If newLISP meets BI

[csfreebird]

Hi, my current job is about big data, we need to create some tools to help other coworkers to search and analyse their business data.

One approach is to teach them to use SQL, we need to provide a standard SQL which encapsulates the differences of Hive, MySQL, Oracle and PostgreSQL. Most of these people are not engineers.

Is it possible to develop DSL for this purpose based on newLISP instead of SQL. Which way is better for my coworkers?

Do you have some experience about this?



I can develop a web site with newLISP Dragonfly and some web front-end technique.

[TedWalther]

My experience is, yes.  I've used newLISP for a lot of data mining, transfer, and recovery.  It is so easy and convenient.  Send me a private message if you'd like more details, most of my gigs are confidential.

[csfreebird]

Thanks for your reply. Let's just talk the programming technique here.

Do you create DSL for your job? Could you give me a simple example?

[TedWalther]

DSL are simple to make, but maybe not the right thing for your application.  I estimate 2 man-months to make a (safe) version of SQL that encapsulates all those versions of SQL you mentioned.  Probably 2 days to have something usable, 2 weeks to have something workable, 2 months to stomp on it and make it really bug-free.



Are there any particular queries or types of queries your users are likely to need?  You could simplify those cases in your web-UI.