Text preview for : 19770916_Preliminary_Pilot_Development_Work_Plan.pdf part of xerox 19770916_Preliminary_Pilot_Development_Work_Plan xerox sdd memos_1977 19770916_Preliminary_Pilot_Development_Work_Plan.pdf
Back to : 19770916_Preliminary_Pilo | Home
XEROX
INFORMATION PRODUcrs GROUP
System Development Division
June 7, 1977
revised June 10, 1977
revised July 8, 1977
revised July 21, 1977
revised July 26, 1977
revised August 5, 1977
revised September 16, 1977
DRAFT - DRAFT - DRAFT
To: file
From: W. C. Lynch
Location: Palo Alto XEROX SDD ARCHIVES
d and understood
Subject Preliminary Pilot Development Work Plan 1 have rea
Keywords: Pilot, work plan pages_ - - - - -TO. - - - - -
Date.___-
File: (Lynch)PilotWp.Memo or (Lynch)PilotWp.Ears Reviewer-----;-ef 0.1 15 b D.- 3,2 6 .
f of Pages,_ _-
Archive document #: Lopez archive number
Introduction
This workplan presents the plan for the creation and delivery of Pilot. in detail, through
release 2.0 on May 31, 1978. This release will both be the first release for the DO' and
contain the function required for the Janus IMO release. This workplan does not cover the
period beyond the release of Pilot 2.0 in nearly as much detail. Pilot 2.0 will provide almost
all of the functions caJ1ed for in the Pilot Functional Specifications but will not necessarily
have either the performance or the main memory consumption required for product release.
Common Software (anything not covered in the Pilot Functional Specs) is not covered by
this workplan. System integration activities (post release support, performance
improvements, and memory requirement reductions) beyond the release of Pilot 2.0 are not
broken down into detail.
Very little in the way of post release support for Pilot 1.0 (Pilot on the Alto) is
contemplated and the mechanisms for post release support of system software products in
general have not yet been worked out When they are, additional tasks will undoubtedly be
identified.
OIS Mesa, as represented by the OIS Mesa Fuctional Specification, will disappear as an
. identifyable project, as wiIJ that document The functions represented are being divided
between the Pilot project and the Mesa project and the portion subsummed by Pilot is
covered by this workplan. The remainder of the functions are to be covered by the Mesa
workplans.
Background data
productivity assumptions - document preparation
Prelim Pilot Devp Work Plan - revised September 16. 1977 - DRAFT 2
This section will list the productivity assumptions upon which this workplan is based. The
background material that was used to arrive at these assumptions is contained in
productivity rates achieved. both to enlarge our data base of productivity information and to
give early warning of potential schedule deviations.
The construction of written documentation is a very significant component of this project
We therefore need an estimate of the productivity of producing this documentation. I
estimate a production rate of 4 pages/day if the material is already thought out and no
background documents are required (RawPage). Otherwise the rate is .5 pages/day
(FinishedPage). The thinking through of the material seems to require 3 SupportPages per
FinishedPage. Each SupportPage is treated as a RawPage
Each RawPage requires .67 hours of Alto usage (AltoHour). We therefore estimate 4.5
AltoHours/FinishedPage. Alto usage for document construction will be based on this figure.
document size distribution
In order to estimate the effort required to produce the necessary documents. we need to
estimate the size of the documents. To this end I estimate that
The design spec is three. times the size of the functional spec
The test spec is 12 pages
The technical documentation is twice the size of the functional spec
productivity of coding and unit test
Again from unit test 800 lines of Mesa source code (loc) per man month (mm.). at the well established
conversion figure of 2.5 bytes of object code per loc we obtain a productivity of 2000
bytes/mm. These figures are also consistant with the estimate given in the SDD Software
Development Procedures and Standards (P+S) which claims that coding and unit test
represent 15% of the total effort and that each person will produce about 1500 loc per man
year overall.
From the data presented in Productivity.memo I estimate that 0.16 AltoHours per loc are
required and this plan is based on that figure.
Alto usage is not likely to be more than 50% effective. I will therefore assume that one
Alto will yeild 100 hrs. of usage per month.
program size distribution
In order to estimate the programming effort we need to estimate the size of the individual
programs. Information of this sort is contained in document estimated the number of locs per procedure and turns that into a number of bytes
generated. By doing a regression on the number of procedures per chapter of the Pilot
Functional Specification we obtain a figure of one thousand bytes of code for every three
pages of Pilot Functional Specification, deleting first the introductory chapter and
appendices and the first introductory 1.5 pages of each chapter.
phase distribution
We are already part way through the Pilot project. I estimate thjat we are 70-75% through
the definition phase. Of the remaining effort, the P+S indicates that we will have
Prelim Pilot Devp Work Plan - revised September 16, 1977 - DRAFT 3
10% in the definition phase
22.5% in the design phase
22.5% in the code and unit test phase
45% in the system test phase
The system test phase will commence with the release of Pilot 2.0. All other phases will be
completed before the release of Pilot 2.0.
Alto/Pilot - Pilot 1.0 (Internal release only)
references
Memo of May 19, 1977 from D. DeSantis to Bill Lynch - subject: Desired Alto-
based Pilot Functions
Memo of June 2, 1977 from J. Szelong to W. Lynch - subject: Alto/Pilot
documents
Alto/Pilot Functional Spec - Small document to point to relevent parts of the Pilot
FS - to be constructed from Pilot Functional Specification 2.0
Alto/Pilot Design Spec - unreleased memo to be constructed from the Pilot design
spec.
Alto/Pilot Test Spec - memo plan due 15 September 1977
Alto/Pilot Tech Doc - none
Standard release Description
Post release support (bug reports) TBD
DO Conversion
An important dependency in this workplan is th reliance on the existance of a reasonable
programming environment on the DO. The attainment of this situation has come to be
known as the DO conversion problem. The current plan is to achieve the DO programming
environment in a stepwise fashion. The crucial point is to create a set of microcode which
will cause the DO to emulate the Alto. The Mesa debugger will be constructed in such a
fashion that the dubugger itself wiIl execute in the well debugged Alto world while
debugging code operating inthe DO princeops world. The microcode will be swapped upon
entering and leaving the Mesa debugger.
With these facilities, DO programs can be debugged with the full power of the Mesa system
available and without regard to how much or how little of the DO system is working. This
dual world system with the two sets of microcode will be released with Pilot 2.0 so that
Applications can also enjoy the benefits of a reliable mesa debugger in their initial DO
efforts.
[Say something about how the critical I/O devices will be debugged and integrated into Pilot. The file system
Prelim PiJo~ Oevp Work Plan - revised September 16, 1917 - DRAFT 4
will begin with a simulated disk really in main memory, followed by the integration of the real Pilot disk
drivers. This decouples the debugging of the Pilot disk drivers from the development of the rest of Pilot.]
references
Memo dated June 6, 1917 from Wendell Shultz to distribution. subject: Conversion plan to
0(0)
(Johnsson >Conversion27 J un. bravo
see (Lynch)DOConversion.Memo
Documents
name prsn 1 date size effort Alto time
015 - Project document List L 7/30177 2 pages .2mm. 9 hrs.
D12 - Preliminary Work Plan L 8/15177 8 pages 1 mm. 36 hrs.
02 - PiJot Functional Specifications U 8/30177 75 pages 4 mm. 340 hrs.
013 - Design Work Plan L 9115177 10 pages 1 mm. 40 hrs.
018 - Alto/Pilot Test Spec L 9/15177 4 pages 1 mw. 9 hrs.
D16 - Alto/Pilot Functional Spec T 10/1/77 8 pages 1 mw. 9 hrs.
D3 - PiJot Design Specifications R(GHUTO)
11/1/77 225 pages 12 mm. 1000 hrs.
014 - Implementation Work Plan L 11/15177 10 pages 1 mm. 40 hrs.
D17 - Alto/Pilot Design Spec R 11/30177 24 pages 3 mw. 36 hrs.
01 - Pilot Concepts and Facilities U 11/30177 40 pages 1 mm. 60 hrs.
011 - Pilot test Plan L 12/1177 8 pages .5 mm. 36 hrs.
019 - Alto/Pilot Standard Release Description
U 12115177 2 pages 1 mw. 9 hrs.
06 - Pilot Test Specs T 5/1178 8 pages 1 mm. 36 hrs.
07 - Std Release Descriptions (2.0) L 5/31178 2 pages 1 mw. 9 hrs.
08 - Std Release Descriptions (2.1) L 8/31178 2 pages 1 mw. 9 hrs.
09 - Std Release Descriptions (3.0) L 12131178 2 pages 1 mw. 9 hrs.
04 - Pilot Tech Manual 11/1179 150 pages 4 mm. 225 hrs.
totals 580 pages 27.95 mm. 1912 hrs.
Prelim Pilot Devp Work Plan - revised September 16, 1977 - DRAFT 5
Programming Projects
project name prsIIt .... req _