Let's' dream!

[agershun]

Sometime I am thinking about a purpose of library and imagine how it can be used for intelligent software in the future.

Here we can have two direction of development: a) very pragmatic - the target is mobile + reliable backend b) less pragmatic - use for AI tasks, for example combine with libraries like this

[mathiasrw]

I love dreaming !

Are you dreming of having AI in alasql or having a vision for how alasql can be implemented into AI projects?

I dream of working on the to-go-to solution for javascript DB - and have a feeling that a modular setup for query generation combined with SQL compatible flavours and modular connection to different backends and a strong test disciplin is the way to go :)

[wuxianliang]

even more

1. Probabilistic programming for the web http://webppl.org
2. A Bayesian database table for querying the probable implications of data as easily as SQL databases query the data itself http://probcomp.csail.mit.edu/bayesdb/

[agershun]

Prolog. I thought about realize Prolog queies along with graph queries.

For example we can think about nodes and edges in the terms of atoms and terms:

CREATE GRAPH John > loves > Mary;
// internaly equivalent to
:- loves(John, Mary);

So we can search over graph:

SEARCH @x> loves > Mary;
?- loves(@x,Mary);

[wuxianliang]

In my understanding today, this kind of query only tells me what I input ever, if so AlaSQL is a database. But based on a lot of json which mention John in AlaSQL, AlaSQL can tell me John is the most probable boy who loves Mary which is not inputed by me. It is AI. Probabilistic programming make sense here.

During revolution of industry, steam engine is not most smart part but transmission device. If AlaSQL is database only, it is a toy. If AlaSQL glues all kinds of libraries in front end together by most simple way to handle all types of data, it is a standard.

Never mind I am dreaming.

[agershun]

@wuxianliang Inspirational! Thank you!

[mathiasrw]

:+1: for gluing

[agershun]

I like this BQL (Bayesian Query Language) http://probcomp.csail.mit.edu/bayesdb/docs/0.2/bql.html. I already worked with Bayesian probabilities and expert system.

[agershun]

Predictions on JavaScript in the next 12 months

Content (I marked with bold the items which affect on AlaSQL):

1. Fewer people will write JavaScript without a compilation step
2. Smaller libraries (and the composing of) will be preferred over large frameworks-
3. Focus on libraries that do one thing and one thing well
4. Large, fully-featured frameworks will remain rightly popular
5. The use of compilers (Babel etc.) will be abstracted for us by generic build tools
6. Running the same JavaScript client and server side will be common
7. As ES2015 implementations complete, we'll be writing ES7 already
8. The rate of new frameworks will begin to slow down

[mathiasrw]

Regarding 4. What in your vision the one thing alasql does that it does well?

[agershun]

... WebSQL compatible replacer on top of IndexedDB working in Corodva

[mathiasrw]

:+1:

[agershun]

So, let's do it :)

Some idea of our next plan:

After we finish with:

• README
• Starting page + Codex
• ; and similar issues
• I have to finish with PIVOT

We can start to prepare to release 0.2, before that we will:

• move all unnecessary parts to plugins
• reduce the number of keywords
• release VALUE keyword
• remove pretty() and console() functions to plugins

Then in version 0.2 we can continue to fight for:

• Work in Cordiva
• Effective work with IndexedDB
• Improve SQLTEST and NIST SQL tests
• Reduce the size of the library
• Add more Postgres/MySQL compatibility
• Create OriendDB plugin

[mathiasrw]

I vote for this plan !

[gregaubert]

:+1: