this particularly interesting:

Mary · SearchIRC Admin Joined: 03 May 2003 Posts: 686

A little bit more research this morning shows that CoSCo's Irchiver joined bots to ALL of Freenode's channels at ONE time on October 30th. 10% of the network's channels noticed the bots and banned them, and apparently this was reported to the network, which glined the bots about 10 minutes later. The leader of the Irchiver project then joined a channel with lilo to discuss their project with all interested parties. The website states that they have since joined IRCnet, but did not log channels there.

I visited Freenode last night to see if the network was cooperating with the Irchiver company or not. Although some staff members were aware Irchiver had visited the network, they did not know details.

CoSCo claims to be both "academic" and "scientific", however they do not show any affiliation with an institution of higher learning, rather the speaker on Freenode referred listeners to a website that shows other CoSCo projects which have commercial advertising applications.

They claim their search engine will be open source, which raises concerns that other interested parties will now be launching bots into IRC channels to log user information and channel text as well. One can only speculate that Google's recent visits to networks are an experiment to see if there is value in a similar project for their own search engine.

The bots on Freenode came from irchive.it.hiit.fi

Asmo · none Joined: 06 May 2003 Posts: 28

Hmm, interesting info Razz

moonman · Lurker Joined: 15 Jul 2003 Posts: 212

honestly, i don't see a problem. irc is an entirely open medium by nature. if an individual channel wants to keep information out of the public eye, they would do best to set the appropriate modes to keep prying eyes away. that's why networks give us +s/p/i/etc.

Mary · SearchIRC Admin Joined: 03 May 2003 Posts: 686

Well... SearchIRC already has bots on all of the networks, we catalogue all the channels, we have a search engine that filters and can mine IRC data (which is way different from html) from info gathered from each and every one of those channels, we have contacts with IRC administrators, and I like to feel we have a good reputation for doing what is right for the IRC community. We could have something like Irchiver is TRYING to do, all set up by midnight.

:P

Trust me, the people on big networks don't need me to tell them why this is a Bad Idea. There have been numerous well funded companies who have tried the same thing as Irchiver and Google. They have all failed - for one big reason. The people who create these projects know about IRC, but they don't know IRC at all.

Mary · SearchIRC Admin Joined: 03 May 2003 Posts: 686

Asmo did an excellent job of reporting on Irchiver in today's IRC-Junkie. He wrote directly to CoSCo, the company that is developing Irchiver, and found out that this project is not purely academic nor scientific as claimed (as we suspected), but totally commercial in nature - that is, bots are sent into IRC channels to to see what everyone is talking about, in order to track trends that inform targeted advertising. (Translation for the marketing impaired: they want to watch us so the spambots will know exactly which nicks to message about that porn url.)

http://www.irc-junkie.org/index.php#newsitem1068809748,58828,

Ville H. Tuulos · Guest

As the main person responsible for the Irchiver project, I feel obliged to correct few misunderstanding here:

1) CoSCo, Complex systems computation group, is a research group in the Department of Computer Science at University of Helsinki. Currently we work as a part of HIIT, Helsinki Institute for Information Technology which is a joint institute by Unv. of Helsinki and Helsinki University of Technology. All this information can be found on our web page.

All our work has a scientific motivation, yet it might (and hopefully will) have use also outside the academic community. We aren't in any means commercial.

With respect to Irchiver, we are especially interested to study IRC as an example of a highly dynamical environment, like a time series, in which we can utilize our statistical models.
That will be part of my Master's thesis.

2) We have no motivation nor intention to publish or collect logs in the long run. We are not going to compete with SearchIRC.com, which IMO is a great service, thank you for that. We want to collect a test data set to study feasibility of the idea w.r.t our models.

We will publish our code in open source to benefit the whole IRC community. As a proof-of-concept we hope that we can provide a public web demo which probably will work with a static data set. However that would require us to cooperate with some IRC network so that we can get appropriate permissions to log and show the data in public. Of course we won't do it without permission.

I'm happy to answer any further questions,

Ville H. Tuulos / Complex systems computation group
tuulos@cs.helsinki.fi

Guest · Posted: Nov 18, 2003 6:36pm Post subject:

Let me get this straight. You want to load massive amounts of bots onto IRC networks, and log channel discussions, for your Master's thesis?

dusk · none Joined: 18 Nov 2003 Posts: 3

Points for originality. Smile

Asmo · none Joined: 06 May 2003 Posts: 28

Guest · Posted: Nov 20, 2003 1:25pm Post subject:

I welcome our new targeted ad IRC spamlords.

I like the idea of robots monitoring my IRC conversations to help a company develop targeted ads (for educational use and hopefully commercial use!)

Guest · Posted: Nov 20, 2003 1:30pm Post subject:

Don't be so harsh.

The notion that the software will "eventually" become open source means the project can do no wrong. After all, with the source code anyone who wants to can launch robots like these without having any clue what they're doing. That is a good thing (tm)

Guest · Posted: Nov 20, 2003 1:40pm Post subject:

We all know anyone can listen in to a cell phone conversation.

Since cellphone users shouldn't Expect privacy, perhaps our new overlords can be convinced to switch to publishing phone conversations on the web, complete with the name and phone numbers of the participants, so telemarketers can better target their advertisements.

Ville H. Tuulos · Guest

Ok, my bad. Maybe I've not been clear enough. In my point of view there're two different things:

First: Developing a search *engine*. That means the code, the kernel, for a search engine. I understand that for most of the people "a search engine" means a web page. But if you try to understand search engines in a technical or scientific point of view, you see that *developing* and *researcing* theory and code is the thing we're interesting in. Not doing some web pages. We happily leave that business for those who know it better than us. You may want to read our FAQ again having this in mind.

However, as any other open source / research project we might want to show and demonstrate our results in public. But that's not our main intent.

Asmo:

Asmo · none Joined: 06 May 2003 Posts: 28

I was unable to find back the log of the chat you had on Freenode Mr. Tuulos. Care to provide a link again to that?

It is also interesting to see some parts of your FAQ have changed. Yes in some parts you added a date to it, but in others you did not...