Findability Blog - Latest Comments

Re: Video: Search Analytics in Practice

Kristian Norling — Sun, 13 May 2012 13:17:47 -0000

Thanks Otis! :)

I liked your Berlin Buzzwords video and I will certainly watch the others as well. Hope to be able to chat with you at Enterprise Search Summit in a few days. Looking forward to meeting you there.

Re: Video: Search Analytics in Practice

Otis Gospodnetic — Thu, 10 May 2012 16:18:28 -0000

Nice talk, Kristian! :)
Here are 2 more, one from Berlin Buzzwords and another from Enterprise Search Summit, both from 2011:http://vimeo.com/25120921 http://vimeo.com/31293628 And a quick one here:
http://blog.sematext.com/20... This one shows how Search Analytics can help with A/B testing:http://blog.sematext.com/20... here are slides:http://www.slideshare.net/s...

Re: Semantic Search Engine – What is the Meaning?

Pawel Wroblewski — Mon, 02 Apr 2012 05:04:50 -0000

I am eager to see how Google will handle this. Nevertheless you can see now some other examples of semantic web search. If you try Bing.com (US version) or Ask.com or Hakia.com you can see how they implement it. For example if you put to Bing "lung cancer" - you will be able to see not only the relevant results but meaningful suggestion to narrow or broaden your search; look for related medications (with proper names!) or seek for related conditions. So Bing somehow inderstood that two-words-query "lung cancer" is a specific disease and proposed several relevant actions to help you discover more info about it.

My point is that it is not so hard to implement it for your own company or enterprise!

My post is about how we could make it using search technologies like e.g. SolR.

Re: Semantic Search Engine – What is the Meaning?

Kerstin Forsberg — Sat, 31 Mar 2012 08:58:11 -0000

Nice blogpost. How would you Google's move towards a semantic (web) 1) relation to what you outline here?

1) http://semanticweb.com/goog...

Re: Semantic Search Engine – What is the Meaning?

Pawel Wroblewski — Fri, 30 Mar 2012 06:37:40 -0000

Thanks Jan for your comment. Nice to see that somebody has even thought about it.

Anyway, what you put is what you get. The quality on ontology is a separate topic itself but I believe there is a way to do it and in most of cases it should be developed. I also found out that there are ontology developers who specialize in it.

Agree that e-commerce is not the best place to experiment with semantic search while margins and conversion rates apply strongly. But there are quite interesting other domains like Knowledge Management, Intelligence, Web Search, Portal Search and maybe more?

Re: Semantic Search Engine – What is the Meaning?

Jan Rychter — Fri, 30 Mar 2012 05:55:30 -0000

Two observations from someone who has designed and built a semantic search engine (at Fablo):

1. What you described is based on strict taxonomies and ontologies. While those exist in many domains, in general you can't assume that level of data quality. This is why we built a semantic search engine based on latent semantic indexing: that way you can throw any data at it and have the engine figure out the rest.

2. What we learned is that in the e-commerce search market this is not a sellable proposition. The market is not ready for semantic search. What people expect is a straightforward and understandable search. There is also a browsability expectation: search is supposed to bring up *all* possible candidates and let users browse through them. In most semantic search solutions you end up with excellent top results, but poor browsability.

Re: Mobile clients and Enterprise Search – What are the Implications?

Björn — Mon, 19 Mar 2012 03:27:23 -0000

Great post btw!

Re: Mobile clients and Enterprise Search – What are the Implications?

Björn — Mon, 19 Mar 2012 03:24:42 -0000

"Use Javascript code to detect what type of device is accessing your search solution and if it is a mobile client you display the mobile interface."

Why JavaScript? With the use of css3 and media queries you should be able to change the CSS to suite any device. You probably need JavaScript to change menus to dropsdowns etc but not for device detection. Or am I missing something? :)

Re: Enterprise search – market overview 2011

choubb — Tue, 08 Nov 2011 08:11:18 -0000

can not agree more with you on the price trend. As customer are more knowledgeable on ES solution and vendors, the marketing is expanding with price drops.

Re: Gartner and the magic quadrants – crowning the leaders of Enterprise Search

choubb — Tue, 08 Nov 2011 08:05:31 -0000

Hi Both

Seems we are the both side. We have put a lot effort on top of solr, like management UI, DB connectors, user authentication. You can try our open source version here http://sourceforge.net/proj....

Re: Google Search Appliance 6.10 released

Margaret — Thu, 12 May 2011 13:29:35 -0000

Thanks for posting this! I couldn't find anything on the 6.10 GSA release anywhere - thanks again!

Re: Solr 3.1 released

Tobias Berg — Wed, 06 Apr 2011 12:01:11 -0000

Thanks Otis, I've added your link to the blog post

Re: Solr 3.1 released

Otis Gospodnetic — Wed, 06 Apr 2011 11:40:31 -0000

There is also a writeup on Solr 3.1 and new developments in Solr covering the last few months at http://blog.sematext.com/20...

Re: Gartner and the magic quadrants – crowning the leaders of Enterprise Search

Caroline A — Thu, 27 Jan 2011 22:46:22 -0000

Hello Jason!
I do agree with you! I hope you did not misunderstand the blog, I was merely summarizing the Gartner report. We are working quite a lot with Solr and have even made contributions to the Apache project.
So far, Findwise, has implemented more than 20 Solr projects and have many more in the pipeline.
..and as you say, it is quite obvious that Gartner chooses to somewhat diminish the open source community, which is really sad. It will be interesting to see if the next years Enterprise Search MarketScope chooses to comment this.

Re: Gartner and the magic quadrants – crowning the leaders of Enterprise Search

Jason Hull — Thu, 27 Jan 2011 00:07:38 -0000

It seems that one of the most obvious search engines is missing. Solr is the most downloaded open source search engine, and it powers some pretty well-known sites such as Zappos and the United States Library of Congress.

To ignore Solr is to turn a blind eye to the trend of adoption, much like ignoring Linux or Apache. I notice that Gartner includes only pay-for-license search providers; could this be because of a pay-to-play model at Gartner?

I daresay that users of the major search engines listed above (Microsoft nee FAST, Endeca, GSE, etc.) will point to weaknesses of Solr and users of Solr will point of weaknesses of the other engines. That is not the point of my comment. The point, though, is to note how a widely adopted, extremely powerful search engine, Solr, is inexplicably missing from the evaluation and reflects a gaping blind spot in the analytical prowess of Gartner. Simply making a hand wave over Solr and Lucene for unspecified reasons beyond "did not meet at least one criterion" is fallacious and pretending that the elephant is not in the room.

Re: OmniFind Enterprise Edition 9.1 – new capabilities discussed over breakfast

Caroline A — Thu, 16 Dec 2010 10:36:34 -0000

Follow the discussions around this post:
http://arnoldit.com/wordpre...

Re: If a piece of content is never read, does it exist?

Dave Maskell — Tue, 14 Dec 2010 12:21:01 -0000

I would like to add another aid to findability. Moving from the reactive Question-Answer exchange to a more praoctive approach of offering the user signposts to the next piece of information that might be of interest. Similar to the Amazon approach of "people who bought this book also bought that book". For example, suppose you are on a health web site and do a search for "Diabetes". Once the search engine has returned the usual list of of keyword matches, it is logical then offer the user some related term suggestions for the next search, such as "Symptoms", "Treatment Centres", "Heart Disease" etc. Integration of the search engine with a metadata taxonomy enables this kind of approach even with unstructured data. Search users are often trying to discover more about an overall subject rather than just answer a discrete question.

Re: If a piece of content is never read, does it exist?

Alexandra — Mon, 13 Dec 2010 23:38:48 -0000

My take on this issue: http://contentperspective.s...

Re: Search is a journey not a destination

Maria Johansson — Fri, 03 Dec 2010 15:01:33 -0000

Sorry that you were disappointed in the level of presentations at the conference. Of course one could look at this lack of inspirational presentations as a sign that we, here in Scandinavia, are not as behind in the innovation and development of search solutions as one might think. :)

Another reason might be that there are no really mind blowing successful search projects to present; everyone is working hard with delivering the basics. Why is that so?

Re: LDAP connector for Openpipeline

Tobias Larsson Hult — Tue, 23 Nov 2010 07:59:26 -0000

Hi Floris,

Yes this is similair to the People Connect feature. Though we feel our connector is easier to configure and allows you to have more control over what is being indexed. If you use Openpipeline to feed data into the GSA, you could use this as a source for People Search in the GSA.

Re: LDAP connector for Openpipeline

Floris Weegink — Tue, 23 Nov 2010 07:46:08 -0000

Hi Tobias, I am very interested in this feature!
Is it to similar the People Connect feature in GSA 6.8?

Best regards,
Floris Weegink, Google Enterprise Professional at VLC The Netherlands

Re: Apache Nutch making use of Open Pipeline

Kalle — Thu, 18 Nov 2010 16:17:40 -0000

Very interesting.. I would really like, at some point, to have even more deatils on how you run openpipeline in embedded mode and how it all integrates with Nutch.

I think it could be useful for other projects we have.

Re: Structure First or Structure Last?

Dan — Fri, 22 Oct 2010 01:07:44 -0000

I always find structuring first to be a nightmare. We have learned that you need to profile and analyze the data first and gain a good understanding of it before you start structuring. Of course, even after you think you've understood it, you will be wrong. But at least you won't have wasted the time up front designing the schema.

Re: Search and Business Intelligence?

Alexandra — Mon, 19 Jul 2010 22:26:53 -0000

We will move in this direction this autumn I hope. A first step would be to have Christian and Simon organise a brainstorm on how to provide log data from Solr in a form that can be imported into Spotfire and voilá - instant search analytics :)

Re: Search and Business Intelligence?

SDM — Fri, 16 Jul 2010 17:51:40 -0000

NO MORE BOXES

Thinking outside the box can often lead to development of just another box with its own unique set of built-in biases and limitations. Such is the case with search and business intelligence. How about throwing out ALL the boxes and starting over? Andy why only business intelligence? Isn't all intelligence potentially business intelligence? Why not connect ALL THE DOTS to reveal as yet undiscovered patterns that become the core of future concepts and theories leading to new opportunities? Enterprise cost savings is just the tip of the iceberg.

It seems to me the inescapable truth of such an undertaking would be to change the underlying fabric of the Internet Itself. Is anybody up to the task? For a glimpse into that future, see my comments at http://www.squidoo.com/Unst...