Wow. I'm really sorry that I caused all this ruckus -- it appears that I'm not fully acquainted with the rules of ettiquite here on k5.
I'd like to apologize if I offended anybody. There have been a bunch of questions posted to the original story and to this one -- I'll try to answer as many as I can.
Why was the story posted anonymously?
I despise cookies, and try to avoid them on any site I visit. I read slashdot, k5, and advogato, but without an account at any of them. After I posted the story, I wanted to check it in the submission queue (ok, and put up a +1 vote for myself <grin>), so I created an account ("megacz") for that purpose. If I had had the account at the time I posted the story, I definately would have used it. Also, lowerbound.org has my contact information on it, so I don't feel that the post was truly anonymous.
Why didn't you post an actual story intead of just propagating a link?
Ok, this is yet another one of my goof-ups. I assumed that the best thing would be to keep my post short and simple, and let those who are interested explore the site. Apparently kuro5hin's style is to write a short-and-sweet teaser for the front page, and then a detailed analysis for the comments page. I'm sorry that I didn't do this. See below for a bit of belated detail.
Why did you post this to kuro5hin?
Honestly? I thought people here could benefit from this technology the same way I have. I figured that the people at kuro5hin would appreciate it.
Was this post malicious?
No, just inept.
Do you consider your post to be spam?
I'd like to think not -- I chose a group of people who I believe are especially interested in a price-metacrawler and let them know about it. They collectively voted it good enough to post.
Untactful? Yes. Spam? No.
Are you making money off of lowerbound.org?
Interesting copyright explanation on the site, btw. Isnt it effectively just a metacrawler, but only for price sites? Nothing new there as far as legalities, I'd think..
No -- metacrawlers (ie dogpile) republish the content they crawl verbatim. Our site actually comprehends the listings it downloads (see the last question in this post) and publishes a summary. It's no more illegal than publishing a review/summary of a book you read (but not the book itself). Comprehension is the difference. Think about plagarism vs. research.
Do you correct for shipping prices?
Unfortunately most of the engines don't post shipping prices (AFAIK, only shopper and pricewatch). So we really don't have enough data about that to post. We're looking into ways to take shipping into account -- if you have any ideas, let me know!
So, does pricewatch really offer the lowest prices?
I guess this should have been in the original post. For pure-commodity items (ram, cpu's), yes, pricewatch tends to whomp on the competition. However, for lower-volume commodities (drives, sound/video cards), we've found ibuyer.net and shopper to be contenders. The other engines clock in at the top every once in a while.
Interesting note: IMHO, ibuyer.net is way too disorganized for humans to use -- I think lowerbound as a front end to their site actually makes it very usable. A good thing, too, since they have really competitive prices.
How many hits have you gotten from k5?
Somebody asked this, and when I went to check I realized that my hosting provider has chosen an asinine setting for the apache log file format which doesn't record referers (I'm having him fix this right now), so I don't really know.
When are you going to release the source?
Heh. Good question. We're still discussing the best way to deal with this request -- keep your eyes peeled later this summer.
Can you tell us a bit about lowerbound?
I started coding lowerbound in the early summer of 1999 when I was working for Sun Microsystems in California. At the end of the summer I asked Tobin Coziahr to help me out. Since then we've been working on it on-and-off between classes at Carnegie Mellon University (To other CMUers, yes, I'm posting this during Carnival...).
Lowerbound implements some really neat technology -- aside from crawling all the pricing sites, it actually understands the product listings that it downloads -- it knows how to recognize that a listing is for, say, a Tyan S1894S/L/A motherboard, and that all such motherboards are ATX form factor with a 440BX chipset, even if the listing doesn't mention those facts.
More interesting than that, lowerbound can actually learn from the listings it downloads! We're currently testing this only in the Audio/Sound_Cards section. We teach lowerbound how to recognize a whole bunch of different brand-name sound cards. Lowerbound then sorts the listing by which brand name they belong to. By scanning the listings for, say, a Creative Labs Sound Blaster PCI128, lowerbound notices that a large proportion of the listings claim that this card is Full-Duplex. If enough independent listings confirm this, lowerbound will infer that ALL SBPCI128's are Full-Duplex, and attribute this feature to all SBPCI128 listings.
Search pricewatch, streetprices, and others all at once with lowerbound.org (now with discussions!)