The single hardest problem with observing CQS is knowing which "functions" [1] you intend to call are themselves observing CQS. Ideally, the compiler would flag an error when encountering them, so you don't have to. Since this isn't happening yet, I propose to write a (conservative) CQS checker, to warn the author of a function.

Sometimes I can be incredibly slow on the uptake.

It was about 5 years ago that I first started using agents. Classes such as ARRAY all had a feature named do_all.

Void safety clashes with DbC (it forces us to indulge in defensive programming).

On November 26th the W3C brought disgrace upon themselves by publishing an eyes-wide-open deliberate lie - the so-called "Fifth Edition" of XML 1.0.

This is actually a new version of XML, masquerading as 1.0. Most other W3C recommendations that refer to XML 1.0 are now automatically non-conforming, including XML Namespaces!

Eiffel is missing from this list. Anyone fancy adding it on?

Following on from my comment in my post about class fibonacci that Haskell manages integers better than we do, I extended my test program for Fibonacci numbers to make a call to a Haskell function.

Timings and output printing F(11199) (compiled with ISE 6.3) are:

Class FIBONACCI (FreeELKS) - 2ms prints: -1880045342

Class MA_DECIMAL - 22

I was somewhat surprised to see this class in FreeELKS.

WHY? It doesn't appear to be used by anything else.

Anyway, I was rather intrigued, and gave it a try (I was interested in timing the n_th number).

I soon spotted that it was giving errors, and it was immediately obvious why. Here is a copy of the code:

	i_th (i: INTEGER_32): INTEGER_32

This week I have again been considering the issue of the correctness of code that receives an agent (the last time I looked at this was page:correctness_conditions_2_for_calling_an_agent).

I return to the original class which provoked my initial ponderings on this - Karine Arnout's (now Karine Bezault) `FACTORY'</ei

Mark Howard, director of software engineering at AXA Rosenberg, describes why we chose Eiffel for our expert system.

I'm just reading Yhc/Erlang/Proof of concept.

Perhaps some enterprising young student would take a stab at compiling Eiffel to BEAM? Then we'd get real concurrency.

Not a trivial task.