2 - 8/31/01 6:35:01 AM - barry: Good morning, everybody. Sorry for the communication problems we're having. Most people are getting the material ok, but there are a lot of different configurations to get to work. We're fixing them as fast as we can.
3 - 8/31/01 6:37:05 AM - barry: I'm going to type in some answers to the most frequent questions now. Feel free to follow up on the answers, resubmit your previous questions if they're critical for me to answer, and to submit new questions. I'll try to operate first-come, first served, except for related questions.
4 - 8/31/01 6:38:11 AM - barry: Q: What changes between a project competing on schedule and one competing on cost?
5 - 8/31/01 6:41:01 AM - barry: 4 A project competing on schedule will focus on reducing effort along its critical path tasks, even when this adds effort off the critical path. Examples are more training and test preparation. It will also try to set things up so that as many people can operate in parallel as possible (e.g., by producing validated module interface specifications.
6 - 8/31/01 6:42:10 AM - barry: Q: Why is a mid-range cost-quality position on the Porter "rollercoaster" curve risky
. - 8/31/01 6:42:52 AM - Frank has entered the room
7 - 8/31/01 6:45:44 AM - barry: 6 It is too easy to be threatened from the left (low-cost producers learning how to improve quality while retaining low cost) or from the right (high-quality producers learning how to retain quality at lower cost). An example is the mid-range Oldsmobile car, threatened by higher-quality Toyotas and lower-priced Mercedes.
8 - 8/31/01 6:46:29 AM - barry: Q: Can MBASE documents be combined for small projects?
. - 8/31/01 6:47:02 AM - hongmin has entered the room
9 - 8/31/01 6:49:06 AM - barry: 8 Yes; it's a good idea. We're working on it. A start is that we are identifying the usually-essential subsets of each document to work on first in this year's CS577 guidelines. Againn, there will be no one-size-fits-all solution, and risk will be your best criterion for what to leave in and drop out.
10 - 8/31/01 6:49:16 AM - Frank: 5 A assume a project competing on cost just tried to reduce cost wherever possible?
11 - 8/31/01 6:51:00 AM - barry: 10 Basically yes, but subject to the risk that too little quality will produce something that nobody will want to use or buy (the downward left end of the curve).
12 - 8/31/01 6:51:30 AM - barry: Q: How does MBASE extend the Rational Unified Process?
13 - 8/31/01 6:53:15 AM - barry: 12 Its main added features are model clash avoidance, stakeholder value propositions and reconciling them via WinWin, business case analysis, and more traceability and crosslinks among the artifacts.
14 - 8/31/01 6:53:50 AM - barry: Q: Does MBASE completely prevent model clashes?
15 - 8/31/01 6:57:06 AM - barry: 14 No; you may not realize that the model clash is there. It may be hidden as an unstated assumption in the user's organization or in a COTS product. A good way to try to find these with users is to show them the model clash concept and examples, and ask if they see any things like these that might compromise the project.
16 - 8/31/01 6:57:59 AM - hongmin: Are we going to have FAQs from discusion sessions posted on our course web?
17 - 8/31/01 6:58:03 AM - barry: Q: With incremental development, is there a single system boundary?
. - 8/31/01 6:59:51 AM - Techsupport has entered the room
18 - 8/31/01 7:00:01 AM - barry: 17 Good question. No, the boundary may be different between the Initial Operational Capability and the Final Opearational Capability. What you put in first depends on how much you know about the nature of the Final Operational Capability up front.
19 - 8/31/01 7:01:19 AM - barry: 16 Yes; we'll post these on the Web site. ***Omer -- please do this for each session***
20 - 8/31/01 7:01:58 AM - Techsupport: 19-Ok Professor
. - 8/31/01 7:02:11 AM - SteveU has entered the room
21 - 8/31/01 7:02:11 AM - Frank: 18 Can stakeholder assumptions change as the project goes on and if so, how whould that be handled?
22 - 8/31/01 7:04:24 AM - Techsupport: just a word regarding the chat tool; plz refresh your browser once in a while to see updated whiteboard and agenda. It seems they are not automatically refreshing. I'm working on it.
. - 8/31/01 7:05:51 AM - SteveU has left the room.
. - 8/31/01 7:05:57 AM - Steve has entered the room
23 - 8/31/01 7:05:58 AM - barry: 21 Yes; this happens pretty frequently as organizations reorganize, people change, budgets get cut, or IT infrastructure changes. The MBASE Life Cycle Plan has an initial section called Assumptions where you can stste the critical assumptions that, if they change, will mean that the cost, schedule, or delivered product will have to be renegotiated.
24 - 8/31/01 7:09:07 AM - barry: Welcome, everybody! If you're just logging on, you can just see the latest section of the material. Click on Threaded Agenda and you'll be able to catch up on where we've been. ***Omer -- can you post this somewhere that new people can see it when the log in?***
25 - 8/31/01 7:10:13 AM - hongmin: In EC-2, Competing on Schedule, Cost, and Quality, there are some acronyms such as RVHL, DPRS, CLAB, RESL, PPOS & RCAP as input to RAD Extension. Where can I find the full tech words of thos acronyms?
26 - 8/31/01 7:11:11 AM - Frank: 22 My chat session does not scroll back. I can see 14 - 25 only now.
27 - 8/31/01 7:12:34 AM - Techsupport: 26-the main chat mesg board can only hold a specific number of lines, if you want to see the lines before that you have to see the Log, which you can see by clicking "Normal" or "Threaded" on the bottom right of the chat tool
28 - 8/31/01 7:13:55 AM - barry: 25 Thanks for bringing this up. The best place to look is the Acronyms and Glossary section of the "Software Cost Estimation with COCOMO II" book, stating on page 477. I notice it doesn't have the CORADMO acronyms; they're defined in Section 5.3 of the bood (P. 214)
. - 8/31/01 7:14:24 AM - Keun Lee has entered the room
29 - 8/31/01 7:15:31 AM - barry: Q: Are there situations where the Waterfall model is still the best thing to use?
30 - 8/31/01 7:19:30 AM - barry: 29 Yes; when there are no real technical or user interface risks, and where the biggest risk is that the people joining the project have different ideas about what its requirements are. Then it pays to work out the requirements first, before the participants run off and design or code incompatible things. Thus, the waterfall model is a risk-driven special case of the spiral process model generator. You can see the special case going down the lower right diagonal of the spiral diagram.
. - 8/31/01 7:21:23 AM - ye has entered the room
31 - 8/31/01 7:22:32 AM - barry: Everybody -- when you come back from a separate screen like the Threaded Agenda, often the cursor is not in the text box and your typing won't be recognized. Click on the text box for the cursor to reappear.
32 - 8/31/01 7:23:48 AM - Steve: Barry, are we going to be using the COCOMO model in this course?
33 - 8/31/01 7:25:55 AM - barry: 32 We are going to use the COCOMO II model in the course. I am working on redoing the parts of Chapters 10-20 in "Software Engineering Economics" that use COCOMO-1981 in their examples. 
34 - 8/31/01 7:26:54 AM - barry: Q: How do you quantify benefits and risks?
35 - 8/31/01 7:32:06 AM - barry: 34 Sometimes data is available for such things as server reliability or downtime, or for the cost of a minute of downtime (e.g., $25K/minute for amazon.com). Sometimes relative ratings are sufficient (e.g., on a scale of 0 to 10). Sometimes you can run user-satisfaction experiments or surveys to get data. We'll cover this more in Chapters 15 and 18 of "SW Engr. Economics."
36 - 8/31/01 7:33:21 AM - hongmin: When the project's requirements are pretty much complete, but some requirements will be changed from time to time. When make decision on process model, Waterfall or Spiral, what is the key?
37 - 8/31/01 7:39:36 AM - barry: 36 The best thing to do is to make your best judgement up front (and to get your customers' and users' best judgements) about what things are likely to change in the future. The MBASE Software And System Requirements Description guidelines has a section called Evolution Requirements where you state these. Then you can use this information to architect your system to make it easier to adapt to these changes. A classic 1979 IEEE-TSE paper by David Parnas has guidelines for how to do this; basically by hiding the sources of change within modules. This will work for both Waterfall and Spiral models. Generally, your predictions won't be perfect, but they will be better than no information.
38 - 8/31/01 7:41:50 AM - barry: Omer -- Can we leave the chatroom running during the 830-950 live session in Studio A? I'd like to elaborate on some of these answers verbally, and to use them to stimulate further questions from the class.
39 - 8/31/01 7:42:40 AM - barry: Q: What is Adaptive Software Development?
40 - 8/31/01 7:43:16 AM - hongmin: Are 830-950 live session on Fridays going to be webcasted?
41 - 8/31/01 7:43:38 AM - Techsupport: 38- Yes Professor You can
. - 8/31/01 7:45:05 AM - andreas has entered the room
42 - 8/31/01 7:50:42 AM - barry: 39 It is one of several "Lightweight" or "Agile" methods that have been emerging to avoid overly bureaucratic "heavyweight" methods for software development. Adaptive Software Development is described in a book of that name by Jim Highsmith (Dorset House, 2000). Other examples are Extreme Programming, Crystal, Scrum, and Lean Development. They emphasize code vs. documents, stories vs. Requirements specs, deep customer involvement, pair programming, small releases, refactoring, continuous integration, etc. We'll cover them later in the course, and post some articles on them.
43 - 8/31/01 7:51:19 AM - barry: 40 Yes, we will webcast the Friday Q&A sessions.
. - 8/31/01 7:51:48 AM - andreas has left the room.
. - 8/31/01 7:53:01 AM - ye has left the room.
. - 8/31/01 7:56:00 AM - Steve has left the room.
44 - 8/31/01 7:57:38 AM - hongmin: Sometimes, new people will be added into the project to speed up the process. Are there any major rules or models to follow to avoid the possible chaos (integrations, commucations among the workforce, ...)?
45 - 8/31/01 7:57:38 AM - barry: We seem to be losing people faster than we are gaining them, and not getting many questions. I'll continue for anither 10 minutes.
46 - 8/31/01 7:58:21 AM - barry: Frank, Hongmin -- While you're on, do you have any suggestions on how to improve the chatroom sessions?
47 - 8/31/01 7:59:07 AM - Techsupport: 46- also if you face any problem or have any suggestions particular to the chat tool , plz let us know
48 - 8/31/01 8:00:46 AM - Frank: The BofA is a project we can look back on, with hind sight, and tell where there assumptions were wrong. Could we try to do a project using for sight? Who are the stakeholders in this class? What assumptions and requirements going into the class? I am seeing a lot of new things being tried here.
49 - 8/31/01 8:01:26 AM - barry: 44 The best thing to do is to architect for ease of having new people contribute (e.g., by giving them modules with well-defined and validated interface specifications, or by organizing testing so that additional testers can do useful tests without conflicting with others).
50 - 8/31/01 8:04:24 AM - Frank: 46 You have only posted nine questions. If you could post the questions ahead of time it might spark more questions.
51 - 8/31/01 8:04:31 AM - hongmin: 46--Professor, I think this works great for me. It could be great that at the beginning of this session if you tell us how many prepared "the most frequent questions" you are going to answer.
52 - 8/31/01 8:07:09 AM - barry: 48 Let me see what we can do about involving you and the other stakeholders in ensuring compatible assumptions. We've done this for CS577, where the stakeholders are the students, faculty, campus clients and users of the team=project products, industry consumers of educated students, and the general software practitioner community in terms of relebvance of MBASE and the other methods we use. There's a paper you can find on the 577a web site called "A Stakeholder WinWin Approach to Software Engineering Education" (approximate title).
53 - 8/31/01 8:08:30 AM - barry: 50, 51 I'll do more of that next week. This week I was away Thursday.
54 - 8/31/01 8:10:43 AM - hongmin: Thanks for answering all my questions today. I certainly will watch today's 830- 950 session later. Have a great weekend to you all!
55 - 8/31/01 8:11:40 AM - barry: Frank, Hongmin -- I think we'll wind up at this time. Omer will keep the session going, and I'll log back on and use the material in the 830 live/webcast session. Thanks for your participation.