... large,¹

ILOG's, for one, intra- and/or extra- company...

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

... configurations.²

The first facet was the elaboration of Jacc, an advanced system for syntax-directed compiler generation [3]. The third facet will be the integration of logic-relational (from Logic Programming) and objet-relational (from Database Programming). A later facet may be to complete the design to enable both LIFE-technology [2] and CPS/LP technology to cohabit.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

... package!^1.1

Thank you Patrick Viry, Frédéric Paulin, Chritian Bracchi, and Chrisptophe Gefflot!... ;-)

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

... object);^3.1

Section 3.3.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

...-calculus);^3.2

Section 3.4.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

... stack;^3.3

Where stack here means ``stack of appropriate runtime sort;'' approppriate, that is, as per the instruction's sort--viz., INT, REAL, or runtime OBJECT.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

... type.^4.1

Or more precisely that $\forall$ never occurs nested inside a function type arrow $\rightarrow$. This apparently innocuous detail ensures decidability of type inference. BTW, the $2$nd order comes from the fact that the quantifier applies to type parameters (as opposed to $1$st order, if it had applied to value parameters). The universal comes from $\forall$, of course.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

... others.^4.2

At the moment, the handled goals are:

• typing goal: $e:T$ ;
• type unification goal: $T=T'$ ;
• base type unification goal: $T=\textit{base}(T')$ ;

Others are expected (and will!) be introduced, e.g., when we support subtyping constraints: $e <: T$, $T\leq T'$, etc., ...

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

... currying,^4.3

For the reader who might wonder what all this has to do with Indian cooking, it does not. It comes from Haskell B. Curry, one of the two mathematicians/logicians (along with ?. Feys) who made extensive use of isomorphism of Equation (4.1)--hence the misleadingly exotic [un]spicy metonymy.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

... names.^4.4

At the moment, there is no name qualification or namespace management. When this service is provided, it will also be through the ilog.language.design.types.Tables class.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

... compiler^B.1

Jacc is a java-based software extending yacc's familiar action-annotated (1) parsing, with several useful extensions. Jacc is the property of ilog but is not part of the software products sold and/or maintained by ilog. upon specific request, and on a per-case basis, compiled java classes (not sources) for Jacc may be made available on an ``as is'' basis if it is worth ilog'sand this author's time to do so.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.