Whenever ads are submitted to Safarri, I host them then submit them to Oodle and Base.
This sends people browsing either site to Safarri.

When people are browsing Safarri, I fill in Base ads under the Safarri ads. They are not quite as relevant, but I have developed some algirithymns that make them pretty good.
Since Safarri’s interface is awesome, this makes browsing Safarri even better than browsing Base. This sends more people to Safarri.

In my opinion, Base is a a win/win situation!
(And if you don’t believe me), check out Safarri.com

The website is http://www.phpbase.org

It has a plug-in style for attributes and schemes. But currently only Housing scheme is defined.

I guess that the larger the database , the more you need to think about ranking. Also, the less relations between the items, the more difficult that has to be.

What about ranking items in an enormous (”world’s”) database where items have virtually no linking between them (ie. products, classifieds, images, etc)?

I think that 2 simple aspects of digg (and other projects) are inspiring and worth exploration:

1.extracting ranking directly from traffic/users
2.owning traffic/users by taking manual submissions (as opposed to crawling/rss, etc).

A not-so-unrealistic scenario could be one in which the leader is not the one controlling the items-db, but the one controlling the largest users-feedback-db and extracting the most value from it.

If that’s at least partially true, the leader could be completely open about search results without sharing its real assets.

I don’t see Google Base as an innovator or even a bookmark. Probably, the only potential value it is showing right now is free exposure (and just because of the brand).

But we need a carrot to motivate webmasters to choose the RSS solution: Yahoo!, upcoming.org and the gang must be able to to something useful with these feeds asap.

Google Base - Atom 0.3 Specification

Why couldn’t I do / use both (and others)?

Microformats indeed needs more pixels to offer a decentralised, empowering version. I will continue to evangelise about MF as a component of what I call SIDE but if a client comes to me and starts asking about Google Base submission, I’m going to balk. IMHO, we (meaning all of us “2.0-ers”) have a duty of care towards the information we sheppard and we should get our collective arses in gear to develop robust open standards so access to the more meaningful, aggregated and categorised lexicon of all human thought and knowledge doesn’t one day get become synonymous with Google.

1st penny– GOOG is launching carefully to test their semantic tagging scheme… which is structured at the high level and unstructured at lower levels. In order to validate their scheme, they first need critical mass of meta-data.

2nd penny– assuming the test works, they’ll open up APIs for the rest of web to mashup (e.g. a la GoogleMaps). However, it’ll be critical for them to tie the API to a GoogleWallet to share transaction revenues. *But*, the transaction dollars will be ad dollars, not consumer purchases… this is the reason why I believe Eric Schmit has been saying that GoogleWallet will not be a direct competitor to PayPal.

More at: http://dannyayers.com/archives/2005/11/17/google-base-more-semweb-bootstrapping/

You’re asking the right kind of questions.

I wanted to direct your attention to a slightly broader topic, that of semantic markup. I wrote an article on Digital Web Magazine about it: http://digital-web.com/articles/writing_semantic_markup/

In this article I mention microformats as well as structured blogging (similar to ROR) and RSS. All of these technologies are doing the same thing: providing a semantic structure for markup. Microformats, though they have a lot of press, aren’t necessarily the best way to solve this problem. I’m not sure what is, but there are definitely alternatives…

Also, thanks for the comparison with Google Base, it’s a very interesting one. My optimistic side hopes that they simply rewrite their templates to spit out more structured code…

-Josh

the angle on microformats vs GoogleBase is right on target: those are 2 distinct alternatives to consider for the future. one is a centrally-hosted option (which has benefits, but significant downside if not searchable by anyone but Google), and the other is a distributed option, but “centrally-searchable”.

in other words, in a microformat world i post my data (job listing, classified ad, restaurant review, etc) on my site or blog, and using tags & microformats, that data is discoverable by anyone who knows to look for those tags & can recognize the data structure defined by the microformat. this is what i call “open access”.

in a GoogleBase (or Craigslist or eBay) world, i upload my data to a hosted DB, specify the structure as i post the listing there, and then the data is discoverable by anyone who uses the associated portal — and, if that data is searchable by other engines & crawlers, then it can be discovered by anyone using other meta-search or vertical search services. but if it is *not* open to search by others, then i’d say it’s more of a “walled garden”.

what is interesting about these 2 alternatives is that at first glance they look somewhat similar — they both allow anyone to post data that could be discovered & searched by a large audience. however, in the 2nd case, unless the data is searchable by other services, the size of that audience is limited by the hosting portal.

furthermore, depending on the web services offering that host portal provides, there could be a limited ability to build application workflow on top of that portal. from one perspective, if the web services offering is very rich, perhaps that could be an advantage you’d trade off for a more limited audience size. however i think as vertical search engines and vertical applications develop, assuming a one size fits all stratey will work is unlikely.

ultimately, i believe the tradeoff is whether you trust a centrally-hosted data service to provide access to data for others to use and enrich further, or whether you’ll buy into a potential ‘walled garden’ due to other benefits (ease of use, richness of web services, audience size, etc).

however, as the variety of other services grow, and data becomes available / searchable / remixable, and other web services offerings debut, it becomes increasingly unlikely that any one portal will offer a rich enough platform to offset the downside of the ‘walled garden’ — unless they choose to be truly open.

as powerful as Google is (or Yahoo or Microsoft), i don’t believe there can be only one.

- Dave McClure
SimplyHired.com | SimplyFIRED.com
Master of 500 Hats

my general take is that

In fact we are shoked to see how Google Base is similar to ROR, except fot the openness issue.

The ROR guys