Moose on the web - interaction
Locked
 
|
Some thoughts on the interaction with a Moose server retained from the discussion with Anne and Vincent (it was also Vincent in my previous mail but i misspelt :-).
With the current implementation, one can start by interacting with a model at any level (allClasses, allMethods, etc.) and then narrow his search on an entity or a set of entities. Every entity returns in the results the list of msgs it implements. We are working on how to correctly choose a specific moose entity from the returned collection (see my previous mail). Scoping information is essential to correctly identify/find an entity. So scoping information can be retained by the way user interacts with the server. For example: But that would mean interaction is restrictive, one has to start with the highest level of container (allPackages in java) and drill down to search for the required entity. Not retained. Nicolas P. suggested to keep a way to search entities by their mooseIDs as well so we'll implement it as well. Specify the number of entities to return.
Be able to specify an element by its position in the collection returned. What do you think? Usman _______________________________________________ Moose-dev mailing list [hidden email] https://www.iam.unibe.ch/mailman/listinfo/moose-dev |
Re: Moose on the web - interaction
|
|
I see space to investigate the use of a graph database such as http://www.neo4j.org in order to store the model. The Moose model is "just" a graph. Did you investigate it? On Wed, Jun 5, 2013 at 4:40 PM, Usman Bhatti <[hidden email]> wrote:
Andre Hora _______________________________________________ Moose-dev mailing list [hidden email] https://www.iam.unibe.ch/mailman/listinfo/moose-dev |
|
|
That looks cool. btw, at 8 minutes [1] shows an interesting comparison
different NOSQL databases that I hadn't seen before.
[1] http://vimeo.com/67244674 cheers -ben Andre Hora wrote: I see space to investigate the use of a graph database such as http://www.neo4j.org in order to store the model. The Moose model is "just" a graph. Did you investigate it? On Wed, Jun 5, 2013 at 4:40 PM, Usman Bhatti [hidden email] wrote:Some thoughts on the interaction with a Moose server retained from the discussion with Anne and Vincent (it was also Vincent in my previous mail but i misspelt :-). With the current implementation, one can start by interacting with a model at any level (allClasses, allMethods, etc.) and then narrow his search on an entity or a set of entities. Every entity returns in the results the list of msgs it implements. We are working on how to correctly choose a specific moose entity from the returned collection (see my previous mail). Scoping information is essential to correctly identify/find an entity. So scoping information can be retained by the way user interacts with the server. For example: http://mooseserver/allPackage/first/allClasses/first/allMethod/first But that would mean interaction is restrictive, one has to start with the highest level of container (allPackages in java) and drill down to search for the required entity. Not retained. Nicolas P. suggested to keep a way to search entities by their mooseIDs as well so we'll implement it as well. Specify the number of entities to return. Be able to specify an element by its position in the collection returned. http://mooseserver/allPackage/3/allClasses What do you think? Usman _______________________________________________ Moose-dev mailing list [hidden email] https://www.iam.unibe.ch/mailman/listinfo/moose-dev _______________________________________________ Moose-dev mailing list [hidden email] https://www.iam.unibe.ch/mailman/listinfo/moose-dev |
Re: Moose on the web - interaction
|
|
In reply to this post by Usman Bhatti
usman
We should discuss with olivier too. We could have <a href="http://mooseserver/allPackage = named: #Graphics/class=named: Point/">http://mooseserver/allPackage = named: #Graphics/class=named: Point/ or <a href="http://mooseserver/allPackage = named: #Graphics/class=named: Point/">http://mooseserver/get?Graphics::Point::x This idea of first is totally bogus and we should not use it. In Moose we got uniqueNames for all the entities based on their structural relationship Graphics::Point::x denotes a unique entity. Stef On Jun 5, 2013, at 4:40 PM, Usman Bhatti <[hidden email]> wrote:
_______________________________________________ Moose-dev mailing list [hidden email] https://www.iam.unibe.ch/mailman/listinfo/moose-dev |
Re: Moose on the web - interaction
|
|
On Wed, Jun 5, 2013 at 9:33 PM, Stéphane Ducasse <[hidden email]> wrote:
Yes.
ok.
Yes. The idea of searching in a group of entities rather than a complete model is interesting. I think I'll spend a day tomorrow working with Vincent to construct different navigation scenarios.
_______________________________________________ Moose-dev mailing list [hidden email] https://www.iam.unibe.ch/mailman/listinfo/moose-dev |
Re: Moose on the web - interaction
|
|
In reply to this post by Usman Bhatti
Andre Hora wrote:
>I see space to investigate the use of a graph database such as http://www.neo4j.org in order to store the model. The >Moose model is "just" a graph. What do you hope to achieve by using a graph database? From my POV it just looks like a very inefficient java based image. 64 bit though. Stephan _______________________________________________ Moose-dev mailing list [hidden email] https://www.iam.unibe.ch/mailman/listinfo/moose-dev |
«
Return to Moose
|
1 view|%1 views
| Free forum by Nabble | Edit this page |