Hello there, rarely-heard-from search_dev group. I'm pondering no-results and low-results pages for online library catalogs. For the product I'm working with,...
I was looking into this yesterday. Amazon shows some recommendations, things I¹ve looked at recently. IMDB suggests that I change my search options. Ultraseek...
Some search engines also let you change the default "AND" to an "OR", if they submitted multiple words. Or open up the search a bit by turning on stemming,...
Back in the day (1998 - the now defunct CDNow) we implemented an iterative approach. Like Mark suggests, you want to start with a pretty strict search, then...
Dear Group, I'm looking for opinions about which search engine vendor has the best Taxonomy story. I know both K2 and IDOL intimately, so I know what they...
Often we will show a list of popular search terms on the no results page. For example: http://halloween-costumes.spirithalloween.com/search?view=grid&w=ipod ...
Some libraries now embed chat widgets that pop up for live "ask a librarian" response. We won't be doing that. and we won't be doing the instructions on ...
I've seen 2 distinct meanings of ngrams. One character based, as you suggest, and another word/token based. Verity had a CHARACTER based ngram? I had thought...
Right. The word/token-grams are also sometimes referred to as "shingles". Otis -- Sematext -- http://sematext.com/ -- Lucene - Solr - Nutch ... From: Mark...
I think that Verity K2/VDK did support n-grams, though it probably wasn¹t used much. Verity K2 had positions in the index, so it didn¹t need shingles. Also,...
Hi all, I am importing XML documents and some of the fields have HTML characters in them. For example, one of my fields is <article-title>Stoichiometric...
I saw that the IDOL's Import Module has some kind of support for legacy K2 "BIF" files. The documentation mentions two parameters: ImportBIFReferenceField and...
... legacy K2 ... Erkan: Don't have a full answer for you but we did an article a couple of months back that talked about how you can convert BIFs info IDOL...
Erkan, It is a known bug in latest version of import module which ships with 7.0 versions. Autonomy is aware of the bug and may fix in next release of import ...
SearchDev.org Dinner at Enterprise Search Summit Exalead and New Idea Engineering invite you to join us at a dinner for attendees of the Enterprise Search...
There was a discussion over in SearchCoP about evaluating search engines with A/B testing. I finally got around to writing up my comments on that. See my post...
we are running into a problem with the Import Module not converting unicode characters. It does convert some characters but far too many are left as is. I am...
Is this XML? XML only does numeric entities and five pre-defined character entities. If you need character entities, you need to define them in the file. See...
Yes, this is XML. Which file do you refer to where I should define the entities? I don't have control over the XML itself. Is there an Autonomy convtable file...
First, this is not legal XML. You should be able to prove that by saving it in a file that ends in ³.xml² and opening it with a browser. It will give a parse...
I am running AutonomyIDOLServer v7.3.9.0.129198 in Linux environment. I can't seem to get transliteration to work. Below is my Language Types section... ...
I have to import XML documents that contain ISO 8879 entities. IDOL only supports ISO 8859 (according to documentation) so many entities do not get converted....
ISO 8879 is SGML. XML is a massively-simplified subset of SGML. No XML parser can parse all of SGML. Even SGML parsers can have a hard time with all of SGML. ...
I'm also having issues with IS0 8879 and UTF 8... this character shows up and when I finally get rid of it...it shows up as 0xA0...Any ideas? Thanks, Kathy ...
New information. Looks like although Autonomy documentation says the PollingMethod parameter can appear in a [job], it cant. When I have [job]s with differing...
Hello, Autonomy suggested PollingMethod= 2 to index all files in certain directory with predefined type. For PollingMethod= 1 is to perform the above job based...
