+ New General Search Engine from Convera Debuts: govmine, Not Just Government Info

The search engine scene never slows down, not even for a moment. Since late 2004, we’ve been hearing and reading about Convera (a major player in the government enterprise search space) preparing several services for mainstream consumption. In March they launched their Excalibur web index product to companies (HighBeam and Clusty for example). Excalibur technology powers govmine.

Yesterday, they launched (beta) and available to all, govmine: The Alternative Search Engine. News release here. Partnerships with Quigo (contextual advertising) and RocketInfo (to power news and blog database). While the service is, “specifically designed for government professionals,” it does not only contain federal or state government web pages. So, we think the name is a bit misleading even when trying to appeal to a governent worker.

From the news release:

Recent market data suggests that over 10 million government workers could benefit from an alternative type of search experience, one focused on professional search needs and less on consumer popularity. govmine can uncover the relationships between seemingly disparate data from across the Internet, supporting more informed decisionmaking.

Think conceptual/semantic searching that’s interesting but (yes, it’s a beta, we know, but is still in need of lots of work). Feedback is welcome and the company hopes to move out of beta by the end of the year.

As we try to do with other engines, we like to give them some time to work the kinks out. Nevertheless, here are some links and things to be aware of as you kick the virtual tires.

+ If nothing else, it’s another large web index (how large, we don’t know precisely) on the web. That’s not a bad thing.

+ Pages are cached with date and time stamp. Not every page or doc is cached. Cached pages also include images. So, govmine can be used as an image cache. Again, useful to the web researcher. Copyright caveats of course. While the IA does the best job it can, the more caches the better.

+ One search box on home page. Click the “go” button to search

+ This page explains govmine’s relevance philosophy. From the way it reads (we might be wrong), link structure plays little to no role in determining relevancy.
From the page:

govmine’s goal is to reduce the amount of time it takes to find answers and it does so by utilizing semantics which rank search results solely on the relevance of the term based on our understanding of the document.

+ Five different databases/catalogs: Web, Images, News, Blogs, .gov sites. .gov sites include federal gov, state governments and .mil sites.

+ Implied “And”

+ You must use the main interface to get to the image, news, blog, and gov tabs via a second click.

+ Advanced interface offers numerous features including limiting by site/domain (site:) and personalization by selecting categories to include or not to include. Interesting feature that I plan to spend more time looking at. However, for those who type 3.2 words in the search box and go, likely too complex.

+ Some PDF indexed. Cached versions of PDF’s converted into HTML docs.

+ Boolean AND, OR, NOT available

+ This help page offers a diagram of a results page and a key to icons used.

+ If needed and as seen elsewhere, disambiguation provided on results page

+ Like we’ve seen elsewhere, results are categorized on results page.

+ Like we’ve seen elsewhere, suggestions to narrow results on results page.

+ Related concepts and recent search also listed on results page

+ Results can be limited to Spanish or French

OK, those are a few of the facts. We ran a few very simple searches. More to come in future weeks.
1) Search Boston
Disambiguation but no mention for the 70’s band Boston.
2) Web Search: Boston: selected news web site category
Third result in this category is The Cal Poly News Web. The word Boston is not in the document, not even a mention of New England.
3) Image Search: Golden Gate Bridge
Not bad but the first 10 results included a picture of a vulture, a pic of a bridge (not the Golden Gate) and a map of Minnesota.
4) Vanity: The official url for this blog is not listed in the top 10 when doing a web search. It was also not found when doing a blog search. It’s listed as a news source. That’s cool.
5) Web Search: Danny Sullivan
Danny is a good name since not only do we know the search guru but don’t forget about the race car driver.
++ The most current page (about the Danny the search guy) in the top 10 was from 3 months ago.
++ The related concepts look good for Danny, our friend. Race car fans are out of luck.
++ The default categorized results will make race car fans happy. (-: Remember, more categories can be added on the advanced search page.

Again, just a beta so stay tuned. One thing we also noticed is that a good portion of popular pages/sites have not been crawled/refreshed in a while.

More Convera News from 2006:

+ Convera and Factiva enter indexing deal
+ UBM and Convera to Co-develop Search Engine for Healthcare Professionals
+ Treasury picks Convera for search

—————
—————

+ Ads coming to Google Maps (via News.com) and Google Blogoscoped
We’re curious to see how long it will take the very popular Firefox plug-in CustomizeGoogle (Free) developers to create an option to remove the coupons. They already offer a button that removes ads from Google Maps pages and just about everything else you could think of. Another option they provide is the ability to restore the right-click button to print images of Google Book Search pages. We’re sure coupons from Yahoo and MSN can’t be far off. MSN just released a free call feature that allows someone to search for a business, click the call button, and have the company, store, service, call them back in a matter of seconds. Our fried Greg Sterling reminds us that local search from A9 also offers this call feature.

This entry was posted on Tuesday, August 15th, 2006 and is filed under Uncategorized. You can follow any responses to this entry through the RSS 2.0 feed. Responses are currently closed, but you can trackback from your own site.