Comments on: Unified Language Specification

By: Ben Gillis

Ben Gillis — Sat, 18 Apr 2009 22:24:27 +0000

The discussion and examples are nice, but is there any reason why this can’t include a complete grammar in ABNF?

By: douglasp

douglasp — Thu, 16 Apr 2009 21:22:06 +0000

M is the DSL for data.

Data == values, constraints, languages/transformations

SQL is one of _many_ possible projections of this data.

CLR is another.

XML is another.

Net, an M constraint, which is all an MSchema is, could be projected and enforced in a lot of different places other than SQL.

Kevin: Have you watch the “What is M?” part of our MIX talk? I think that does a reasonable job of framing this story in terms of the current bits. We are spending a lot of time on the CLR runtime and projection these days and also providing something more productive for the SQL runtime and projection.

By: Kevin Dente

Kevin Dente — Thu, 16 Apr 2009 17:49:57 +0000

Got to say I’m with Shawn on this one. Generally I think the excessive focus on MSchema in most Oslo talks I’ve seen only muddies the water about the value of M.

By: Shawn Wildermuth

Shawn Wildermuth — Thu, 16 Apr 2009 10:47:29 +0000

Also, I understand the desire for MSchema/MGraph being one language, but I don’t get the One “M”…isn’t MSchema just a DSL for data? MGraph and perhaps MGrammar make sense, but MSchema seems outside the box there. I don’t know why MSchema would be special…

I *understand* that the Repository/Data story is important and that MSchema fits into that, but at some point it feels that MSchema is a SQL Gen tool and just a DSL.

Your thoughts?

By: Shawn Wildermuth

Shawn Wildermuth — Thu, 16 Apr 2009 10:44:45 +0000

Just a typo…I think in the first para, you mean “MGraph and MSchema” not “MSchema and MSchema”