Slashdot Mirror
Google To Create "Blog" Search; Potentially Remove From Main
Skyshadow writes "Google, search engine of choice for pretty much everyone, has announced that it will begin a seperate index for blogs and remove them from the normal index, handling them instead in much the same way as their usenet archives. This will hopefully put an end to the recent difficulties locating primary source material among the mountains of blogs which are clogging the ratings system." There's been comments from elsewhere that says they won't be removing them - but that remains to be seen.
Is there any chance of having an RSS feature for journals, for everyone or even just subscribers?
I, for one, am sick of searching material only to find that the page is some asshat's blog. Nothing against blogs, but you never know where this material came from.
OTOH, what constitutes a 'blog'? Is Slashdot a blog? Is this a blog? The lines are constantly being blurred, and I'm not sure it'll be easy for google to make that distinction.
My journal has hot
Most of the useless information people put into blogs. Although, when you search for information, would you want to search 2 different locations? This is the whole claim to googles fame. I have found that many times people post how-to's in thier blogs along with other information.
If it ain't broke...don't fix it
-Rob
I also like the analogy made by the article to the voting system where a page votes for a topic: an expert site on turtles voting for turtles once a day every year vs. a blog mentioning turtles once in that same period leads to the expert site winning.
I hate liberals. If you are a liberal, do not reply.
I wonder if this is also intended to stop Googlewashing? Google has a history of trying to 'play fair' - and the power of a few well connected blogs to basically 'take possession' of any term works against that philosophy.
Am I the only one who thinks it is funny to see all the anti-blog comments everytime a weblog related story is posted? IMHO, Slashdot is a weblog.
I think I originally found Slashdot on RobotWisdom-- yet another weblog. But that was a couple of years ago...
One of the biggest newspapers in Norway, where I live, has recently said they believe blogs to be the new 'killer app' for delivering information on the net. The problem with that is that the treshold for publishing 'news' is so low, anybody can do it. This makes it very difficult for people to find the info they are looking for. At the same time there is no guaranty the info is useful or even correct. A good reputation will be more and more important for businesses and sites on the net.
/. would get though ;)
This move by google tells me newspapers in norway aren't the only ones seeing how influental blogs will/could become.This is a truly great step forward if Google could come up with a way of rating the different blogs. That way you could easily find serious tech-blogs.
Wonder what rating
Be like the twenty-second elephant with heated value in space-Bark!
As a previous poster briefly mentioned, what exactly is a blog? Would Slashdot forums be considered a blog? What about the myriad of ezboard message board forums out there, as well as other discussion websites? If the answer is no, it would be seemingly difficult and perhaps only of minor benefit to seperate just the true "blog" sites while ignoring the other sites.
And what about ebay? Quite often I am searching for info on an old piece of electronics I've picked up someplace, and I do a goole search, hoping to find information about the item. Well, all I get in return are ebay links to a similar item that was sold on ebay a few months ago. And even then, I click on the link, hoping to see what the item sold for (and thus get an appraisal), but the auction has been removed from the database due to it being several months old. Why index ebay pages? It's really frustrating.
Loomis
"The television is the retina of the mind's eye" - Videodrome
It's a reasonable solution, I think. Is it worth tainting the vast majority of the search results with useless blog entries just so that the (very) few blogs with good information will still show up?
This solves their problem with bloggers manipulating search results, yet still keeps the information available to those who want it. Granted, you have to know to look for it, but it seems to me like a fair trade-off.
I really don't mind finding blog links when I search for something, as they usually at least link to some relevant sources.
On the other hand, it is really a pain to search for help on something, and instead of getting a useful, authoritative document, I'll get a half-dozen archived unanswered mailing list posts from people with the same problem. I would much rather Google address this dilution from mailing lists.
The requested URL
What exactly is a Blog? Can anyone answer? What makes it different from a person web site full of links and comments, such as has existed on the Web for more than 5 years? What makes it something "new"?
I work at a company that has a blog-like recap of political news of interest for our clients and friends. If google tries to separate all sites with blog-like content, won't this naturally reduce my rank without actually increasing the source of information? Or am I missing something? How is google going to search for blog-like sites?
I love this idea... and I have been waiting for something like it for some time...
Think about it... I would love to search the blogosphere to see how widespread certain news items have become, or how widespread a certain opinion is...
You could use something like this to measure the spread of ideas (at least within a vocal and technologically suave minority).
Alright, fair enough - but how do you identify a weblog? They can do this for blogger/blogspot/whatever that they bought, and maybe standard software like moveable type etc. But what about sites based on slash, phpnuke or totally custom code? And where does a weblog begin and a news site end?
Filtering out usenet news is relatively easy, but weblogs? Mhhh, I shall remain sceptical until I see it implemented.
Somehow I can't drop the feeling that this will be very similar to a spam filter...
"I only speak the truth"
Karma: null(Mostly affected by an unassigned variable)
There is a wealth of categorization systems out there. Generally, they "position" the sites in an imaginary, highly-dimensional space, depending on whether keywords occurr (and how often/prominent etc.), and on certain structural properties of the documents. You can then try to define separating hyperplanes, which are functions that devide the ("feature") space into separate compartments, so you can group documents together.
Usually, these systems are trained on a set of sample documents that are already categorized, in this case, for instance, a thousand blog pages and tenthousand non-blog pages.
An example for this would be Support Vector Machines and Joachim's text classification algorithm.
Relevant keywords (from the field) to look for include "Maximum Entropy Models", "classifiers", "categorization", "Bayesian *" (whatever), "Neural Network Classifiers", "Data Mining"...
Hmm, I'm hoping the results are excluded, and blog is a "tab" just like the web, images, groups, directory, news are now.
I've found this mechanism to be really effective in helping me find what I want.
I use the google toolbar - this defaults to a 'web' search. 95% of the time what I'm looking for comes up on the first page. If not, I can click on the 'groups' tab, where my search is repeated (like when I'm trying to figure out an error message or somesuch).
If the thing I'm looking for is a business, or a product or something likely to be listed, then the 'directory' tab will give me good results.
Having a 'blog' tab (and keeping the results out of the main web results), seems like a good arrangement to me. Most of the time I'm not interested in results from blogs, and it doesn't seem too much extra work to just click one more time on the main results page to repeat the search in a blog-specific area.
I've found some of the best information on blogs.
I think it depends on the kind of info you are searching for. In my experience, most of the blog results aren't helpful. I've wanted a way to filter them out (usually putting in -comments -posted or similar helps).
There is much cruelty in the universe, John.
Yeah, we seem to have the tour map.
but filtering out ephemeral content in general would be good -- blogs would be included in this. so would mailing list archives, news stories, online stores, auctions, discussion groups, etc.
when i'm searching, i almost always prefer a page that somebody authored and put up as a permanent resource (or as permanent as the web allows). the top-level pages of the ephemeral sites would probably be good to keep in the main index, though i'm not sure how you index, e.g., the /. homepage.
-esme
I don't know, my blog has some very useful information that Google serves out to a lot of people needing help, for instance, this page is a lifesaver when you hose your Win2000 install using Easy CD Creator, and a lot of people still e-mail me, 2 years later, to thank me for writing it up.
When in danger or in doubt, run in circles, scream and shout. --Robert A. Heinlein
Well, I guess I shouldn't have specified live journal. My guess would be that easy publishing websites (blogger, live journal, etc...) are more often (but not always) used by people who just want an online journal. Also, the name "live journal" implies that it's a journal, not a blog.
In fact, my first 'blog' was hosted on blogger. It was mostly a journal. Then I switched to hosting it myself with more advanced software (movable type) and my blog migrated into a more news-oriented feature. As a result, I split my blog into a more journal-oriented blog and a news/science/politics blog.
I agree completely that a blog is about getting what you want to say out there. That's what I use mine for. I was merely responding to a comment that indicated that all blogs were just about mundane things that happened during the day.
neurostarRather than separating stuff, why not make it a series of choices using check-boxes. Example:
Include Web-pages: [X]
Include Blogs: [ ]
Include Usenet: [X]
And so forth. You can get better combos this way, If they add other "web types" in the future, you can combine searches without having to go to each one. They could still include a dedicated listing if they want, but I hope they don't hard-wire their data that way to prevent or reduce multi-factor searches in the future.
Even more generic would be to have a pull-down list of the "strength" of each search. Thus, if you wanted weblogs included, but given less weight, you might assign it a lower number. Zero would be the same as a no-check above. However, this is perhaps too confusing to most users.
Table-ized A.I.
...they remove the blogs from main, then re-incorportate the highest-hitting blogs from the new search back into the main? Then you may not miss a relevant and useful blog while avoiding the one that is mainly about some highschool girl and funny text messages that she got from her friends?
My blog can kick your blog's ass