Especially when adressing the elite, it’s important for startups to do things right. This is something the guys behind IRSeek had to learn the hard way. But let’s start at the beginning. Some time in mid-2006, Eran Cohen and Ariel Berkman, leading officers of israeli based B&C Advanced Solutions LTD. came up with the idea of a Full-Text Search Engine for public IRC conversations.
IRC, which stands for Internet Relay Chat, is a typically plain text form of real-time Internet chat or synchronous conferencing, where one uses one of the various clients available, connects to servers and joins discussion forums, so called channels.
So, in March 2007 they went online with a site and sent bots to various channels on various large networks such as Efnet, Freenode and GameSurge. Besides a pretty unsuccessfull digg submission, it went largely unnoticed by the users, but they just kept on logging and feeding their database.
It dragged on like that for a while until on the 30th of November 2007, all hell broke lose with a TechCrunch article asking: Will IRSeek Have A Chilling Effect on IRC Chat?. Although IRSeek stated that channels would be dropped when file sharing activity or private conversation is detected, it quickly raised concerns in the community. Users did some research and were shocked finding logs containing their nicknames in the IRSeek database.
On Freenode, the largest network for the free software community (and 20th largest in the world), several IRC users from Germany gathered and opened #irseek-victims, a help channel to collect information about the IRSeek bots and general activities. The first bots were identified (and mostly banned) and the news began to spread around the world wide web. To keep better track of the gained knowledge, a wiki page provided by Arbeitskreis Vorratsdatenspeicherung was created within a few hours.
Despite the fact that they haven’t ever been asked or even notified of any logging activity, which was in clear violation with Freenode’s Privacy Rules, the community was disturbed by the stealthy behaviour of the bots. As the bots were using TOR exit nodes to connect to the networks, had human-like nicks and identified as ordinary IRC clients it was quite difficult to spot them. The usage of Tor was later on explained with the need of avoiding connection limits (per IP).
While Christel, Freenode Head of Staff, discussed further actions with the team and wrote a blog post on the staffblog to clarify the situation, chx, who happened to be in Israel at that time, got in direct contact with Ariel Berkman.
Soon after the news hit the slashdot frontpage and Ariel Berkman and Eran Cohen joined the #irseek-victims channel and got in direct contact with the community. The site was taken down shortly thereafter and, together with the IRC users, channel operators and network admins, new service principles were worked out.
Since January 4th, 2008, as announced in the newly created IRSeek Blog, the IRSeek site is back up. Due to the popular demand, the majority of the logged content has been removed and the following changes have been made to the proceeding:
- Opt-in and opt-out:
- Formerly, the IRSeek staff just decided which channel to log and which not and put a bot in it. Now, after the policy change, channel operators or contacts that want their channels monitored can freely chose to do so by submitting a form on the IRSeek website.
- If a channel operator or contact decides to abandon the service, he can do so by using that form. Even historical conversations can be removed that way.
- Easily identifiable logging bots:
- Every bot belonging to IRSeek will use a “IRSeek” prefix on its nickname and username so that they can be easily spotted.
Personally, I think it is nice to see that IRSeek listened to the community and willingly changed the policies to what everyone can accept and live with. Of course, this is essential for a startup like this, but hey, they could’ve gone just another way.
Since those days, Ariel and Eran are available for any question on the Freenode Channel, looking for channel operators which find their archiving service
useful and being available for any question. It was there where we found the time to do the following quickinterview:
Why are you doing IRSeek?
IRSeek: We believe that people on IRC are constantly producing a lot of quality content (people would argue that a lot of it is useless talks by friends - which is also true for other mediums on the net. We think that there’s a lot of that too, but still a lot of high-quality content produced by experts helping others and discussing various interesting topics that can be of great value to others) that is accessible to a very small number of users. Namely, the users that are taking part of a certain conversation at a given time. Most of the people on the Internet are not aware of IRC (or perhaps aware but not familiar with how to use it properly), and therefore can’t enjoy this medium.
Why is such a service for the community and not against it (e.g. exploiting it for financial purpose):
We definitely think that this service is for, not against, the users. If you are referring to the privacy concern of the users, then you probably know that IRSeek isn’t the cause of the state-of-privacy on IRC, it was simply the trigger that demonstrated that there is no privacy on public-channels on IRC. If, as a result of IRSeek, people are now more aware of the actual privacy on public-channels on IRC, then I think this uproar had already resulted in a positive outcome. In anycase, IRSeek has changed its policies and is now asking for approval for channels that are being archived in order to avoid any confusion or misconceptions.
Our goal is to bring this wonderful content to the entire Internet community to enjoy and learn from. We also think that as people get to know the service, they’ll understand the benefits of IRC and start using IRC (or use it more often), which will in turn increase the benefits of other users from this service.
Are there any plans on making money with this service?
Our search result pages and conversation pages contain Google-ads for the purpose of covering the hosting costs, but as you may guess, it is far from serving this purpose at this time. Essentially, at this time we are focused on building the knowledge-base while taking into consideration the requests of the users.
