135.text

From: Geoff Croes
Subject: aips++ management
Date: Tue, 28 Apr 92 16:35:54 EDT



	Organisation of the aips++ project management

	It is my pleasure to announce the appointment of
Gareth Hunt to the position of aips++ project manager.
With the indefinite involvement of Bob Hjellming as the
project astronomer and Brian Glendenning as the project
computing scientist, the triad of senior aips++ staff
that will determine the face of aips++ in the years ahead
has now been determined. Although the administrative centre
of the aips++ project will be in Charlottesville, actual
work done for the centre will be shared between Charlottes-
ville and Socorro, as appropriate and as resources permit.

	The three senior staff members will act in close
consultation on all matters related to the conduct of
the project. Each one, however, has his own specific
authorities and responsibilities. These are detailed below.

			Project Manager

	The prime responsibility of the project manager is
the administration of the project. He will therefore:

- guide and monitor progress of the aips++ project, both at
	the centre and at the consortium sites, through
	the establishment of milestones, time lines, etc.;

- coordinates development and documentation activities between
	the centre and the consortium sites, including the
	establishment of "software contracts";

- write monthly progress reports for the aips++ steering
	committee, weekly bulletins for more general
	consumption and a summary progress report for each
	committee meeting;

- establishes manpower and money budgets for the aips++ centre,
	and administers these once they have been allocated;

- with the Chairman of the aips++ committee, establishes the
	agenda for each steering committee meeting, including
	the procurement of all necessary papers, editing of
	minutes, etc.;

- organizes the logistic support for staff at the centre;

- attends and actively participates in all technical meetings
	of the aips++ centre, in particular those dealing with
	analysis and design issues.

	In order to fulfill this task, the project manager
has the authority to:

- assign centre staff to various parts of the project;

- establish "software contracts" between the centre and the
	consortium members;

- disburs any monies allocated to the aips++ project.



			Project Astronomer.

	The prime responsibility of the project astronomer is
to ensure the astronomical correctness of the aips++ product.
To fulfill that task he needs to:

- take full responsibility for the establishment of requirements,
	including consultation with the consortium partners and
	the astronomical community at large;

- spearhead the object oriented analysis of these requirements, so
	that they are guaranteed to be accurate, complete and in
	a form suitable for object oriented design;

- participate in the object oriented design of the product;

- ensure that suitable astronomers work with the project whenever
	that is indicated;

- ensure that proper end-user tests are done to ensure that the
	astronomical content and form of the end-product is
	appropriate;

- ensure that all end-user oriented documentation (e.g. cook books,
	reference manuals, help systems) is written and that it
	is correct.



			Project Computer Scientist.

	The prime responsibility of the project computer scientist
is to ensure that the implementation of the aips++ project is
optimal. To fulfuill this task he needs to:

- take full responsibility for the impementation of aips++;

- spearhead the object oriented design of aips++, so that trade-offs
	are properly chosen, and easy maintenance and extensibility
	are promoted;

- participate in the object oriented analysis of the requirements;

- ensure that appropriate computer scientists participte in the
	project whenver indicated;

- ensure the formal correctness of the end-product;

- ensure that all system-oriented documentation (e.g. class
	documentation, programming guidelines, etc.) are written
	end correct.



			Conclusions.

	Although, as I have said before, a major requirement is
that the senior people work closely together on all major
decisions, each of them has a prime area of authority:

	the project astronomer decides WHAT needs to be done;
	the project computer scientist decides HOW it is done, and
	the project manager decides by WHOM and WHEN it is done.

	If, in the course of time, some irreconcilable differences
of opinion develop (a quite unlikely occurrence), the points at
issue will be conveyed to the aips++ steering committee for arbitration.


				Geoff Croes