C-Linda

C-Linda

(language)
The most widely used variant of Linda, with C as the base language. It is available from Sci Comp Assocs <linda@sca.com>.
Mentioned in ?
References in periodicals archive ?
The authors' claim that C-Linda results in code which is more readable than Parlog86 or Crystal is true, but that it is different from message-based systems is not true.
In principle, such tools could be built for C-Linda, Fortran-Linda, etc.
C-Linda, for example, relies on some sophisticated compiler optimizations based upon a global analysis of the restricted ways in which classes of tuples are used.
Shapiro has attempted to embed the coordination aspects of C-Linda within (the complete language) FCP; a comparable experiment would attempt to embed the coordination aspects of FCP within C-Linda.
They are comparable in length (in both cases initialization code is omitted, amounting to several lines in C-Linda and probably about the same for Parlog86).
A Parlog86 proponent would almost certainly call Parlog86 a "higher-level" language than C-Linda; but we've shown that it is somewhat easier to solve the client-server problem in C-Linda than in Parlog86, and much easier to solve the dining philosophers problem.
Although the two programs do not differ by much, the C-Linda version is slightly longer.
The C-Linda version isn't a set of equations; it is a fine grained explicitly parallel program.
The C-Linda version is longer, but the difference in compactness seems minor.
In principle, the kinds of analysis used in the Crystal compiler would probably work for a program like the C-Linda DNA comparison also, but we haven't investigated them and don't know for certain.
One efficient approach to the problem is easy to express in C-Linda (Figure 4).
In short, if the definition of a "higher level language" is a language that is relatively closer to the programmer's way of thinking, it is impossible for us to accept the claim that functional languages are necessarily higher-level than C-Linda.