[wp-hackers] Plugin Development Request: gada.be

Roy Schestowitz r at schestowitz.com
Sun Oct 16 12:35:54 GMT 2005


_____/ On Sun 16 Oct 2005 10:11:16 BST, [Chris Pirillo] wrote : \_____

>> I know this does not fit in this mailing list directly, but if gada.be
>> was ever to be considered as a WordPress plug-in, it's worth taking a
>> moment aside for assessment (or worse -- critique). As I continue to
>> explore your service, I find that some aspects of the results (other
>> than major search engines) can be filled with spam. Depending on the
>> nature of use, this can become a real deterrent because there is too
>> much noise for results to be worthwhile.
>
> I just posted something about that on my blog. At least, it's directed at
> Blogspot. I'm suggesting they deploy a captcha mechanism in their publishing
> system (not just for comments, but for entry submissions themselves). It's
> the only thing that could "easily" stop the bots - though I know they'd be
> throwing out the metaweblog baby with the bathwater.


Blogspot's problems extend (or /used to/ because Google started acting 1-2
months ago) beyond the issue of post, comment, and ping spam. One of the
notorious participants in our SEO forum was acquiring deleted blogs 
immediately
after disappearance, thus inheriting PageRank, which using the sidebar could
fuel his public domain content garbage (nearing a million pages). He was
Googleating for quite some time, accumulating over 15,000 visits per day.
Recently he has been less vocal so I imagine that Blogspot prosperity is over
for him.


> Since gada.be is acting as an intermediary, once we have user registration
> for our system, it's likely we could flag links as spam - but that, too,
> could be abused (think: competition). If we figure it out, though, we'll be
> sure to share our tactic with the rest of the industry. ;) del.icio.us has
> been spammed, too - although not like you think it has. We're certainly open
> to suggestions.


Yes, I am aware of del.icio.us' issues too. Just let that be used as a 
directory
a la DMOZ, for the purpose of improving SE's, then see how quality takes the
dive. It was bound to happen all along.

Even collaborative spam flagging for E-mail and comments are going to suffer
from link greed (blame backlinks algos). The technology is new so it is not
targetted yet, much like blogs which were not suffering from comment spam 5
years ago. Worse is fact is the fact that false positives in spam 
flagging lead
to disappearance of genuine messages, not obfuscation of Web content.


> One of the reasons why gada.be places a rel="nofollow" attribute into each
> and every outgoing link is because I don't want it to be directly used to
> game anything.


I posted the following URL in this mailing list before:

http://del.icio.us/robots.txt


>> The services you feed on might have to do likewise (or disposed of
>> altogether until they sanitise their sources) because you only
>> serve as an 'RSS bridge' here.
>
> I've already suspended / killed four of 'em for doing so, and contacted the
> ones who made it easy to contact. Seekitall was one of the worst offenders,
> and I'm about one day away from killing Broadsnatch.
>
>
>> By the way, there appears to be a bug as sometimes the navigation
>> footer does not show up.
>
> That's odd. We've got an update to our fetch code (100% scratch) coming
> within the next couple of days, making it easier to parse non-compliant
> feeds - of which there are millions. ;) Could be part of the problem. I'll
> keep my eye out for it, and register it in our bug-tracking system.
>
> I'm honored and humbled that you would take the time to help, Roy.


The pleasure is mine.

Roy

-- 
Roy S. Schestowitz      | "Black holes are where God is divided by zero"
http://Schestowitz.com  |    SuSE Linux    |     PGP-Key: 74572E8E
  1:25pm  up 52 days  1:39,  5 users,  load average: 0.76, 0.41, 0.20
      http://iuron.com - next generation of search paradigms



More information about the wp-hackers mailing list