This is an example from Google Maps France :
As a key test this search on Google Maps: “etoffes et sofas.”
Click the View from 154 from “etoffes et sofas”.
Then, click View maps of the community.
Finally, click on the link Maison & Jardin – rss.ohmybuzz.com.
You can start a search by typing this query: rss.ohmybuzz.com.
The link to the RSS file will be this: http://rss.ohmybuzz.com/search/?l=fr_FR&c=3.

It is necessary to modify the URL : http://maps.google.com/maps?hl=fr&tab=wl

But this request is OK from Google Maps in English : ohmybuzz
or this one : childrens clothing and accessories ages
Click on the link : See community maps
We’ll got this mention : Kids, baby & Family – rss.ohmybuzz.com

Sorry for my very bad english…

Havent really investigated the ranking that closely, other than it seems to do a reasonable job 😉 Those links seem to suggest it likes links that are credited with atom tags, but know content is indexed without them.

I just remembered Google created a map (now out of date) showing the reach of its GeoIndex, with a intro here:
http://www.gearthblog.com/blog/archives/2007/07/heat_map_of_kml_and.html

so it been indexing geo content before people started putting it in their sitemap (realising googlebot would take notice of it), just picking it up from normal weblinks I guess.

Thanks to both of you!

It is clear that Google has been and will be gathering more and more geodata for map building. It is starting to get to the point where POSM (plain old search marketers) need to familiarize themselves with the technology and its implications for search.

It also appears to me that it portends a large transition to Maps as their primary search interface on mobile hardware and many future devices.

What blogs do you recommend for us non geogeeks to gain a better understanding of the technology?

Mike

Indeed…. nice post. Once this is easy to understand and implement by the average user, Google will have to produce search results page exclusively for Google Map results – outside of the existing google/local search…. there will be a need for more comprehensive results.

Re blogs,
of course the Google LatLong blog already mentioned, also the Google Maps API blog, often contains (but slightly more technical) post about Maps and KML in general (rather than just the API).

There is also lots of good stuff on
highearthorbit.com, geoRSS, Mikel Maron to name a few.

Another challenge facing Google in building this GeoWeb thingy is that lots of Web2.0 sites, create lots of (potentially unique) geolocated content (wikimapia, tagzania, plazes, the GEC, flickr to name a few) , that could be aggregated, but is all deeply buried and quickly changing. Of course, getting sites to publish their content in KML is a good way – as long as it can be spider-ed efficiently. For smaller sites this works well, but for the larger sites wonder if there is a better way, personally think we have it working fairly well on geograph.org.uk but again lots of good content is being missed, due to the inherient ‘crawler’ discovery process.

Re spidering RSS, there is a ping mechanism somewhere behind the scenes somewhere (and handled automatically by most blogging software), which thought various properties Google do pick up on, so that accounts for that I think. There isnt such a mechanism for KML (yet ;))

> You say that doesn’t happen as well on larger sites, why is that?
I guess ‘well’ is relative. The main issue is scale, bigger sites of course have more content, so that is harder to keep track off, and can take longer to crawl. A small site can be crawled quickly, but a site rapidly updating, might not easily be kept track off.

> … create a KML file for each business listing as opposed to a consolidated file for all the listing. That seems like it might present other benefits/problems?
I agree, again speaking from experience (on geograph.org.uk), we publish multiple feeds (for recently uploaded content), grouped files (actually for Google Earth, but useful for the GoogleBot too!), and individual files (one per photo). All seem to have their place, and seem to compliment each other (although there is duplication). The feeds and grouped files, are good for get images indexed in bulk and quickly, but ultimately the individual files seems to end up being the best ranked (possibly as they seen as the authoritative).

Time will tell which would be the best strategy, (I thought the grouped files – for efficiency (speed of crawling and bandwidth), the success of the individual files caught us unawares – and was only included in the sitemap to curb their success – they where getting ranked above the photo pages themselves! and how KML files and indexing is better understood this could be reevaluated?)

It begin with it was in the main web search results, where the KML files where showing in preference to the website – that partly has been made better by the websearch being more aware of what KML file is (or rather that its not a webpage – much like feeds are now generally excluded from web results) – prioritising with a sitemap was a quick win there (unless it was a red-herring and Google tweaked it their end anyway!)

Also individual files had great success in the geoindex, partly I guess as these have been in the main web-index for longer (they had been on the site for a long time, but the grouped files where only added around the same time), and being crawlable from normal webpages, gave them authority. On the other hand the grouped files are in their own content island, and where only really discoverable from interlinks between files, and the sitemap.xml entries.

As time has progressed and the grouped files have had time to be indexed, they now do about as well as the single files, and at a quick glance its about a 50-50 split. So I guess it is tending toward ‘bulk’ files, we just had early success with single files, as we caught the uprising in the geoweb 😉

– I really should put together a timeline, to summerise all this 🙂

Flush Brackets – wont bore you with the details, suffice to say they are markers installed by out national mapping Agency Ordnance Survey, to record and help measure height above sea level. Its a pastime to go and visit these 🙂 [I am more into the Trigpoints – the top one – myself]

Hills – not totally sure what you mean, but the source for that layer called itself the ‘Database of British hills’, and doesnt have a nice accessible website. Also never got round to making a nice accessible version of that mashup, it could be so much more!

🙂