__/ [Big Bill] on Monday 21 November 2005 14:31 \__
> On Mon, 21 Nov 2005 13:52:20 +0000, Roy Schestowitz
> <newsgroups@xxxxxxxxxxxxxxx> wrote:
>
>>__/ [Big Bill] on Monday 21 November 2005 12:24 \__
>>
>>> It looks as though Google are favouring inbounds from article sites
>>> and ignoring them from freebie directories. Early analysis has it
>>> that, despite article site not being exactly choosy about what goes in
>>> them, Google are affording their links a decent weighting.
>>> So then, prepare for article spam. Sheesh!
>>> I've never bothered to sit down and do articles because really, so
>>> many of the existing are no-brainers, wrong or both so where's the
>>> point? Where's the actual benefit?
>>> Pooh.
>>> That's all I can say. Probably repeatedly on approximately an A4 piece
>>> of virtual paper.
>>>
>>> BB
>>
>>What /is/ an article? What is a blog item? What is news? What is dross? If
>>Google intend to distinguish between one and another, they are yet to face
>>big challenges. mainstream media like the BBC or the New York Times rarely
>>link. Any blog these days can pretend to be a news site too. I see many
>>sites that are driven using blog software underneath, but have been
>>modified to disguise that. Moreover, such sites have link bars, which often
>>contain blogrolls.
>
> Blogrolls, eh? Maybe that explains all the crap that's written on
> them.
Not quite. My point was that an article is 'decorated' with a variety of
links, which can even be selected at random. It would be hard for search
engines to pick up just the links within articles without wrongly accred-
iting external links that are automated.
If articles were written in a well-structured manner e.g. under <div
class="article"> as opposed to <div class="links">, all would be simpli-
fied, assuming the Webmaster isn't deceitful. What a search engine /could/
also do, on the other hand, is decide to just stick to a small number of
'trusted' sites and pick up all links from there. This will lead to prob-
lems with niches and rarities though. It's a similar issue to training of
a spam filter using collaborative flagging. You must establish trust.
Roy
--
Roy S. Schestowitz
http://Schestowitz.com | SuSE Linux | PGP-Key: 0x74572E8E
3:20pm up 18 days 11:14, 4 users, load average: 0.16, 0.25, 0.50
http://iuron.com - next generation of search paradigms
|
|
