Cheops and its following incarnations were an occasion for me to develop my ideas on software engineering. This worl led to an elaborate architecture for complex informations visualisation, which we called Giza. I wrote a fairly long article [GizaArticle.pdf] to explain my approach, but it was in the somewhat shortened form of a "Tech Note [TechNote.html]" that it was finally presented at OOPSLA [http://oopsla.acm.org/], with a demonstration [oopsla_demo.html]. I have also presented Giza at Université de Montréal, in the software engineering group seminars ( GELO [http://www-labs.iro.umontreal.ca/~labgelo/previous-site/menu.html] ), at Ruedi Keller [http://www-labs.iro.umontreal.ca/~keller/]'s kind invitation.
In one word, Giza's architecture allowed to build custom facades on-the-fly, provided access to data went through accessor objects, which could be combined at need.
Today, I woud say that the approach was over-architectured; but those were the times, and it corresonded to our attempts to develop a generic library before we had clients.
Giza, like Cheops, was not developed further after we left CRIM, but I was quote pleased to see the software engineering ideas similar to those I had sketched in Giza brought to fruition in others projects, such as in HyperJ [http://www.research.ibm.com/hyperspace/HyperJ/HyperJ.htm] from IBM, or in a different form within PyProtocols [http://peak.telecommunity.com/PyProtocols.html]. Louis Vroomen had also pushed those ideas further on his own.