EXHIBIT 99.158 Here are my notes for tonight. One point from today's "Energy Central" news, do we need to include the new California APX with its 7-day ahead forward market in our work? The project will be in three phases. We will describe the project in the proposal but it will be noted as a means for SCE to protect itself from any aggressive business activities of competitors and to allow SCE to take advantage of any business opportunities that the new market provides. I will write a (estimated six page) paper that will go to a SCE lawyer (and then presumably to Heller) explaining what the software can really do. I will not imply that Perot will help use with the gaming issues, only that Perot will design the ISO simulation to parrot the existing system and that Perot can clarify our understanding of the existing ISO/PX protocols and system. I will have this done AM tomorrow. From Perot's conflict of interest perspective, I agree that the ISO must be made aware of any "holes" that Paul discovers. (Will Paul be off the ISO project by December such that any "discoveries" after that time are not Perot's problem?) If there is a paper trail of the ISO being notified of every "discovery", it is unclear to me, that our "knowing" about the problems as well constitutes a conflict of interest situation. The first phase is the "functional model phase." We develop the basic model and basic options. It contains all the basic data including placeholders for the information we think we would get from the ISO and PX. The system can be tested by SCE staff at this time for any missing features and reasonableness of results. (The model would generate its own test cases based on simulated computer players.) This phase is done by September #. The second phase is the "operational model phase." In this phase we determine the available and actual data from the ISO/PX and add the analysis capabilities to take the advantage of the data stream in the model. We again "make-up" test data streams and train the SCE staff how to use the system plus make sure the system behaves as planned. This phase ends December 31. The third phase is the "real-time implementation phase." We monitor the system during the early days of operations. We have "fall-back" positions when the model produces anomalous results. We fine tune our algorithms and add any required features. We benchmark the model compared to optimal or "other utility" performance. (Do we want to test this in Phase 2 or 1 to show what savings we might produce under a given optimality assumption of what the other players will do?) As an intrinsic part of the system, the CIGMOD engine will allow SCE to test the impacts of others entering the California market or the impacts of mergers, acquisitions, and takeovers among competitors or by SCE. This could be phase four consulting because the consolidation (or dilution) changes how you need to game the system. I have PEROT, Paul, Jeff and George broken out. Jeff is often Jeff's shop and Perot is the Perot shop. In many instances, the designated time is a very rough guess and the time per party arbitrary. The party doing the task and the effort can be change as desired. Just get things going I have assumed my rates for Jeff and I ($250 per hour) and $450/hour for all Perot. Phase 1 would then be $560K, and Phase 2 as $448K, Phase 3 depends on how much support we need to give. The total is too high for Heller unless Hemant does some nifty marketing, but it is near the million mark and can be "refined" down easily. Also we don't have enough hours among us so we have to drop the person weeks to something more realistic. Exclusivity: We need to say something. The less the better, but we do want to do as needed to keep our future marketing options intact. Two options so far if we are forced to give (some) exclusivity: we get % of market profits (loses too?) or project price is 3X of original price. Tasks of Phase I Review PX and ISO documents for protocols and business opportunities. 2 weeks George, 1 week Paul (might be two weeks for Paul. I am just assuming he knows much if it already.) Convert to reduced-form ISO model Paul (4 weeks?) Jeff's or Alan's time for database and system integration (2 weeks) Enhance PX model to contain California-specific protocols. Jeff (two weeks), Perot (1 week) Incorporate strategies into systems. George (3 weeks), Jeff (3 weeks) Add short-term forecasting structure for all players in and out of California using Jeff's "live-weather" Farsight model. Jeff (2 weeks) Enhance interface for SCE gaming/system-control/SCE data-entry/SCE biding strategy output. Jeff (3 weeks), Perot (1 week) Calibrate CIGMOD to competitor and SCE data. Jeff (2 weeks) Add plant specific data to model Jeff (2 weeks) Test model with SCE staff and update algorithms/interface. Perot (2 weeks), Jeff (2 weeks) Add SCE proprietary data Perot (3 weeks), Jeff (1 week) Consultant test of basic strategies to determine realistic use of strategies. Perot (4 weeks), Jeff/George (4 weeks) CIGMOD Software $50K (Heller already told this as part of initial McMenamin proposal for the "basic" CIGMOD gaming effort.) Phase II Determine PX/ISO data stream. Use of data and transfer of data to system. Perot (2 weeks) Determine use of data stream and implement AI/statistical code Perot (4 weeks), George (4 weeks) Develop real-time data retrieval system Perot (2 weeks), Jeff (1 week) Develop portfolio rules for real-time and over-all strategies Perot (1 week), George (1 week), Jeff (2 weeks) Make test data stream and test with SCE staff (training) Perot (3 weeks), Jeff (3 weeks) Consultant test of real-time strategies to determine realistic use of strategies. Perot (4 weeks), Jeff/George (4 weeks) Any language conversions for speed and better integration??? Phase III Develop default fail-safe algorithm for bidding, add any AI that would warn of possible "bad calls." Perot (4 weeks), George (4 weeks), Jeff (2 weeks) Monitor early use of system. Determine and fix any areas needing tuning or enhancement. Support SCE staff on usage. Benchmark savings