Just getting the day started – been up for an hour or so now – checking email, deleting spam, checking the blog, deleting spam…agh. I still haven’t done anything productive yet, but things are organized and tidy now just waiting for me to tackle them.
What a waste.
Anyways – while enduring the latest trackback spam delete marathon I started to think about Blogware’s trackback implementation, the rather naive assumptions we made during its design and implementation (blog spam wasn’t really on the radar while we were building v1.0 so we made a lot of assumptions that we could simply clean up the spam on a centralized case by case basis. Not a really great model when trackback and comment spam become the norm…) This led me to start thinking about some of the new trackback and comment spam controls that the product management team are building for an upcoming release whereupon I realized how much a pain the ass comment and trackback moderation will be…I don’t want to have to moderate 500 trackback spams anymore than I want to delete them.
This led me to start thinking about different ways to tackle the problem which took me down the scorecard route. There is probably enough collective intelligence in the Blogware mob to build a scoring model that will rank trackbacks and comments based on the actions that other Blogware users have taken to deal with those comments. Might be a bit of work, and it would take a certain level of interaction on behalf of Blogware blog publishers, but fairly elegant and well-bounded.
But then I realized that there might not be enough real trackbacks to make all of this worthwhile. I mean, I’ve got a fairly popular blog with a fair amount of traffic – I’m no Wonkette, heck, I’m no AccordionGuy, but nevertheless, I get my fair share of comments and linkbacks. Thing is, I don’t really get a lot of linkbacks – most of them come from myself (when I link to something that I’ve previously written) and other Blogware users (AccordionGuy sends me a lot of love…). From what I can tell, not a lot of people are sending trackback notifications when they link to me, which means that I’m not noticing a lot of trackbacks. This has lead me to start using other tools like Pubsub and Technorati to see who is linking to me and what they are talking about.
So if trackbacks are only telling me part of the story, and the value of that data is far less than the investment I have to make it maintaining that data (by spending too much time cleanup up spam) then why am I still accepting trackbacks? I mean, no one else seems to be supporting them – why should I? For that matter, why should Blogware?
I’m not saying that there isn’t value in the function, but simply that there seems to be little value in the current implementation. I bet the blogosphere can do better.
I’m not going to turn off trackbacks right away. I want to do some more thinking and poking around – quantify things a bit. I’d be interested to find out what other bloggers are doing – does your blog tool send trackbacks? Do you accept them? Do you care about knowing who links to you? Do you tend to use tools like Pubsub and Technorati to find out who is linking to you instead of trackbacks?
Can't blogware tell if hundreds of trackbacks (across all blogware sites) in a short period originate from one IP address/block or point to the same web site? What would happen if just those trackbacks (or comments?) were held for approval by a moderator (and discarded if approval doesn't come within a certain timeframe designated by the blog owner -- so the response to spam is "do nothing")
As I understand it, the main point of trackback spam is googlebombing, so if they need to point to the same web site/page.
I'm sure that some sort of a signature analysis method could be developed that would allow Blogware to separate the trackback wheat from the spam chaff at a high level. The problem with an approach like this is exactly the same problem with signature based spam filtering. It depends on reactive intelligence which by nature will always be less intelligent than the state-of-the-art used by the threat source.
My central question is not whether the problem can be solved or how the problem can be solved, but really whether or not the current implementation is worth saving. Is the value of 'trackback" high enough to warrant any investment? Not just by Tucows, but the blogosphere in general....
I personally feel that trackbacks provide a ton of merit where they can almost act as an article verification providing a third party filter of what you are writing about. For example, if I make a claim about a topic that you agree or disagree with; comments shouldn't be the only way to chime in since they can only provide so much context without understanding where the writer is coming from.
Trackbacks have some of the same net effects as comments but allow the added ability to incorporate the tune of your own blog into the mix. I’ve found a lot of very interesting blogs by following trackbacks that I wouldn't have come across otherwise, and for me that provided a whole different spin on what they wrote in the first place.
That in turn tells me there is value provided all the way down to the reader so scrapping them entirely may end up cutting out a very important part of the value chain. Now the amount of time I spend deleting trackback spam is increasing every day so I do feel your pain, but like spam for email, I don’t think the best solution is to eliminate them completely but to find better ways of validating the authenticity of the context in the first place.
We're agreeing. I'm asking not if there is value in the application (i.e. finding out when people reference your website) but if there is value in the implementation. Trackbacks seem to be horribly broken in their current state and I'm wondering whether or not there is enough value in the current implementation to try and save it or whether using something different to serve the same application might be warranted...
I'm not only talking about people just referencing my site, but also when I read other blogs and follow a trackback without necessarily being in on the dialogue.
I think there is always a better mousetrap out there that we should be striving for, and of all the people I know you are probably top of my list of someone who could figure one out for trackbacks, but I would love to understand what about the current implementation is "horribly broken" past the ability for people to abuse the intention?
I mean, how else would your blog communicate with another when you link to them, and reference that communication publicly? That public reference of the exchancg of content is important, and that will always be a carrot to the spammers due to the links so not sure how you could achieve the same net effect without breaking it completely.
"what about the current implementation is "horribly broken" past the ability for people to abuse the intention?"
Precisely that. There are no mechanics in place to prevent the abuse.
"how else would your blog communicate with another when you link to them, and reference that communication publicly?"
Doesn't matter. There can be other ways - there are other ways. The technology lives separate from the application - the important part is figuring out whether or not there is value to the application. Only that can tell you if making an investment to save, improve or replace the technology is appropriate or not.
"Precisely that. There are no mechanics in place to prevent the abuse."
There were no mechanics in place to prevent the abuse of email, but we didn't scrap that when unethical marketers figured out how they could force their messaging on an unsuspecting population. This is why we have so many innovative business models baked around the spam defense vertical which is turing into a pretty impressive industry all on it's own.
"the important part is figuring out whether or not there is value to the application"
If we go back to the drawing board you'll either have to acquire mass adoption by all vendors since there are always two people in a converation or scrap it altogether. Trackbacks bring too much value to the conversation to eliminate completely, and I think you'll soon see innovators that will take trackback / referral / comment spam and create business models around helping the rest of us prevent it. You gotta love the free market...
BTW, ditto on seeing you behind the wheel again, it's great to have you back in force ;)
In a perfect world we (bloggers) would have some type of legal recourse for trackback spam or spam in general. Wouldn't it be great if you could go after the guys putting up the spam and sue them. If the upside of spam only had a downside ($$$)
Take a bite out of spam (excuse the pun) Join a class action lawsuit against -insert name here- for $100 000 000 bucks or better yet 2 to 5 years in a small room to contemplate your actions.
Sadly, in the end all I am left with is a delete key and a public rant as recourse.
TrackBack URL:
http://pokerboy.journalmine.com/blog/_trackback/1285640
P.S
Nice to see you blogging again Ross.
happy trails on the new wheels...
R
P.S Nice to see you blogging again Ross. happy trails on the new wheels...
Thanks dude! Nice to be back.