SV-BC Separate Compilation discussion 7/15/03 Attendees 000000 876654 012012 553928 000000 333333 aaaaaa Karen Pieper --aaaa Randy Misustin a-a--a Dennis Brophy -----a Danny Jacobi aaaaaa Brad Pierce -aaa-a Dave Rich ----aa Francoise Martinolle -a-aa- Matt Maidment -a-a-- Tom Kiley --a--- Johny Srouji aaa--- Jay Lawrence -a---- Arturo Salz -a---- Peter Flake Karen discussed several other alternatives to $root, including turning $root into an implicit module which could then be imported to other $roots. The concensus we reached was that that would make a bigger mess rather than clean things up, so we will focus our initial efforts on just providing namespaces as an alternative to $root. We acknowledge that this will leave the problems that exist with $root, but will provide a clean path for tools that choose to implement separate compilation to have a subset they can separately compile. We did discuss that namespaces are different from modules because they contain global information. If we instantiated them with parameterization, then we would create a situation where we could have use before definition, a situation we would like to avoid, so we should not instantiate namespaces, their definition should be their instantiation. What about the extern proposal? Does that solve the problem as well? We will plan to address the extern proposal after we have agreed to the namespaces proposal. A few issues remain in the namespaces proposal. We need locations in the manual to add the text. Need to clarify compilation order for namespaces. Does reparsing require recompilation of everything? What about variable initialization? Are initial blocks required? Namespaces already has a definition in 1364, how about using the term package instead? Jay agreed to check his design suite to look for the word package as an object name. Karen will try to update the proposal reflecting the issues mentioned above and get Randy's buy-in. Once that is accomplished, we will hold another meeting on the following Tuesday morning.