MADEFAST Process

MADEFAST Process Decisions

Early Organization

Early attempts at process management included time-lines, task/resource allocations, ect. However, recording decisions as they happened, or later rationalizing them, turned out to be more appropriate to the way the process actually occurred.

Process Decisions

This is a highly informal representation of some of the PROCESS decisions being made in the MADEFAST project.

21 July 94

Results of the meeting on 21 July 94.
We are planning a short (10 minute) demonstration for the ISAT meeting in Woods Hole on or about the 25th. We decided to pick 2 examples that illustrate various facets of the Madefast capabilities, now and future.

Mark wants to redesign the seeker using a single mirror - he uses the Madefast documentation to retrieve the relevant design decisions. Then he contacts Sam with the collaboration services. Sam says that he has indeed thought further about this possibility and shows Mark some design ideas. They collaborate on a redesign. Glenn Kramer, Charles Petrie, and Carolyn Valiquette will work on this scenario.
One approach will be to use ABSML (Glenn Kramer) tags that correspond to the Dedal model (Vinod Baya). We will also investigate using DME (Yumi Iwasaki).
We then bring in the MD*/SLA or MSU composite services and show how they could be used to prototype parts of the new redesign. ACaPs is quickly shown to be an example of this use of the NII for rapid prototyping.

5 July 94

This is the state of the process as of the meeting of 5 July.

Action Items

Utah (Sam Drake) will machine the mirrors but Stanford (Mark Cutkosky) will attempt to have mirrors fabricated by Sandia Labs at Livermore also. Utah (Carolyn Valiquette) will also attempt to have MSU fabricate the casing with composite materials and CMU fabricate the detector shroud using SLA.

Larry Pfeffer is in charge of the control systems and is conducting tests.

George Toye will make a sketch of the current two mirror design option for posting to the web.

Jayachandra Reddy will work with Charles Petrie on the design and process documentation (such as this), which is all part of the process.

Notes

The 5 July participants will continue to work on the August presentation. Larry Leifer captured rough notes in PowerPoint and distributed them in binhexed email. There needs to be an alternative to requiring all of the participants to acquire a Mac to share the slides. A ServiceMail application may be in order. However, in the meantime, Charles Petrie has acquired and is learning to use a Mac in order to be able to read the binhexed PowerPoint mail. There is no common repository yet for these working notes.

Note from Carolyn Valiquette: When collaborating over the net, team members are not able to walk into their project leader's office and talk him/her into trying a new idea. At a late stage in a project, such an idea would have to be very well presented and justified. This is easier to do in person than via e-mail. The cultural effect is that team members who want their ideas to be considered will have to become very good at presenting those ideas clearly and concisely using the net.

21 June 94

This is the state of the process as of the meeting of 21 June.

The Madefast prototype demonstration and presentation will likely be at an ARPA meeting called "ISAT" at Woods Hole on August 25/26.

The Madefast scenario has four parts:

Services such as making models and animating them. (This may be a Utah service - Sam Drake is evaluating.)
Directory of Services, such as the fledgling DesignNet.
Shared Repository of design and catalog information such as the Madefast web.
Actual Design and Prototype to be built, including the design documentation.
Technology Infrastructure such as the tools being developed by EIT, the various Madefast services that exist now, those services yet to be developed, and the shared experience of the Madefast participants in collaborating. See the discussion in the meeting of June 21.

The general goal is to develop a presentation that shows results in each of these areas. One of the most important is to develop a directory of useful services, many of which may not exist by the end of this experiment but which point toward needed development. What are the services we wish we had had?

We also need to say what tools we've tried and what tools we think future Madefast-like collaborators will need. Design documentation tools will be a major item.

We must address the collaborative incentives and social interactions as well as the infrastructure - lessons learned. What are the barriers to collaboration?

Organization

One result of the 21 Jun 94 meeting was to organize as follows for the remaining two months of Madefast:

General Manager - Mark Cutkosky
Design Manager - Sam Drake
Enabling Technology Manager - Glenn Kramer
Process Manager - Charles Petrie
Assistant to Mark - Carolyn Valiquette

Early in the project, there was an attempt to provide a flowchart and distribution of tasks. This being a completely distributed and non-hierarchically managed experiment, this ordering was not followed.

Charles Petrie