Nicolas, Thanks for your reply. You bring up a good point: one the one hand, the relational model is king when it comes to the tools actually being used: odds are you've got a RDBMS on the back-end and SQLite embedded in your browser. But on the other hand, there are many people who don't want to necessarily think about, or interact with, their data as a typical RDBMS would normally require.
So we're trying to address this in two ways: - On the server-side, allow people to maintain their files are "ordinary" raw data files (CSV, JSON, MS Excel, etc) but have a process that is baking these into an optimized database for you so that you get query across all the files & optimized access. - On the client-side, create the ability to smartly know when to pull down and cache this data (to fuel an Exhibit, or whatever else you're running), but also to allow client-side developers to interact with the data in a comfortable way programmatically. For this we will definitely need to support query, and at the moment we've been toying with the idea of trying to stay as neutral on the issue as possible, possibly passing along an extra parameter specifying the langauge the query is in. This comes with some pros and cons, of course. Thanks for the link to Cubicweb -- we will definitely check it out! Best of luck finishing the SPARQL additions. Regards, Ted On Apr 2, 3:36 pm, Nicolas Chauvat <[email protected]> wrote: > On Thu, Apr 02, 2009 at 07:43:43AM -0400, Adam Marcus wrote: > > module/python script that will turn datasets into sqllite databases on > > the server side, so that if a dataset is too large to send across the > > network, all queries can hit the server. > > So you store information in a database on the server... > > > The step after that will be to allow exhibit to speak with this > > data-aware webserver to decide whether to run the query remotely or > > And provide a web service for your javascript client to fetch the > information it needs on the fly... > > am I understanding correctly? > > If that's what you are trying to do, I bet you will want your client > to send a query like "give me that part of the dataset", which means > you need some query language. Since you've probably heard about SQL > injection, you know sending SQL requests from the client to the server > is not your best option. Then SPARQL comes to mind... > > ... but wait, did I mentionhttp://www.cubicweb.org? > > We usually run it on top of larger SQL databases like Postgresql for > better performance and larger datasets, but all the automated tests > are run with sqlite. > > An example with Timeline would be sending a request like: > > Any X WHERE X publication_date >= "2008/01/01", > X publication_date <= "2008/06/30" > > and getting the result as JSON. Drag the focus in Timeline and you can > send a new request with a different date. MVC in Timeline, I think I > read it is almost there. > > Or maybe with Exhibit: > > Any X,T,D WHERE X name LIKE "a%", T tags X, X pub_date D > > and getting the result as JSON, to display the X with names starting > with letter 'a' and using tags and dates as facets. > > These examples are with the RQL query language. SPARQL will be > available within a couple months. > > Licence is LGPL, book being written athttp://www.cubicweb.org/doc/en/ > > If I understood correctly and that's the kind of thing you were > thinking of, please be our guest :) > > -- > Nicolas Chauvat > > logilab.fr - services en informatique scientifique et gestion de > connaissances --~--~---------~--~----~------------~-------~--~----~ You received this message because you are subscribed to the Google Groups "SIMILE Widgets" group. To post to this group, send email to [email protected] To unsubscribe from this group, send email to [email protected] For more options, visit this group at http://groups.google.com/group/simile-widgets?hl=en -~----------~----~----~----~------~----~------~--~---
