The founder of RealNames, which rose and fell with the fortunes of the bubble and Microsoft back in the late 90s, has regained his domain, launched a limited search engine, and has plans for more…..here’s and excerpt from his blog…
have recently reaquired the RealNames domain name – realnames.com. This is some 30 moths after we were forced to close the company.
It feels good to have 100% ownership back of a thing I spent 5 years creating. To be honest I’m not yet sure what I will do with it.
Anyway, I have all of the old data and have created – over a weekend – a new search engine based on the RealNames data. Yes I coded it myself – and it shows….
….For what it’s worth I believe there are enormous opportunities to innovate in search today. The crawl and index technology that has done such a good job in dealing with the staic web is very poor at daling with today’s web. New challenges mean lots of potential to innovate.
Additionally the problem addressed by RealNames – that is the poverty of the DNS as a naming and navigation system for the world’s internet users – remains unresolved.
Google’s direct navigation via Keyword feature [ in the Google toolbar], and Microsoft’s version of the same thing (try typing a natural language Keyword in the IE browser that does not have the Google toolbar installed] are both falling far short of what is needed – a standard, natural language, naming system, available through all browsers, and embedded as a sub-index in all search engines, with the ability to have names registered in all human readable scripts.
Additionally the problem addressed by RealNames – that is the poverty of the DNS as a naming and navigation system for the world
As one of your readers I would prefer not to encounter explitives, including modifications thereof, in your posts.
Buh? Where’s the explitive (sic) in any of the above?
RealNames had at least two major flaws. The first was a lack of sufficient interest by the market… sure, if everyone in the world decided that RealNames ought to be the canonical reference service, they all might have paid, tens of millions of keywords would have been registered, network effects would have kicked in, etc., etc. But the market doesn’t seem all that keen on ceding such a monopoly voluntarily. Network Solutions (and now that part of what had been NETSOL resides at VeriSign) was granted a cash-cow monopoly by the Dept. of Commerce, and that’s unlikely to be replicated, absent government mandate.
The second fatal flaw was the same thing afflicting the domain name system: name contention and collision. If Ford wants “Explorer,” but so do the descendents of Lewis, Clark, Magellen, and Capt. Spaulding, who mediates the fight?
And if RealNames wasn’t a compelling solution, pre-Google, I don’t see how its position would be any better now that there’re pretty impressive free text search services.
Realnames had two business:
1)Trying to replace the URL with natural language in the same way URLs replaced IP addresses. It was a good idea and still is a good idea, but last time around AOL, unlike MSFT, wasn’t willing to be a party to it because of their desire to keep AOL users *inside* AOL. I fault AOL’s lack of vision, as monetarily it would have made all the sense in the world for them to forgo AOL Keywords for Internet Keywords. One need only look at AOL’s lackluster subscriber growth to know that walling off their users from the internet was a strategic mistake.
2)Paid search. No need to wonder about that business model, as GOOG & YHOO have proved its merits. RealNames ***gave*** the idea for paid search to Bill Gross, and RN has no one to blame but itself for having done so. It’s also worth noting that RN was the company that showed Google what PPC revenues could look like, prior to which Google had no idea how they were going to make money off their cool search engine, no idea at all other than their search appliance. I know this to be true because I witnessed the selling process RN had to take Google through to warm them to the idea. When Keith turned down NSI/Verisign’s $1B+ offer to be bought, he said RN was worth tens of billions. He was right, but not in the way he thought he was. He thought business #1 (above) was worth tens of billions, and it wasn’t, at least until he corralled AOL; business #2, though, *was* worth tens of billions, as YHOO/Overture & Google have gone on to prove.
I’d have to disagree that “replacing the URL with natural language” was or is a good idea, because of the inherent messiness of natural language. RN, and its ilk, have to necessarily create 1-to-1 mappings (e.g., of “Ford Explorer” to http://www.ford.com/lines/explorer/), and the world doesn’t really work that way.
Things ought to map to URLs, but it isn’t 1-to-1 keyword tags, at least not like RN tried it.
Actually RealNames never insisted on 1-1 mapping. We had that – and in the browser that was the implementation. However, our API supported fuzzy match with relevant result sets. AltaVista, for example, implemented it that way. We delivered the first 1-n results for them and sold these on a CPC (we called it PPC then – Price per click) basis.
Chris Zaharias is right about a lot of things but not that I turned down a $1 bn offer for the company. No offer was ever made for RealNames – by anybody. Lot’s of hints of offers, but never an offer. And secondly, the CPC model really DOES miss the point about the long tail. There are billions of potential Internet Keywords – in English, Chinese, Korean, etc. This is seriously a big need and a big business. But Chris is right that I seperated navigation and search in too srtict a way. My product manager at the time was Diego Fonstad and he argued that at the time, as did Chris and his brother Nick. The browser deal with Microsoft – in 1999 – blinded me to the search business we had already built. We were in every search engine except Yahoo! Way ahead of Overture in that sense.
I should have managed both opportunities, not only the browser one.They are not only compatible – they need each other – as Ross states.
It sounds like the white paper I’ve been working on for CommerceNet ought to be out in the next week or so; it addresses product codes (e.g., UPCs) as unique identifiers. But not, a la the code-scanning ventures like DigitalConvergence (the :CueCat) or Barpoint, as something end consumers need be explicitly aware of. (We all use ASINs–Amazon’s semi-proprietary codes–when we navigate the Amazon site, but I don’t know or even note ’em, they’re just the unique identifiers that drive transactions internally.)
I also don’t think such things as the apparent Neomedia/SAIC collaboration will exert any monopoly over resolving such codes… there’s decades of prior art in using codes as lookups to information (and companies like GXS, owned by Francisco Partners, and QRS, just bought this Monday by Inovis, have whole product catalog businesses based on this). Perhaps some patent claims on particular business processes, e.g., delivering multimedia presentations to customers in store environments, but not “gimme a code, get some information,” macro-level processes.
The most obvious reason why realnames failed is because it tried to reinvent the wheel. The hierarchial domain name system is easy to understand and uses the dot notation to signify depth and structure in a single piece of text (e.g. explorer.ford.com and magellan.explorers.info make perfect sense in their own right). It’s a little patronising to imply non-techies may not get this 🙂
Also, the market seems to like a clear distinction between domain names and sponsored search – merging the 2 doesn’t help anyone as people usually know whether they want to go to an actual website or look for sites that may have what they’re after. The 2 are simply NOT the same activity – which is one reason I never use the “I’m feeling lucky” button on Google.
While I generally agree, I’d add that problems that the hierarchical domain name system does have also applied to RealNames too, e.g., only one company gets to be “Ford.com,” and getting caught in the middle of who gets to be called what is inevitable, and unhappy.
Product codes, on the other hand, both the previous generation’s (UPC, EAN, etc.) and the coming Electronic Product Code (EPC), are pleasantly free of all that, as some root authority (e.g., the UCC for UPCs, and EPCglobal for the EPC) makes the assignments, the numbers are merely semantically-void serials*, and there’s no disputes as to who owns what.**
* There is the issue of current UPC owners wanting to be assigned EPCs that match, but I don’t think that EPCglobal is going to do anything dumb, and not have that be the case. If P&G “owns” the UPC prefix 037000 according to the UCC, EPCglobal will doubtless award them 37000 as an EPC Manager Number.
** I have run across an interesting little issue… there are a handful of companies that will sell you a unique UPC code, e.g., to allow you to put one specific UPC on your garage band’s CD for $20 or so. This is specifically prohibited in the UCC’s licensing (one is not allowed to “sublet” one’s UPC assignment); when I asked them about it, they noted that yes, it was not permitted, yes, they knew it was going on, and cautioned that buying UPCs thusly could result in your products being confused by retailers and such. They have contract on their side, but either don’t care, or don’t care enough, to go through the effort to shut down this (minor) abuse.