[phylum]<CTamarin>Doc>Memo>Tamarin-Future.bravo!1

The Tamarin project is in a period of crisis. I am unable to get the software help I need on the Tamarin project. Lately, I have been very unproductive and have made almost no progress on Tamarin in the last two months. The completion date has slipped substantially.

At the beginning of the Tamarin project, it was clear that a large amount of software work would be required. I attempted to structure the project to delay the amount of software work as long as possible. The intent was to allow AISBU to obtain additional staff for software. This new staff would provide trained Lisp talent to work on Tamarin and would also offload the "super-experts".

This staffing has not happened!!! A year ago, I expected that AISBU would have adaquate staffing for all its projects by now. In reality, almost no software staffing has occurred. This is already hurting Tamarin and will hurt much more in the future.

I have outlined the software tasks required for Tamarin in a separate and enclosed memo. My urgent needs are for help in the "Software Definition" phase. I feel that Bill van Melle is the only current staff capable and willing to perform this task.

A couple of months ago, I formally asked for some of Bill van Melle’s time. You indicated to both Johan and me that you had arranged for Bill to spend 25% of his time on Tamarin. Since that time, I have received essentially zero help from Bill. When I specifically asked Bill about this 25% number, he explained that he has many other responsibilites, particularly CommonLisp, on which he feels that AISBU puts a higher priority. In talking to Bill, I got the expectation that he will be working full time on CommonLisp for at least the next year.

I cannot do any more work on the Tamarin processor chip without help on the "software definition" phase. I believe three options exist at this point: continue working on the processor chip using 50% of Bill van Melle’s time, working on the I/O chip, or suspending the whole project.

The first option would keep Tamarin on the fastest completion path. At least 2 solid man-months of Bill’s time would have to be allocated to Tamarin.

Working on the I/O chip would allow me to continue to be productive. However, it would be detrimental to the overall schedule by about six months. Completing the processor chip would allow software development to proceed in parallel with my work on the I/O chip. Working on the I/O chip first would delay the processor chip and the start of the major modifications to Lisp.

The responsibility for prioritizing tasks belongs with AISBU. It would be inappropriate for me to say what is best for your markets. However, you have to realize that delay in providing software help will delay the completion of Tamarin.

Solving the problem of the System Definition phase would keep things going. However, much more software work will be required beyond this first phase. AISBU simply does not have the staff required to port Lisp to new hardware, be it Tamarin or some high-end processor. The software tasks for Tamarin need trained people, not just slots for people.

My role in the Tamarin project has been one of hardware design. I have avoided performing management functions for the project. Performing project management would prevent me from having time to work on hardware design. Secondly, I am not a part of AISBU and do not have the ability to provide direct input on priorities and directions.

Yet the Tamarin project needs someone to perform the overall planning and resource procurement working for AISBU. This has to be performed by someone in AISBU, most likey in product planning. This project needs to be "championed" by someone who wants to see it happen. This person needs to look after the medium and long term future of the project. He needs to be able to obtain the resources required.

In conclusion, I need guidance on how to proceed in the short term, i.e. which of the three options to chose. Secondly, I need a clear indication of whether Tamarin still fits within AISBU needs. Thirdly, the project needs a mechanism for planning the medium and long term.