Unfortunately WPF sample does not work
1) Please, include SampleApp1.2.zip to the downloading package on github
2) When you create task (FileUpdateTask) using AppcastReader, task’s LocalPath field is not initialized.
Then when you do the preparation of FileUpdateTask you check if LocalPath is null and if it is you return. In this case it is always null and it should not be like that

“What’s the point if you still need to build out your objects?” – Don’t you have to have a c# model in Raven as well?

“Just so you don’t have to worry about your ORM” – If you try to see Siso as a replacement of an ORM you are looking at it the wrong way since it’s about how to look at the data being modeled. It’s not intended to replace ORMs since you will have to change your mental model when using documents, just as with Raven. E.g producing view specific documents duplicating data, where the latter would give a traditional RDMS ORM user, cold feats.

Also there have been a lot of changes targeting some of the “issues” in this post: http://sisodb.com/wiki/release-notes

//Daniel

I’m planning to run a knowledge base site and I want to use HebMorph.
How can we discuss about the details for comercial use?

Thanks,

Tomer

Also feel free to email me and I would tell more how we are using Hebmorph with Lucene.net

Anyway i will download the src and compile them against the new lucene.net version

Out of curiosity, how do you use HebMorph? can you please email me with the details?

I just downloaded the new 2.9.4 lucene.net, and now my project fail to build becuase lucene.net.analysis is missing the 2.9.2 reference.

Are you plann to make a fix or i need to download the src and compile the solution again?

The current version of HebMorph already shows very good IR performance even though it doesn’t use any advanced morphological or statistical tooling. As I mentioned in the articles, the super-advanced tools you referred to still miss a lot and the cost of perfecting and maintaining them is becoming insane because of that. And they will still never be perfect.

We still have a lot of work to do, but I believe we can find a method that is working very well and costs quite little to build, and this is exactly what HebMorph is here for.

I am a great believer in N-grams but not for hebrew. This is because hebrew has greater ambiguity than arabic both due to general lack of vowel and spelling variations. This spells an advantage to context savvy morphological analyser over algorithmic ones.

You have stumbled on another well-known fact that there is a unique challenges when analyzing queries.
However a well-engineered IR system can “cheat” to overcome these. By performing periodic analysis of historical queries and the user’s result choice user) it is possible query time analyser which will lear to adjust its rank based on the user’s expectation.