Slashdot Mirror
A9.com with Syndicated Search
joeykiller writes "Search Engine Watch reports that Amazon now lets you add your own search to their A9 search engine. Users can opt-in to use additional search engines in addition to A9.com's own when searching. Amazon has chosen to use an extension of RSS 2.0 for this, and hopes that this format will enable search syndication in the same way RSS did for content. Several add-on searches are available already, among them New York Times, Wikipedia and NASA."
I can't wait until Google does the same caus' I'd like to use this featrue to evaluate some of my client's sites popularity, day after day.
Trolling using another account since 2005.
I've been using Firefox's configurable search feature for a while (you can add plug-ins for things like IMDB and Dictionary.com, and then select the engine you want before searching.) While it's a great idea in theory, it also is a little more clumsy than you'd normally want. I've ended up doing too many searches for "Linux USB XYZ-123 driver" in Dictionary.com or Wikipedia or whatever, because it doesn't reset itself after each search. After a while, you stop using it - it's just not quite what it needs to be.
A9 will improve itself with this kind of feature. But, more importantly (like I said, two levels, by adopting this RSS-extension, it will encourage others to do likewise, which means other developers can put together tools for this kind of quick-search without having to learn a hundred different search engine APIs. This means there'll be enough tinkering with UIs for it to be virtually certain someone will come up with something usable.
At the moment, searching on the Internet's a little like a bunch of 1960s style fast food outle[tt]s. You get your basic cheese burger virtually every where, but it's literally just a bun, a beef burger, and a slice of cheese. A few places are adding pickles (like Google, you might call news the pickle, or groups the tomato ketchup), but we're a long way away from, say, the engines offering delicious Whoppers. If we want to have search engines that give us the full lettuce, mayo, ketchup, onion, tomato, etc, we need to standardize on protocols in much the same way as outfits like MacDonalds and Burger King were able to create efficient food transportation systems for raw ingredients beyond simple buns and beef. This is Amazon giving us the lettuce of Wikipedia and mayo of the NYT to us, asking us if we want the fries of NASA, and it's a step towards them offering "have it your way." Awesome.
You are not alone. This is not normal. None of this is normal.
can't you use the google api:
http://www.google.com/apis/
?
This is an interesting move by Amazon, and it's certainly an improvement over delivering static HTML pages from a database, an increasingly worthless exercise.
However, why do they need to extend RSS? I fail to see what the extensions are for, when all they really need is a XSL transformed RSS document or perhaps an XSL transformed XHTML document with an accompanying or alternative URL to fetch RSS directly.
Anyway, I like this because hopefully the next move is for the big three to start offering straight XML results; this in addition to or in leiu of the data APIs Google and Yahoo already make available to web authors.
Thing is, if they just published search results in RSS, then non programmers could jump in the game and start utilizing the data for different applications. It would open up the field considerably, but unfortunately, it would mean a lot more abuse too from search spammers would would seek to capitalize on the data.
I Want To Believe
I don't come to amazon to search the web. I don't come there for the billions of crap banners they splash over their pages. I just go there to buy books and read reviews on potential future purchases. They should make there interface simply and clean, like google, rather than the cluttered piece of crap they have now.
They try to do everything from a single page, and end up with a nasty mess.
This is a little off topic... ...but I have been running a protein structure comparison service (TOPS, if anyone cares :) that is a /kind/ of search engine.
What I think is best about our system is the presentation of the results, not necessarily the engine itself (oh well!).
So I was thinking that it would be cool if other 'search engines' (comparison services) could run a search and submit the results to our servers for display.
Kind of like this? Would need some kind of protocol for describing matches.
I use IceRocket for most of my non-technical searches these days, and use Google for technical searches. Nothing beats Google when you know a few rare keywords guaranteed to be on the page you want, but I find its utility has been on the wane for general interest pages.
Read: Besides, they are only extending the RSS 2.0 namespace... something done quite often. http://www.xml.com/pub/a/2003/07/23/extendingrss.
Get your Unix fortune now!
echo -n blabla | md5sum | cut -b 1-5
I have looked at A9 OpenSearch in the past, and what I can't see is any way to retrieve the OpenSearch description documents that people have submitted.
At the moment, it looks like A9 are keeping these to themselves. So you can make your search engine compatible with A9, but you can't use other people's engines without their OpenSearch document.
I think A9 should release these documents!
You cant make anything foolproof, they'll only invent better fools.
Well, I can see how this can be commercially exploited: Imagine best-deal/coupon/bargain search engines syndicated with A( OpenSearch and people willingly subscribing to these searches because they love good deals? The only difference will be that now they won't have to leave their primary seaarch engine. A9 has something here with the potential to drive a lot of business! And I guess they can collect a lot of information and simply use it to sell more products on amazon.com, being the largest/broadest online store and all that. Cool.
If web services (broad sense : google, amazon, ebay, blogger, wikipedia...; not the WS-* sense) standardize their input/output they are commoditazing what they make a living of.
Their added values are going to drop and new entrants will offer new services built upon the commoditized ones.
The problem is that nobody expects the new services and everybody will recognize them when they appear. It's a hard turn to take for the current rulers.
Is Amazon starting to shoot itself (and its peers) in the foot?
http://search.msn.com/results.aspx?q=slashdot.org& format=rss&FORM=ZZRE
Really, thats the only Ireason go there once a week to keep the small amazon discount that they give your account. Its not much but if its enough to push the price of something below a local store then I'll order it from amazon to avoid tax.
On the other hand, for engines like A9, whose business model relies on drawing clients to their main source of revenue (their store), this could be yet another traffic generator.
Haven't you ever heard of >a href="http://www.pubsub.com/">PubSub?
It lets you search for a keyword in millions of syndicated and nntp articles and get the results as rss.
To paraphrase David Spade:
I liked OpenSearch better the first time, when it was called...OpenSearch!
Er, this sort of lightweight protocol sounds
potentially really useful, but have we gotten
any assurances from Amazon that this is really
going to be "open" as we understand it?
In particular, has Amazon stated that they're
not going to try making patent or other proprietary
claims on it, like they have with one-click and
other questionable patents?
Right now, what's to stop them from letting the
protocol get popular, and then turn around and
say "sorry, guys; now that you're hooked,
you can't write your own OS producers or clients
without getting our permission or paying a fee"?
What i'd like to see is search engines utilising my websites own search facility on it's own.
/search.xml, like it already checks for the existance of indexing guidelines in robots.txt.
Google, for example, never indexes everything, but could easily extend it's "More results from " hyperlinks to access my website's own engine, for example
This would mean search engines index lots of very complete indexes instead of trying to index the entire web from the HTML and only managing to create a partial index.
That is awesome. Thank you!
You are not alone. This is not normal. None of this is normal.
and use their datamining to its fullest
the reason a9 exists at all is to help them not you, you are simply a resource to be used and abused at will because billions of dollars clear profit just isnt enough
My own site's search engine has been providing RSS feeds of its results since last December - not that anyone's actually using them. Having a standard for doing this seems like a pretty nifty idea; I'll have to update my code to generate this format. I don't have high hopes for it being ad-free when provided by a commercial site like Google, but I've got software that can filter ads out of RSS feeds, too.
Jon Udell has also has written about OpenSearch.
CK
You will forget this sig before you next see it
LOL! hahaha I think France (one way or another) has managed to piss of Yahoo!, ebay, amazon and google now!
Well done AFP! (oh c'mon that was on topic right?)
#hostfile 0.0.0.0 primidi.com 0.0.0.0 www.primidi.com 0.0.0.0 radio.weblogs.com
I am still waiting for a9.com to load. You cannot slashdot google.
A9's is the easiest yet. You don't have to register for any special account or get permission. You just go to the submission page, give them the URL of your site description, and they validate it (as well as your sample search), giving you a preview.
The XML itself is a few simple extra tags to your RSS 2.0-compliant feed. If you already have a site search engine (like this one) it takes a few minutes to add an XML output option.
Once you get it working, it's immediately available as an a9.com column. It's that simple.
Incidentally, at IMVU, we use a customized version of the open-source osCommerce package. It'd be pretty neat if OpenSearch came standard on projects like this.
Can your IM do this?