9607 1 IN THE IOWA DISTRICT COURT FOR POLK COUNTY ----------------------------------------------- 2 JOE COMES; RILEY PAINT, ) 3 INC., an Iowa Corporation;) SKEFFINGTON'S FORMAL ) 4 WEAR OF IOWA, INC., an ) NO. CL82311 Iowa Corporation; and ) 5 PATRICIA ANNE LARSEN; ) ) TRANSCRIPT OF 6 Plaintiffs, ) PROCEEDINGS ) VOLUME XXXV 7 vs. ) ) 8 MICROSOFT CORPORATION, ) a Washington Corporation ,) 9 ) Defendant. ) 10 ----------------------------------------------- 11 The above-entitled matter came on for 12 trial before the Honorable Scott D. Rosenberg 13 and a jury commencing at 8:38 a.m., January 22, 14 2007, in Room 302 of the Polk County 15 Courthouse, Des Moines, Iowa. 16 17 18 19 20 HUNEY-VAUGHN COURT REPORTERS, LTD. 21 Suite 307, 604 Locust Street 22 Des Moines, Iowa 50309 23 (515)288-4910 24 25 9608 1 A P P E A R A N C E S 2 Plaintiffs by: ROXANNE BARTON CONLIN 3 Attorney at Law Roxanne Conlin & Associates, PC 4 Suite 600 319 Seventh Street 5 Des Moines, IA 50309 (515) 283-1111 6 MICHAEL R. CASHMAN 7 LINDSEY A. DAVIS Attorneys at Law 8 Zelle, Hofmann, Voelbel, Mason & Gette, LLP 9 500 Washington Avenue South Suite 4000 10 Minneapolis, MN 55415 (612) 339-2020 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 9609 1 Defendant by: DAVID B. TULCHIN 2 SHARON L. NELLES JOSEPH E. NEUHAUS 3 JEFFREY C. CHAPMAN Attorneys at Law 4 Sullivan & Cromwell, LLP 125 Broad Street 5 New York, NY 10004-2498 (212) 558-3749 6 KIT A. PIERSON 7 Attorney at Law Heller Ehrman, LLP 8 333 Bush Street San Francisco, CA 94104 9 (415) 772-6000 10 HEIDI B. BRADLEY Attorneys at Law 11 Heller Ehrman, LLP 333 South Hope Street 12 Suite 3900 Los Angeles, CA 90071-3043 13 (213) 689-0200 14 DAVID E. JONES Attorney at Law 15 Heller Ehrman, LLP One East Main Street 16 Suite 201 Madison, WI 53703-5118 17 (608) 663-7460 18 BRENT B. GREEN Attorney at Law 19 Duncan, Green, Brown & Langeness, PC 20 Suite 380 400 Locust Street 21 Des Moines, IA 50309 (515) 288-6440 22 23 24 25 9610 1 RICHARD J. WALLIS Attorney at Law 2 Microsoft Corporation One Microsoft Way 3 Redmond, WA 98052 (425) 882-8080 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 9611 1 (The following record was made out of 2 the presence of the jury at 8:38 a.m.) 3 MR. CASHMAN: Just a couple of quick 4 housekeeping matters, Your Honor. 5 THE COURT: Okay. 6 MR. CASHMAN: Just for the Court's 7 information, Plaintiffs are going to play the 8 remainder of Mr. Barrett's deposition from 9 Caldera before we have a live witness by the 10 name of David Bradford. 11 And that's a little bit different than 12 what I think I told the Court on Friday that we 13 were going to interrupt Mr. Barrett's 14 deposition. 15 THE COURT: Oh, okay. 16 MR. CASHMAN: And then just secondly, 17 this morning the Plaintiffs filed a motion to 18 supplement one of the proposed preliminary 19 instructions that you issued an order on 20 Friday, I think, concerning the 1995 consent 21 decree between Microsoft and the government. 22 And just for courtesy, I'm going 23 to hand-deliver a copy to Mr. Tulchin and a 24 copy -- 25 THE COURT: What is this? 9612 1 MR. CASHMAN: -- to you. 2 THE COURT: Okay. 3 MR. CASHMAN: And Ms. Conlin, when she 4 arrives mid-morning to handle the examination 5 of Mr. Bradford, I think she'd like to address 6 this particular motion over the lunch hour, if 7 possible, before we resume in the afternoon. 8 THE COURT: All right. 9 MR. CASHMAN: I'll let her bring that 10 up with the Court. 11 THE COURT: We'll see. 12 MR. TULCHIN: Just for the record, 13 Your Honor, Mr. Cashman handed this to me just 14 one minute ago, and, of course, we had no 15 notice that anything like this was coming. 16 MR. CASHMAN: Nothing further, Your 17 Honor. 18 THE COURT: Get the jury, please. 19 (The following record was made in 20 the presence of the jury at 8:44 a.m.) 21 THE COURT: Everyone else may be 22 seated. 23 And I guess Carrie will get the 24 lights. 25 MR. CASHMAN: Thank you, Your Honor. 9613 1 Plaintiffs continue with the testimony 2 of Philip Barrett. 3 (Whereupon, the following video 4 resumed playing to the jury.) 5 Question: Windows 3.1 came to market 6 in April 1992. Does that sound right to you? 7 Answer: Yes. 8 THE COURT: Would you hold up a 9 minute? 10 (Whereupon, the playing of the video 11 to the jury was stopped.) 12 THE COURT: Sorry about that. 13 (Whereupon, the following video 14 resumed playing to the jury.) 15 Question: What did you do after that? 16 Answer: Actually, slightly prior to 17 that, I moved to a different group. The group 18 was called multimedia and consumer systems, I 19 believe. 20 Question: What prompted the move? 21 Answer: A number of factors. 22 One is I was tired of working on 23 Windows. I had been working on it for six 24 years, four years, whatever. 25 Question: Long enough? 9614 1 Answer: I had been working on it, 2 yeah, a fairly long time, six years. And, 3 frankly, I did not get along that well with 4 Brad Silverberg, the VP of the division, and 5 felt it was time to move on anyway. 6 Question: Do you recall when you made 7 the move? 8 Answer: It was just before the 9 release. It was before the release of 3.1, so 10 it would be in the first quarter of 1992. 11 Question: The product by then was in 12 its final beta release? 13 Answer: Yes. It was basically being 14 completed. 15 Question: Thanks. 16 When did you leave Microsoft? 17 Answer: In October of 1994. 18 Question: And you went to what is now 19 RealNetworks at that time? 20 Answer: That's correct. At that time 21 it was called Progressive Networks. 22 Question: Did you initiate the 23 contact with Progressive Networks? 24 Answer: No. 25 Question: Who called you? 9615 1 Answer: Rob Glaser. 2 Question: Had you known him when you 3 were at Microsoft? 4 Answer: Yes. 5 Question: He had been a Microsoft 6 employee? 7 Answer: That's correct. 8 Question: What position did you take 9 at Progressive Networks? 10 Answer: Vice president of software 11 development. 12 Question: Is that the senior 13 technical person at Progressive Networks? 14 Answer: That's correct. 15 Question: Is that still your 16 position? 17 Answer: Yes. 18 Question: You hesitated. Is there -- 19 Answer: Well, the company has grown 20 so big that I believe that all the senior 21 management would still say I'm the senior 22 technical person at RealNetworks, but there is 23 no title that reflects that within the company. 24 Question: I want to back up to, we 25 discussed Windows 3.0, which you had primary 9616 1 responsibility for; correct? 2 Answer: Correct. 3 Question: And Windows 3.1, which you 4 had primary responsibility for up until right 5 before it was released; is that correct? 6 Answer: Perhaps I misheard that. I 7 had primary responsibility for 3.0 and then 8 primary responsibility for 3.1 up until shortly 9 before it was released. 10 Question: Sure. And before Windows 11 95, if you wanted to run Windows, you had to 12 buy DOS, install DOS, and then buy Windows and 13 install Windows; correct? 14 Answer: That is correct. 15 Question: And the two products worked 16 together? 17 Answer: That is correct. 18 Question: DOS provided certain 19 functions and Windows provided certain 20 functions? 21 Answer: Yes. 22 Question: And they worked together 23 even though they were installed separately? 24 Answer: Yes, that's correct. 25 Question: And then Microsoft released 9617 1 Windows 95 August 1995 and changed that? 2 Answer: Yes. 3 Question: And at least to the user, 4 it appears to be a single product, but performs 5 the functions that it used to perform using a 6 separate DOS and separate Windows? 7 Answer: That is correct. 8 Question: But it's your 9 understanding, isn't it, that there is still 10 DOS in Windows 95 and there's still Windows in 11 Windows 95? 12 Answer: Yes. You can find -- you can 13 actually bring up a Windows 95 as DOS. 14 Question: I think when you and I 15 talked about it before, you described Windows 16 95 as a DOS and Windows stuck together with 17 baling wire and bubblegum? 18 Answer: That is a fair if colloquial 19 representation of it, yes. 20 Question: And what do you mean by 21 that? 22 Answer: That basically, yes, there is 23 DOS on the underlying -- under the hood there 24 is DOS. There is a form of DOS, a version of 25 DOS that was -- and I don't know all of the 9618 1 details of what was developed. I don't 2 understand all they did there, but you can 3 actually produce a bootable DOS diskette. 4 There is still 16-bit code inside. 5 Question: And when you said they were 6 tied together with baling wire and bubblegum, 7 you were referring to the amount of integration 8 between DOS and Windows in Windows 95? 9 Answer: Yes. 10 Question: Is it your understanding 11 that the relationship between DOS and Windows 12 in that the relationship between DOS and 13 Windows in Windows 95 is the same as it was 14 between DOS and Windows in prior versions such 15 as DOS 6.X and Windows 3.X? 16 Answer: I wouldn't say that they are 17 exactly the same. To answer your question 18 directly, no. 19 Question: What is your understanding 20 of the technological benefits that Microsoft 21 achieved in combining DOS and Windows in 22 Windows 95? 23 Answer: The main technical advantages 24 I believe are in better management of DOS 25 applications running in the Windows 9619 1 environment, and a more clean install, a better 2 installation process. 3 Question: Because it's a single 4 installation process? 5 Answer: Yeah, because it's a single 6 installation process. 7 Question: My question was, are there 8 any other reasons that it was a better install 9 process other than the fact that it was a 10 single install process? 11 Answer: I believe the answer would be 12 no. 13 Question: And when you say better 14 management of DOS applications running under 15 Windows, what do you mean? 16 Answer: The way that DOS applications 17 are run, DOS applications have a lot of 18 knowledge of the internals of MS-DOS and 19 because of that they have to be handled very 20 specially in order to run well. 21 And Windows 95 did a better job of 22 having specific knowledge of how DOS 23 applications utilize the internals of MS-DOS. 24 Question: Was it required that it be 25 one product for it to do that? 9620 1 Answer: No. 2 Question: It could have been done as 3 two separate products? 4 Answer: Yes. 5 Question: So to the best of your 6 knowledge, the only technical advantage that 7 couldn't have been achieved releasing it as two 8 products is the single install program? 9 Answer: Yes. 10 Question: This idea of combining cost 11 and Windows in the same box with a single 12 install program dated all the way back to 1991 13 as we saw this morning, didn't it? 14 Answer: Yes. 15 Question: Do you have an 16 understanding of what happens when someone 17 turns on their computer that has Windows 95 18 loaded on it, what happens first? 19 Answer: Yes. 20 Question: What happens? 21 Answer: Do you want chapter and verse 22 or do you just want general overview? 23 Question: I want general overview 24 first and I'll ask for more detail if we need 25 it. 9621 1 Answer: Okay. When you turn on your 2 computer, the first thing that happens is the 3 ROM BIOS gets control, and it's called the 4 POST, power on self-test phase. 5 Once it goes through that phase, once 6 it successfully completes that phase, it looks 7 to the hard drive to find the operating system. 8 And based on configuration of that system it 9 will look at different hard drives and that 10 sort of thing. 11 But anyway, it looks at the hard drive 12 to find the operating system and through a 13 somewhat tricky process will load the operating 14 system into memory. 15 And it looks for a specific place on 16 the hard drive to gain a pointer to that 17 operating system that loads into memory. 18 Once the operating system is loaded 19 into memory, it transfers control of the 20 operating system, and the operating system, 21 which typically is a small but first part of 22 the operating system then loads the rest of the 23 operating system into memory. 24 Question: And in the case of Windows 25 95, what loads up first after the ROM BIOS does 9622 1 its work? Is it the DOS portion? 2 Answer: I honestly don't know the 3 answer to that question, but you certainly 4 piqued my curiosity. 5 Question: So you don't know what's 6 happening while I'm staring at that screen that 7 says Windows 95 and has the clouds and all 8 that? 9 Answer: Yeah, actually I do know. 10 What happens during that phase is 11 various device drivers, DOS level device 12 drivers are being loaded, such as CD-ROM 13 drivers and that sort of thing. 14 Question: You may be the most 15 technically savvy witness we've seen in this 16 trial. 17 Would it be difficult for somebody 18 like you to write an installation program that 19 was capable of installing DOS and then Windows 20 if they were offered as separate products? 21 Answer: I don't believe it would be 22 technically difficult. 23 Question: What would be the 24 challenges associated with doing something like 25 that? 9623 1 Answer: Understanding the detailed 2 process of the boot-up sequence, understanding 3 -- or not the boot-up sequence, I'm sorry, the 4 detailed process of where the systems software 5 has to live. 6 And quite frankly, the hardest part of 7 it has nothing to do with what operating system 8 you're installing, but rather understanding the 9 widely varied hardware platforms on which it 10 would be running in order to be able to hit the 11 broad part of the market. 12 Question: Let me be a little more 13 precise. 14 Wouldn't the challenge be the same if 15 you were installing DOS and then Windows or if 16 you were installing DOS and Windows from a 17 single installation program? 18 Answer: I believe the challenge to be 19 equal in those two scenarios. 20 Q. Technically, the challenge of 21 installing DOS and Windows is the same whether 22 you install DOS and then install Windows or 23 whether you have a single installation program 24 installing DOS, then Windows; true? 25 Answer: Yes, that's true. 9624 1 Question: Now, your company now, you 2 sell your products to end users? 3 Answer: Yes, we do. 4 Question: Do you also sell your 5 products to OEMs? 6 Answer: Yes, we do. 7 Question: And do you communicate with 8 the OEMs at all? 9 Answer: Yes. 10 Question: In your dealings with OEMs, 11 have you gained any understanding of how it is 12 they go about loading software on new machines? 13 Answer: Yes. 14 Question: And is it your 15 understanding that they typically will have a 16 standard disk image that they will put on the 17 machine to send out the door? 18 Answer: Yes, they typically will 19 create an image that contains all the necessary 20 software and copy it to the disk, yes. 21 Question: And the disk image will 22 include all the programs that are going to be 23 on the computer? 24 Answer: If it's shipped out -- yeah, 25 absolutely, that's the way they do it 9625 1 typically. 2 They will have a disk image of the way 3 they want the disk to look, the hard drive to 4 look when the customer turns on their machine, 5 and it will place that on the disk through a 6 disk copy mechanism. 7 Question: If you know, is it more 8 difficult for them to place this disk image on 9 the machine if it has five programs instead of 10 two programs on it? 11 Answer: There's no difference. 12 Question: For the benefit of the 13 jury, would you just briefly explain what a 14 beta program is? 15 Answer: A beta program is a way to 16 gain real-world experience with a product 17 before it's released to determine whether or 18 not the product is, frankly, ready for release 19 and to find ways to help make that product 20 ready for release. 21 Question: And when a program is still 22 in beta test -- is that a fair way to say it, a 23 program is still in beta test -- the program is 24 still being tested and still being modified? 25 Answer: That's correct. 9626 1 Question: And you use betas here in 2 your current position, don't you? 3 Answer: Yes, that's correct. 4 Question: And you distribute betas 5 widely? 6 Answer: Very widely, yes. 7 Question: One of the reasons for 8 doing that is to test the product? 9 Answer: That's correct. 10 Question: And another reason is 11 nothing more complicated than marketing, is it? 12 Answer: That's correct. 13 Question: And the reason it works as 14 a marketing vehicle is because it gets the 15 product out in the hands of users? 16 Answer: Yes. 17 Question: Users beyond those that are 18 actually going to test the product? 19 Answer: That's correct. 20 Question: And even the press will 21 look at beta releases, won't it? 22 Answer: Uh-huh, yes. 23 Question: It's your understanding 24 that companies will also look at beta products 25 to evaluate whether to purchase them? 9627 1 Answer: Yes, that's commonplace. 2 Question: So a successful beta is 3 important in the life of a product? 4 Answer: Absolutely. 5 Question: And this is all true for 6 RealNetworks, the same was true when you were 7 at Microsoft? 8 Answer: Yes, sure. 9 Question: Mr. Barrett, could you take 10 a moment and look at what's been marked as 11 Exhibit 1103 and identify it for me? 12 Answer: This is the Windows 3.1 beta 13 plan dated December 17, 1990. 14 Question: Is this something that was 15 distributed to you? 16 Answer: I recall getting a copy of 17 this. 18 Question: We were talking earlier 19 about the different purposes of beta programs. 20 On the page that is stamped X 548718, it says 21 types of beta programs at the top? 22 Answer: Uh-huh. 23 Question: There's a technical beta, a 24 prerelease and a preview. Do you know what 25 those terms mean? 9628 1 Answer: Yes. 2 Question: Could you explain them to 3 me, please? 4 Answer: Pretty much as it states 5 here. 6 The technical beta is for finding 7 bugs, for those people that are actually going 8 to use the product and uncover problems to be 9 fixed. Prerelease is for ISVs so that they can 10 start writing their -- or modifying their 11 programs to work on the new version of the 12 product. 13 Question: And ISVs are independent 14 software -- 15 Answer: Independent software vendors, 16 yes. 17 Question: And the preview release? 18 Answer: Yeah. It was a new term. 19 I think this is the first place where 20 I ever saw the term preview used, but a preview 21 is essentially for -- it is a marketing tool to 22 get the product into the hands of people making 23 buying decisions. 24 Question: And into the hands of 25 people that would be reviewing it like press 9629 1 reviewers, product reviewers? 2 Answer: Yeah, the press would also 3 get a copy of the preview release. 4 Question: And down in the section 5 where it says objectives at the bottom there, 6 it says, provide early exposure to the product. 7 Is that what that means? 8 Answer: Yes. 9 Question: And in the Windows 3.1 10 release, did you go through all three of those 11 types of betas; a technical beta, a prerelease 12 beta and a preview beta? 13 Answer: I do not recall the details 14 of that, to be honest. I was not managing the 15 beta release at that time. 16 Question: Who was? 17 Answer: David Cole. 18 Question: But you had an 19 understanding, didn't you, of the importance of 20 a successful beta to the success of Windows 21 3.1? 22 Answer: Yes, I understand the 23 importance of that. 24 Question: Take a moment and look at 25 what's been marked as Exhibit 1104. If you 9630 1 could identify that for me. 2 Answer: Okay. 3 Question: Can you identify that, 4 please? 5 Answer: It's a piece of E-mail from 6 me to someone named Chuck Straub. 7 Question: Is that somebody who worked 8 for you? 9 Answer: Yes. 10 Question: And this was in -- 11 Answer: It's the end of 1991. 12 Question: December 17, 1991? 13 Answer: Yes. 14 Question: Do you know why there would 15 be a date -- I'm looking at the first page of 16 Exhibit 1104, why there's another date there of 17 Wednesday, February 24, looks like 1993? 18 Maybe it was forwarded at some later 19 time or -- 20 Answer: I have no idea. 21 Question: But looking at the content 22 of the E-mail, it makes sense to you that this 23 was written December 17, 1991; is that right? 24 Answer: That's correct. That's 25 absolutely correct. 9631 1 Question: I won't read the first 2 paragraph, but it appears that you're unhappy 3 with the person's performance; is that fair? 4 Answer: Yes, that's correct. 5 Question: Everybody was under a lot 6 of pressure in December of 1991? 7 Everyone in your group, the Windows 8 3.1 team. 9 Answer: Yes. 10 Question: And it says in the middle 11 of that first paragraph, this beta release is 12 critical to the success of Win 3.1. What did 13 you mean there? 14 Answer: Our goal was to get the beta 15 release out on time. When you make public 16 commitments or even private commitments to 17 outside parties outside your company or even 18 outside your group within your company and you 19 don't perform the way you say you're going to 20 perform, particularly with respect to schedule, 21 people will become disillusioned with you. 22 Question: Hurts your credibility? 23 Answer: Yes, hurts your credibility. 24 Question: Any other reason that you 25 can recall that this beta was critical to the 9632 1 success of the product? 2 Answer: Not in this context, no. 3 Question: Did you have any role in 4 determining the size of the beta program, the 5 number of beta test sites? 6 Answer: Not a direct role, no. 7 Question: Indirect role? 8 Answer: Yes, I had conversations with 9 people about it should be big. 10 Question: And why is that? 11 Answer: Because I believe that big 12 beta tests are very important. 13 Question: And the reason for that? 14 Answer: To gain widespread exposure 15 both on the technical and preview sense, 16 particularly to be able to have a wide range of 17 different machines that you test on in order to 18 find those situations that you are able to test 19 on in internal labs or developer systems. 20 Question: Let me see if I can 21 separate your answer into the two pieces I 22 think I heard, which is there are technical 23 reasons, technical advantages to having a large 24 beta program and then there's marketing 25 advantages too. Is that fair? 9633 1 Answer: Yes, that's fair. 2 Question: And on the technical side, 3 it's because your hope is that you'll have 4 problems reported that you wouldn't otherwise 5 find? 6 Answer: That's correct. 7 Question: And whether that happens or 8 not is going to depend in part on, for lack of 9 a better word, the quality of the beta test 10 sites? 11 If you understand that. 12 Answer: Yes, I do. Yes, absolutely. 13 The relative quality of the beta 14 testers is the single most important factor in 15 the quality of the feedback. 16 Question: And so the jury 17 understands, what do you mean by when you say 18 the quality of the beta test site? 19 Answer: Their ability to -- well, 20 first of all, install the product, their 21 ability to observe a problem and report it in a 22 clear and cogent fashion, the ability of that 23 person or site to specify the environment or to 24 recreate the actions or activities that they 25 went through in order to -- basically in order 9634 1 to create this bug that allow developers or 2 testers the ability to recreate the problem and 3 demonstrate it in their laboratories. 4 Question: So at least for the 5 technical objectives, we've discussed a beta 6 site that either can't install the program or 7 doesn't know how to recognize errors or doesn't 8 report errors doesn't do you any good at all, 9 does it? 10 Answer: From an engineering 11 perspective, you're correct. 12 Question: Even if you never hear back 13 from a user for whatever reason, there may be a 14 marketing value in distributing the product 15 widely? 16 Answer: Yes. There was a joke that 17 ran around that there were a lot of people that 18 were beta groupees. They just loved to get 19 beta software even if they never installed it. 20 So, yes, the answer was yes. 21 Question: The size of the beta 22 program, would it typically increase from the 23 first beta through the preview beta? 24 Answer: Typically, yes. 25 Question: Why is that? 9635 1 Answer: I think it's just a fact of 2 nature. You generally don't cut people out of 3 a beta program. So there's always demand for, 4 you know, new people being added to the beta 5 program. So, therefore, it grows rather than 6 shrinks. 7 Question: And the objective of the 8 last beta, the preview beta, is to get it in as 9 many people's hands as you can? 10 Answer: Uh-huh, yes. 11 Question: That was true back in the 12 Windows 3.0/Windows 3.1 time frame and is true 13 now at the company you're at? 14 Answer: Yes, it was. It is. 15 Question: Do you recall the size of 16 the beta programs -- back up. 17 Did you have responsibility for the 18 beta program for MS-DOS 5.0? 19 Answer: Yes. 20 Question: Do you have any 21 recollection of the size of that program? 22 Answer: I've thought about this and I 23 do not recall the exact size. It was in the 24 thousands, though. 25 Question: Somewhere -- just I want 9636 1 your best recollection. Somewhere more than 2 1,000 and less than 3,000? 3 Answer: I think it was closer to 4 3,000, as I recall. 5 Question: Was that at the end of the 6 beta cycle, start of the beta cycle? 7 Answer: We're scraping the bottom of 8 the barrel on my memories. I think it was 9 towards the end of the beta cycle. I think 10 that's right. 11 Question: Some were larger than the 12 DOS 5.0 beta program? 13 Answer: Yes. 14 Question: Which ones? 15 Answer: Windows 3.1 and Windows 3.0 16 were both larger. 17 Question: Windows 3.1 was somewhere 18 in the neighborhood of 10- to 15,000 beta 19 testers, wasn't it? 20 Answer: I recall 14,000. 21 Question: Digital Research was not 22 permitted to receive a copy of the Windows 3.1 23 beta, was it? 24 Answer: That's correct. 25 Question: That wasn't your idea, 9637 1 though, was it? 2 Answer: No. 3 Question: Do you recall whose idea 4 that was? 5 Answer: Yes. It was Brad 6 Silverberg's. 7 Question: What was the reason for 8 that? 9 Answer: They were a competitor, were 10 his exact words. They are a competitor, no 11 way. Actually, I think there were a few 12 expletives in that statement also. 13 Question: Was that a conversation 14 that you had with him or was that an E-mail 15 exchange? 16 Answer: It was a conversation. I 17 actually overheard someone else asking him. 18 Question: Did you ever discuss the 19 issue with Mr. Silverberg? 20 Answer: No. 21 Question: Did you ever express the 22 opinion to anybody at Microsoft that you 23 thought Digital Research should get a copy of 24 the Windows 3.1 beta? 25 Answer: I believe I did, but I don't 9638 1 recall when or how. 2 Question: I understand it was a long 3 time ago. Why was that your view? 4 Answer: Well, I believe that it made 5 more sense to -- I just didn't think it made 6 sense. I don't know if I had a good -- I 7 didn't have a good technical reason. 8 Question: Do you have an 9 understanding why it was important for 10 Microsoft not to give it, not to give the 11 Windows 3.1 beta to DRI? 12 Answer: Yeah, because the executive 13 vice president said not to. 14 Question: Do you have any further 15 understanding besides that? 16 I'm not quibbling with you. That's a 17 good enough reason by itself. 18 Answer: Well, in general, products -- 19 to make one product compatible with another, 20 you need a version of that product. 21 Question: Why is it important to get 22 the beta version rather than the production 23 release version? 24 Answer: Time to market. 25 Question: What do you mean? 9639 1 Answer: It takes you a fixed amount 2 of time to do any task. 3 Compatibility engineering is a 4 sequence of tasks, therefore, the sooner you 5 can start on those tasks, the sooner you'll be 6 done. Therefore, you would be in the market 7 with a compatibility-engineered product sooner 8 based on getting a version of the product 9 sooner, the product you are compatibility 10 engineering against. 11 Question: So let's take the case of 12 Windows 3.1. 13 If DRI wanted to be compatible, on 14 their product compatible with Windows 3.1 the 15 day Windows 3.1 was released, it's your view 16 that they needed to receive a Windows 3.1 beta? 17 Answer: Yes. 18 Question: By the way, DRI had been 19 provided a copy or had been provided access to 20 the Windows 3.0 beta but they were excluded 21 from the Windows 3.1 beta program. Do you know 22 what changed in between those two beta programs 23 that would cause Microsoft to exclude DRI from 24 the program? 25 Answer: I was aware that they were in 9640 1 the 3.0 beta. Silverberg made the decision not 2 to include them in the beta, and Silverberg was 3 not on board during -- Silverberg joined right 4 after Windows 3.0 shipped. 5 Question: And have you told me 6 everything that you can recall about 7 discussions with Mr. Silverberg about this 8 issue about excluding DRI from the Win 3.1 9 beta? 10 Answer: No. 11 Question: Tell me what else you 12 remember about it. 13 Answer: His comment was -- I recall 14 his exact words, I won't say his exact words, 15 but he said that was a complete foul-up that 16 they were included in the 3.0 beta. 17 Question: Did he explain to you why? 18 Answer: Well, yeah, he did. They're 19 a competitor. 20 Question: And did you have an 21 understanding of in what markets they were a 22 competitor of Microsoft? 23 Answer: In the DOS operating system 24 marketplace. 25 Question: Not in the Windows 9641 1 marketplace? 2 Answer: There was a product from them 3 called Gem, but I don't think anybody at 4 Microsoft took it seriously. 5 Question: It wasn't your 6 understanding, was it, that Mr. Silverberg 7 wanted to exclude DRI from the Win 3.1 beta 8 because they had this Gem product, was it? 9 Answer: No. 10 Question: It was because DRI was 11 marketing a competing DOS, wasn't it? 12 Answer: That's correct. 13 Question: You have 15-plus years 14 experience in the software industry? 15 Answer: Yes. 16 Question: Developing some 17 applications, developing operating systems? 18 Answer: That's correct. 19 Question: You've been involved in 20 some of the most significant product releases 21 in the last ten years? 22 Answer: Yes. 23 Question: Would it be fair to say 24 that every major product that you've worked on 25 was released with bugs? 9642 1 Answer: Yes. 2 Question: It's inevitable, isn't it? 3 Answer: It's a fact of life in the 4 software industry, yes. 5 Question: Can you explain to the jury 6 why that's true? 7 Answer: Computer software is an 8 extremely complex mechanism. It's not voodoo. 9 It's a machinery with many levers and many 10 dials and many knobs, and sometimes the gears 11 grind. The simple fact is that software is a 12 highly complex piece of machinery that has 13 evolved over time. 14 As new pieces are added or new 15 features are added, they sometimes conflict 16 with earlier components, and so it's very easy 17 to create a product that has features that 18 conflict with other features, and thus bugs are 19 born. 20 Question: When you say bug, what do 21 you mean? 22 Answer: I'm sorry. Bug is a 23 malfunction in some way of the software 24 product, of the machinery of the software 25 program. 9643 1 Question: And bugs, for lack of a 2 better term, can be minor, they can be severe. 3 Is there a whole range of types of bugs? 4 Answer: Yes, there's a standard 5 categorization scheme that has to do with 6 severity. And a severe bug will cause an 7 application or operating system to crash. A 8 minor one might be a blemish on the display 9 that you see. 10 Question: Or the product might not 11 work quite in the way that the user expects? 12 Answer: That's correct. 13 Question: Or the product might not 14 work as the documentation says it will work? 15 Answer: That's correct also. 16 Question: And that would also be 17 considered a minor bug? 18 Answer: Probably. It depends on the 19 recipient. If your product is based on some 20 other product that has a bug in it and your 21 product doesn't work, it may not be that minor 22 to you, but it may be very minor to the person 23 that developed the product you're basing it on. 24 Question: When you are considering 25 bugs, you consider not just the severity of the 9644 1 bugs but whether they affect all users or some 2 small subset of users; is that fair? 3 Answer: That is correct. That is 4 correct. 5 Question: Why is that? 6 Answer: Getting a product to market 7 is an exercise in resources, in managing 8 resources. 9 When you have X number of bugs to fix 10 and you only -- you can only fix X divided by 11 two bugs, half of those bugs in the time that 12 you have to release a product, you have to 13 choose which bugs you fix. 14 You choose those bugs based on how 15 severe they are in terms of the overall impact, 16 and you fix them based on the number of people 17 that will be affected, the number of customers 18 that will be affected. A small number of 19 customers are affected, may not be sufficient 20 enough to cause you to want to fix the bug. 21 Question: So in essence, you're 22 running a triage operation? 23 Is that a fair characterization of it? 24 Answer: Yeah, in fact that exact word 25 is used frequently, triage. 9645 1 Question: How do you make the 2 decision to release a product that even though 3 it still has bugs, it's time to release it? 4 Answer: It's a subject of great 5 debate within the industry. You have a set of 6 metrics, a number of bugs, a number of severe 7 bugs. The -- 8 Question: Can I stop you for one 9 second? 10 When you say number of bugs, you 11 typically have a list of bugs that you're 12 tracking against a product? 13 Answer: Yes. Typically, there is a 14 bug database. 15 Question: And that bug database has 16 bugs that you have discovered through internal 17 testing and bugs that you've discovered through 18 the beta test process? 19 Answer: That's correct, bugs from all 20 sources. 21 Question: Do you ever run into the 22 situation where you have bugs that were 23 generated by outside testers, by beta testers, 24 that when you turn to evaluate them, you find 25 out in fact they weren't bugs? 9646 1 Answer: Absolutely. 2 Question: Why is that? 3 Answer: The very nature of reporting 4 a bug is an exercise in opinions and in exact 5 behavior. 6 People will look at a product and 7 think that it should do X. When it does 8 something different from that, they think it's 9 a bug, even though the specification might say 10 it will do X. 11 Question: I'm sorry. I interrupted 12 you. 13 You were describing the 14 decision-making process that results in a 15 product being released even though it, of 16 course, still has bugs. 17 Answer: Yeah, the idea is that you 18 want to make sure that -- the ideal is that the 19 bug count goes to zero. How does the bug count 20 go to zero? Bugs that we can decide not to fix 21 them because they're not important enough to 22 fix or not severe enough to fix or simply, you 23 know, too minor a problem to really fix, or it 24 may even be risks inherent in doing the fix. 25 And typically, what will happen is you 9647 1 will look at the risk of introducing new bugs 2 by fixing a previous bug. And if that risk is 3 very, very high, then you decide not to fix the 4 bug even though it may be a significant bug. 5 Then you classify that bug as 6 postponed and you get your bug count to zero. 7 Once your bug count goes to zero, you ship. 8 Question: Now, the bug count being at 9 zero doesn't mean there's no bugs in the 10 product? 11 Answer: That's correct. That's 12 simply an active bug count, the bugs you are 13 considering for fixing. 14 Question: And you mentioned that even 15 when there's a severe bug you may leave it in 16 because fixing it could cause more problems. 17 What do you mean by that? 18 Answer: Basically when a bug is in 19 the central piece of code, say a tricky piece 20 of code that gets executed every time somebody 21 does something, the risk of breaking that code 22 could mean that you break the product such that 23 it doesn't work for anybody. 24 So there's a risk associated with the 25 bug fix. Every bug fix, every activity, every 9648 1 human activity has a risk associated with it. 2 Bug fixing activities are similar in 3 that the higher the risk of introducing new 4 bugs, the less you want to fix it. I don't 5 know if that answered your question or not. 6 Question: It did. 7 You then -- you take all this 8 information, and I want to talk now 9 specifically about when you were running the 10 Windows and DOS group. 11 Who would make the decision it's time 12 to release the product? 13 Answer: Generally, it was a group of 14 people; myself, Silverberg typically had last 15 say. The development lead, Tom Lennon, on DOS; 16 David Cole on Windows who was the program 17 manager. 18 Typically it's a group of people that 19 you think of as running the project. They're 20 the ones that make the decision, made the 21 decision to ship various products. 22 Question: Do you remember any 23 significant bugs in MS-DOS 5.0 as of the time 24 it was released? 25 Answer: I don't recall any specific 9649 1 bugs. 2 Question: Do you recall generally 3 that there were severe bugs or significant bugs 4 in the product when it was released? 5 Answer: Yes, I do recall there were 6 bugs in the product. 7 Question: How would you describe 8 those bugs, using the terms we've been using 9 here today? 10 Answer: There were some severe ones, 11 there were some minor bugs. I mean, it was a 12 range. Nothing stands out in my memory, 13 though, as being particularly different from 14 any other project one way or the other. 15 Question: Now, I want to back up a 16 little bit. 17 Microsoft didn't develop -- I don't 18 know whether it's PC-DOS -- should I call it 19 PC-DOS or MS-DOS for -- was it a Microsoft 20 product? 21 Answer: No, it wasn't. No, that's 22 not true, that's not true. There was some 23 MS-DOS for development that happened. 24 Question: Was that after PC-DOS 4 25 came out? 9650 1 Answer: It was concurrent with PC-DOS 2 4, but I believe we pulled back and let IBM 3 release PC-DOS 4 first. 4 Question: And then Microsoft came out 5 with MS-DOS 4.01, didn't it? 6 Answer: Yes, I believe that to be 7 correct. 8 Question: And under any definition of 9 the term, that was a very buggy product? 10 MR. STEINBERG: PC-DOS 4 or MS-DOS 11 4.01? 12 MR. HARRIS: Both. 13 Answer: Yes, both had problems. 14 Question: Severe problems? 15 Answer: Yes. 16 Question: If you would take a moment 17 to look at what's been marked as Exhibit 1107 18 and specifically the message -- I believe it's 19 message 284 that's in the middle of the page 20 there. 21 Answer: Okay. 22 Question: Is that an E-mail message 23 from Scottq addressed to you? 24 Answer: Yes, among other people, yes. 25 Question: Do you know who Scottq is? 9651 1 Answer: Scott Queen. 2 Question: Was that somebody that 3 worked for you? 4 Answer: Yes, he worked in my group. 5 Question: What was his role in your 6 group? 7 Answer: He was an engineer. He was 8 on Windows at one point and then he did move to 9 DOS. 10 Question: The subject line is DR-DOS 11 -- well, let me back up. 12 It says to dosdev. That was a group 13 of DOS developers? 14 Answer: Yes, an E-mail alias. 15 Question: An E-mail alias for a group 16 of DOS developers? 17 Answer: Yes. 18 Question: And doswar, do you know 19 what that was? 20 Answer: Yes, doswar, that war team 21 alias. 22 Question: What's the war team? 23 Answer: The war team is a concept of 24 a group of managers that get together and worry 25 about or deal with specific issues. They're 9652 1 essentially the leaders of the project. That 2 they're responsible for, among other things, 3 bug triage as well as handling any 4 extraordinary events or other issues that come 5 up. 6 Question: Were you on the doswar 7 team? 8 Answer: I believe I was. 9 Question: And can you tell me 10 anything more specifically about what the 11 doswar team did than what you just described? 12 Answer: Not really. I mean, 13 generally war teams function that way. 14 Question: So the subject on this 15 E-mail message is DR-DOS CIS message traffic. 16 Do you know what that refers to? 17 Answer: Yeah. CompuServe Information 18 Systems. This was a CompuServe news group -- 19 or chat group specifically related to DR-DOS. 20 Question: And was that something that 21 Microsoft was monitoring? 22 Answer: There were individuals at 23 Microsoft that did monitor that, yes. 24 Question: Were you one of those 25 individuals? 9653 1 Answer: From time to time, yes. 2 Question: And why would you monitor 3 that chat room? 4 Answer: Because they were a 5 competitor. 6 Question: And you wanted to learn 7 about your competitor's product? 8 Answer: That's correct. 9 Question: The message says AHEM. 10 Before we slam DR-DOS with bugs, we should be 11 aware of which ones we suffer from. In 12 particular, Super-PC-Kwik is also unsafe to 13 load high in the MS-DOS 5.0. 14 First of all, what is Super-PC-Kwik? 15 Is that a utility? 16 Answer: Yeah, I think it was a 17 caching utility. 18 Question: And do you know what Scottq 19 was referring to when he said it was unsafe to 20 load high in MS-DOS 5.0? 21 Answer: Yeah. I think it was a 22 caching utility. 23 Question: And do you know what Scottq 24 was referring to when he said it was unsafe to 25 load high in MS-DOS 5.0? 9654 1 Answer: Yeah, it would malfunction 2 under those circumstances. 3 Question: Do you know how it would 4 malfunction? 5 Answer: No, I don't. 6 Question: You just know there was a 7 problem? 8 Answer: Yes, there was a problem. 9 Question: Do you know if this was a 10 problem that DR-DOS was having too? 11 Answer: I believe so. 12 Question: Would you monitor that 13 CompuServe chat group on a regular basis? 14 Answer: Yes. 15 Question: And what would be a regular 16 basis? Weekly? Monthly? 17 Answer: Over a period of a few 18 months, I probably looked at it multiple times 19 a week, three, four maybe. 20 Question: What would you look for 21 when you monitored the discussion group? 22 Answer: Well, everything. I mean, I 23 wanted to know everything that was going on 24 with the competitor; positive things people 25 were saying, negative things people were 9655 1 saying, bugs, issues, suggestions. 2 Question: And you felt that doing so 3 would help you in your efforts to develop a 4 better DOS at Microsoft? 5 Answer: Yes. 6 Question: What was the time frame 7 which you were doing this monitoring? 8 If it's helpful to you, look at 1107. 9 In fact -- 10 Answer: Okay. I remember. 11 It was sometime shortly after the 12 release of Windows 3.0 in May, so June or July 13 through to the release of DOS 5. So that's the 14 time frame that I mentioned. 15 Question: Why during that time frame? 16 Answer: Because that's the height of 17 the DOS 5 development efforts. 18 Question: Were you monitoring any 19 discussion groups that dealt with products 20 other than DR-DOS? 21 Answer: Yes, I was monitoring the 22 Windows discussion groups. 23 Question: To get feedback on how 24 Microsoft Windows was faring in the market? 25 Answer: Yes, that's correct. 9656 1 Question: What about, did you monitor 2 the discussion groups for any DOS competitors 3 other than DR-DOS? 4 Answer: Not to my recollection. 5 Question: Were you at that time 6 concerned about DR-DOS as a competitor? 7 Answer: Yes. 8 Question: Why? 9 Answer: I'm competitive by nature. 10 Question: Were there other reasons? 11 Was it because DR-DOS was having some success 12 in the market? 13 Answer: Well, yes. 14 Question: Was it because DR-DOS at 15 that time had a product on the market that had 16 features not available in Microsoft's DOS 17 offerings? 18 Answer: Yes, that's correct. 19 Question: Do you recall as you sit 20 here today any of those features? 21 Answer: No, I actually don't. I do 22 not recall. 23 Question: Do you recall discussing 24 DR-DOS as a competitive threat with others at 25 Microsoft? 9657 1 Answer: Yes. 2 Question: Who do you recall 3 discussing that issue with? 4 Answer: Tom Lennon. I'm sure I 5 discussed it with others, but I do not recall 6 specifically. 7 Question: What about Brad Silverberg? 8 Answer: Yes, I did. 9 Question: What do you recall about 10 your discussions with Tom Lennon about DR-DOS 11 as a competitive threat? 12 Answer: How we would respond, 13 basically. How to respond, how to add features 14 to the product. 15 Question: Was there any analysis 16 being done of DR-DOS products with the intent 17 of responding to their features? 18 Answer: I'm not sure what you mean by 19 analysis. 20 Question: Were you looking at DR-DOS 21 5 to see what features it had? 22 Answer: Me personally, no. 23 Question: People in your group? 24 Answer: Yes. 25 Question: Who? 9658 1 Answer: Tom Lennon. 2 Question: Do you know if he had a 3 production release copy of it or a beta release 4 copy, do you recall? 5 Answer: I don't think he had a copy 6 of it. I do not believe he had a copy of it. 7 Question: How was he doing his 8 analysis, then, do you know? 9 Answer: He had an outside contractor 10 that was doing some analysis. 11 Question: Do you recall the name of 12 the outside contractor? 13 Answer: No. In fact, I never knew 14 the name. 15 Question: Was this something that you 16 had asked Mr. Lennon to do? 17 Answer: I don't believe I asked him 18 to do it. 19 Question: He just did it on his own? 20 Answer: I don't really recall the 21 genesis of that. 22 Question: Before the lunch break we 23 were talking about MS-DOS 5. And I believe 24 it's your testimony that it was released with 25 some bugs, just as any software product is, 9659 1 some severe and some not so severe; is that 2 correct? 3 Answer: Correct. 4 Question: What about Windows 3.0, it 5 was released by Microsoft with some bugs? 6 Answer: Yes, it was. 7 Question: Bugs that were known about 8 at the time it was released? 9 Answer: Yes. 10 Question: Some severe? 11 Answer: Yes. 12 Question: Some minor? 13 Answer: That's correct. 14 Question: And what about Windows 3.1, 15 was it released by Microsoft with some bugs? 16 Answer: Yes, I believe so. 17 Question: Some severe? 18 Answer: Yes, I believe so. 19 Question: And some minor? 20 Answer: Yes. 21 Question: If you could, would you 22 please give the jury a short version of what a 23 software development cycle looks like. 24 What's involved in a software 25 development cycle? 9660 1 Answer: Okay. Every product 2 development, every project is different, but 3 basically the ideal is that you go through a 4 process initially where there's some sort of 5 marketing requirement stated, need for the 6 product, why does the product need to exist. 7 Then there are a set of engineers or 8 technical people assigned to it. A 9 specification is generally written and that 10 specification details out detailed behavior. 11 From that specification, engineers are 12 assigned responsibility for different parts. 13 Typically at that point an 14 architecture is done that describes the major 15 components and how they will interact. 16 And then developers start working on 17 writing the code for each of the pieces that 18 they're responsible for. 19 Once the code is complete, there's an 20 integration phase where the pieces are put 21 together to see if they work. Generally they 22 don't, so there's amount of rework that goes on 23 to make the product play together, then the 24 components play together. 25 And at that point you have something 9661 1 that typically is called an alpha release, 2 which is -- sort of works, has many problems, 3 maybe doesn't have all the features completed 4 or doesn't have all the features completely 5 working, but it's close enough that people can 6 start testing against it, building test plans, 7 some limited amount, perhaps even showing some 8 limited numbers of customers. 9 Then bugs are fixed, you know, 10 features are completed, whatever. And that 11 takes you to a beta phase, beta release phase. 12 And in each beta release phase, as 13 feedback comes in, more bugs are fixed and you 14 reach convergence on what appears to be the 15 right product to ship. 16 At the end of that cycle when the bug 17 count is low enough and the customer value is 18 clearly there, you release the product. 19 Question: Is there a point in time in 20 that process where the feature set is frozen? 21 Answer: Ideally, yes. 22 Question: In practice? 23 Answer: In practice, no. Feature 24 sets are often changed. 25 Question: And more specifically, how 9662 1 would that happen at Microsoft? 2 Answer: Well, we would have a Bill 3 review, a Bill Gates review, a review of the 4 project with Bill Gates. And he will decide if 5 the product is not competitive enough or 6 doesn't meet other goals of his. That's one 7 way. 8 Other ways are the team might do a 9 similar sort of evaluation and say it's not 10 good enough, and so we'd fix it and move 11 forward. 12 Question: When you say fix it, you 13 mean change the feature set? 14 Answer: Address the deficiencies 15 typically changing the feature set, yes. 16 Question: And sometimes that would 17 involve adding new features? 18 Answer: That's correct. 19 Question: And sometimes you would 20 drop features off? 21 Answer: Yeah, but that's less likely 22 to occur. 23 Question: Why is that? 24 Answer: You don't really drop 25 features. You might just stop talking about 9663 1 them. 2 Question: And these what you called 3 -- were they referred to as Bill reviews or 4 should I call them Bill Gates reviews? 5 Answer: Internally they were Bill 6 reviews. They were called Bill reviews. 7 Question: Do you recall any of these 8 Bill reviews for the MS-DOS 5 product? 9 Answer: Not specifically, no. 10 Question: Generally? 11 Answer: Yeah, generally. 12 Question: What do you recall about 13 them? 14 Answer: They were painful. 15 Question: Why were they painful? 16 Answer: Because Bill is a very 17 forceful character and oftentimes makes things 18 very personal. 19 Question: Were these reviews that you 20 yourself would attend? 21 Answer: Yes. 22 Question: These reviews happened 23 before the product was released, didn't they? 24 Answer: That's correct. 25 Question: And the purpose of the 9664 1 review was to go over the proposed feature set 2 of the product? 3 Answer: That's correct. 4 Question: Were there other purposes 5 to the review? 6 Answer: Well, the reviews were very 7 wide-ranging in that they were feature set, 8 project status, bug lists, bug status, quality 9 status I guess is a better way of describing 10 it. Just about anything was fair game in a 11 review. 12 Question: Can you give me an estimate 13 of the number of these reviews that you 14 attended in the DOS 5 project? 15 Answer: Three, maybe four. 16 Question: What do you remember about 17 the first one? 18 Answer: I suppose you're going to 19 show me an E-mail. 20 I think there was one where I think 21 the word disaster was used a number of times, 22 referring to the IBM DOS 4 stuff. 23 Question: What did you understand 24 Mr. Gates to mean -- it was Mr. Gates -- 25 Answer: Oh, he was extremely unhappy 9665 1 with the results of the DOS 4 project or the 2 product that was shipped. 3 Question: He referred to the product 4 as a disaster? 5 Answer: I believe he did, yeah. 6 Question: And why was he -- if you 7 know, why was he telling your group that? 8 Answer: I think it's Bill Gates' 9 unique form of motivation to the group, telling 10 them that they screwed up and they have to go 11 fix the problem. 12 Question: Do you remember anything -- 13 do you remember Mr. Gates ever talking about 14 the competitive product being offered by 15 DR-DOS? 16 Answer: During his reviews, no, I 17 don't recall. 18 Question: What about at any other 19 time? 20 Answer: My recollections are that he 21 was complaining about them getting good 22 reviews, but I don't recall the circumstances. 23 I don't recall the situation. 24 Question: Complaining about the fact 25 that DR-DOS 5 was getting good reviews? 9666 1 Answer: Yes. 2 Question: Good reviews by product 3 reviewers? 4 Answer: By trade press. 5 Question: When you were brought on to 6 the -- to run the DOS 5 team, was your 7 understanding that you were to get the product 8 to market as quickly as possible? 9 Answer: Yes. 10 Question: And at least according to 11 what I've seen, the product was first announced 12 in April 1990, to be shipped, I believe, in 13 August of 1990? 14 Answer: I believe that to be the 15 case. 16 Question: But the product actually 17 didn't make it -- excuse me. 18 The product wasn't actually released 19 until some ten months later in June of 1991? 20 Answer: That sounds about right. 21 Question: I want you to take a moment 22 and just page through and look at exhibit -- 23 what's been marked as Exhibit 1109. Another 24 trip down memory lane. 25 Answer: Oh, yeah. 9667 1 (Whereupon, playing of video to the 2 jury was concluded.) 3 MR. CASHMAN: Your Honor, this would 4 be a good break for the morning recess. 5 THE COURT: Very well. Stop the tape, 6 please. 7 At this time we'll take a recess. 8 Remember the admonition previously 9 given. You can leave your notebooks here. 10 We'll be in recess ten minutes. 11 Thank you. 12 (A recess was taken from 9:51 a.m. 13 to 10:08 a.m.) 14 THE COURT: Everyone else may be 15 seated. 16 Please continue. 17 (Whereupon, the following video was 18 played to the jury.) 19 Question: In a minute I have some 20 specific questions on just two pages, but first 21 if you can just by paging through it, if you 22 can identify the document for me. That's what 23 I would like you to do. 24 Answer: Yes. 25 Question: What is it? 9668 1 Answer: This document is a postmortem 2 analysis of the DOS 5.0 project. 3 Question: What does that mean, a 4 postmortem analysis? 5 Answer: Well, an after-the-fact 6 review of the performance of the team and the 7 processes and that sort of thing. 8 Question: Do you know who MikeDr is? 9 Answer: Mike Dryfoos. 10 Question: Was he the author of this 11 report? 12 Answer: Yes, he was. 13 Question: Was that somebody that 14 worked for you? 15 Answer: Yes. 16 Question: Was this report prepared 17 under your direction? 18 Answer: Actually, no. It was 19 prepared with my blessing, but it was actually 20 undertaken by Microsoft and several of the 21 project team members. 22 Question: Okay. If you could turn to 23 page 9 of the document that's been marked as 24 Exhibit 1109, it's got Bates number 7020987. 25 Answer: Uh-huh. 9669 1 Question: Up in the history section, 2 I think it's labeled prehistory, in the first 3 paragraph it describes how the DOS 5 project 4 came together. And at the end of that 5 paragraph, it says, to provide an attractive 6 replacement for DOS 4.01. Was that MS-DOS 7 4.01? 8 Answer: Yes, I believe so. 9 Question: Eliminating the quality 10 problems and the poor user and industry 11 perception of the latter. 12 Was that your understanding of how the 13 project came together? 14 Answer: Yeah. Yes. 15 Question: Now, down at the bottom of 16 the page, it talks about the early stages of 17 the development of DOS 5. 18 And it says in the last sentence, the 19 development spec was largely complete by the 20 end of December. The schedule projected coding 21 would be complete by the end of March and the 22 product ready for final release in early 23 August. 24 At what point did you enter into the 25 project? 9670 1 Answer: Well, the project -- the team 2 always reported to me during all of this period 3 of time. It wasn't until May, I believe, that 4 I really took an active role. 5 Question: And that was because you 6 had finished with Windows 3.0 at that time? 7 Answer: That's correct. 8 Question: If you would turn to the 9 next page, page 10 of Exhibit 1109, right 10 before the title changing product -- changing 11 schedule -- well, let me stop right there. 12 Do you have an understanding of what 13 the author meant in changing product, changing 14 schedule? 15 Answer: Well, not in the specific 16 case. In general, this just simply means that 17 when you change the product it takes longer. 18 When you change the product definition, it 19 takes longer. 20 Question: When you start adding 21 features in the middle of the development 22 cycle? 23 Answer: Correct. 24 Question: It lengthens the 25 development process? 9671 1 Answer: That's correct. 2 Question: Well, if you jump ahead to 3 page 15 and you look at the section inadequate 4 specifications, if you could just take a minute 5 and read the first three paragraphs of that 6 section and tell me when you're done. 7 Answer: Okay. 8 Question: Would it be fair to say 9 that the slippage in the schedule from an 10 original projected release date of August 1990 11 to an actual release date of June '91 was due 12 principally to changes in the feature set? 13 Answer: I'd say that that is a major 14 component to that slip. 15 Question: What were the other 16 components, if you recall? 17 Answer: Well, I think the quality. 18 He mentions it here, but the fact is that the 19 quality, the base that they were starting from, 20 had many problems too. 21 Question: Was the starting point 22 actually the DOS 4 code base? 23 Answer: Yeah, it was. 24 Question: So not only was the 25 addition of features necessary -- well, the 9672 1 addition of the features in this changing 2 feature set, but there was also the problem of 3 having buggy code to start with? 4 Answer: Yes. 5 Question: One of the most important 6 stimulants for adding new features was the 7 competitive pressure from DR-DOS, wasn't it? 8 Answer: It was a stimulus. 9 Question: Well, turn with me to page 10 10 back to the section changing product - 11 changing schedule. 12 Answer: Uh-huh. 13 Question: And the end of the first 14 paragraph there it says, one of the most 15 important stimulants for adding features was 16 competitive pressure from DR-DOS 5.0, which we 17 first learned of in the spring of 1990. The 18 DR-DOS feature set led us to add UMB support, 19 task swapping and undelete. 20 Answer: Uh-huh. 21 Question: Is that an accurate account 22 of what happened? 23 Answer: I think so, yes. 24 Question: And where it says the 25 DR-DOS feature set led us to add UMB support, 9673 1 task swapping and undelete, is that one of the 2 results of your analysis of what DR-DOS was 3 offering at the time? 4 Answer: Yes. 5 Question: Other than this change in 6 the feature set stimulated in large part by 7 competitive pressures from DR-DOS, do you have 8 any other explanation for the ten-month slip in 9 the schedule in the release of DOS 5? 10 Answer: Well, a new disk cache 11 utility, SMARTDRV, was added. And that took 12 much longer to do than was anticipated just 13 because of the fact that it had the potential 14 of damaging the user's hard disk if done 15 incorrectly. 16 Question: When was that introduced, 17 if you recall? 18 Answer: I think in the fall. I think 19 in the fall of '90, maybe late summer. In that 20 time frame. 21 22 Question: I'll ask the question this 23 way. 24 Did you first turn your full attention 25 to the project in May 1990? 9674 1 Answer: Approximately in that time 2 frame. 3 Question: Following the completion of 4 Windows 3.0? 5 Answer: Yes. 6 Question: And hopefully a short break 7 afterwards? 8 Answer: Yes. 9 Question: Was there -- what was the 10 status of the project at that time? 11 Answer: I think it was in trouble. 12 It was in trouble. 13 Question: What do you mean in 14 trouble? 15 Answer: The schedules had not been 16 really well thought through. There was a lot 17 of engineering work that had to be done; 18 testing, test plans. 19 There was no beta test plan. There 20 was just a lot of stuff that needed to be done. 21 A lot of activities that needed to be completed 22 in order to ship a product that had not even 23 been completed, not even been started at that 24 point. 25 Question: Did you communicate your 9675 1 views that the project was in trouble to 2 anybody else? 3 Answer: Yeah. I had discussions with 4 Silverberg as well as Tom Lennon and a few 5 other people. 6 Question: What did you tell them? 7 Answer: Well, that the schedule was 8 wrong, we needed to rethink the schedule. And 9 features were coming in, new features were 10 being added. And as a result of Bill Gates 11 reviews features get added, as well as Steve 12 Ballmer and Brad Silverberg got their licks in, 13 so to speak. 14 Question: And in May of 1990, 15 features were still being added to DOS 5? 16 Answer: Oh, absolutely, yes. 17 Question: And that continued on until 18 when? 19 Answer: Probably early fall, mid 20 fall. 21 Question: And once that process 22 stopped, you still had to finish all the coding 23 and testing and beta testing -- 24 Answer: That's correct. 25 Question: -- before the product could 9676 1 be released? 2 Answer: That's correct. 3 Question: Did Mr. Silverberg, 4 Mr. Gates, or anyone else ask you in May 1990 5 if you had a more realistic schedule to 6 propose? 7 Answer: I don't know the answer to 8 that question. It could have been but more 9 likely wasn't. 10 Question: Did you develop what you 11 thought was a more realistic schedule at that 12 time? 13 Answer: Maybe not in May. June or 14 July would have been the time frame. 15 Question: June or July would have 16 been the time frame in which you would have 17 developed a more realistic schedule? 18 Answer: Yeah. There was a series of 19 meetings, now that I think about it. That 20 document reminded me. 21 There was a series of meetings that 22 were about getting our act together on DOS, 23 getting the feature set nailed down, getting 24 the schedule nailed down, project plans, all 25 that sort of thing. 9677 1 That came out of the result of a 2 conversation that I had with Brad Silverberg 3 probably -- that was probably the first major 4 thing that he did when coming -- after coming 5 on board, which I believe was early July, I 6 think. I think he joined after the 4th of July 7 holiday. 8 And so one of the first things, one of 9 the first conversations we had about was DOS 5 10 is in trouble and there needs to be a lot of 11 replanning. And so he basically encouraged a 12 replanning exercise. 13 Question: Did you have any belief at 14 that time that DOS 5 would be released to the 15 market by August of 1990? 16 Answer: In that time frame, no, I did 17 not believe that. 18 Question: Did you believe that it 19 could have been released, DOS 5 could have been 20 released -- I'm still talking about in May of 21 1990 when you turned your full attention to 22 this product -- that it would be released by of 23 the end of September 1990? 24 Answer: I honestly don't know what I 25 was thinking in May. I don't recall what I was 9678 1 thinking in May. 2 Question: Do you recall what you were 3 thinking in June or July when you developed a 4 more realistic schedule? 5 Answer: Well, in July it was pretty 6 clear that we weren't going to make September. 7 That was a shared view of the team. 8 Question: And did you have a view of 9 what a more realistic release to manufacturing 10 date -- or I'm sorry, release date would be? 11 Answer: Release to manufacturing is a 12 correct term. 13 I don't recall. I'm pretty sure we 14 had a realistic view -- what we believed to be 15 a realistic view. I don't know what the date 16 was, though. 17 Question: I'm going to give you what 18 was previously marked as Exhibit 822, 19 thankfully a shorter document. 20 Answer: Okay. 21 Question: It's an E-mail from, is it 22 Russ Werner? 23 Answer: That's correct. 24 Question: To Bill Gates and Steve 25 Ballmer? 9679 1 Answer: That's correct. 2 Question: It says, I have asked Phil 3 Barrett to spend a lot of time making DOS 5 4 happen fast, is the subject. 5 This is what we were discussing 6 earlier when you turned away from -- after the 7 completion of Windows 3.0 and focused on DOS 8 5.0? 9 Answer: Yes, that's correct. 10 Question: And it says, given the 11 aggressive stance taken by DRI this week, I 12 have asked Phil Barrett -- Philba -- to spend 13 more of his time on DOS 5 than on Win 3.1 14 planning, like 60 to 70 percent, to make sure 15 this product happens as fast as it can. 16 Was that your understanding at the 17 time, that it was the aggressive stance that 18 was taken by DRI that was motivating this 19 focus? 20 Answer: Yes. 21 Question: And it says, we are also 22 planning some extensive PR around our beta ship 23 in about three weeks to get across the message 24 that the product is a now thing versus a later 25 thing. 9680 1 Why was that important? 2 Answer: Digital Research was making 3 -- as it mentions in there, making a lot of 4 noises about DOS 5 being vaporware. 5 Question: So it was important to let 6 the market know that DOS 5 wasn't vaporware? 7 Answer: Yes. 8 Question: And one way to do that was 9 to let the market know about the beta that was 10 shipping in about three weeks? 11 Answer: I think that's one 12 interpretation, but I think the correct 13 interpretation was to make noise about the beta 14 when it shipped. That's typically the case. 15 Question: I stand corrected. 16 Wasn't there a beta that was shipped 17 sometime shortly after that? 18 Answer: I don't recall when that beta 19 shipped. 20 Question: Sometime in the summer? 21 Answer: But there was a beta that 22 shipped. 23 Question: Yeah, in the summer of 24 1990? 25 Answer: In the summer, yeah. 9681 1 Question: Did that beta have the 2 features in it that eventually showed up in the 3 product that was released in June '91? 4 Answer: Oh, boy. 5 Question: It couldn't have, could it? 6 The feature set was still being changed through 7 the fall of 1990. 8 Answer: It had some of the features. 9 It did not have all of the features of the file 10 in release of the file. 11 Question: I want to jump back in time 12 to 1988. 13 Answer: Okay. 14 Question: You're at Microsoft. 15 You're responsible for the Windows and DOS 16 development, correct? 17 Answer: Correct. 18 Question: Actually, start in 1989 and 19 work our way backwards. 20 I'll give you what was previously 21 marked as Exhibit 23. After you've had a 22 moment to look at it, tell me, and I've just 23 got a few questions. 24 Answer: Okay. 25 Question: Do you recognize that as a 9682 1 memo from Bill Gates addressed to, among 2 others, you? 3 Answer: Yes, that's correct. 4 Question: In November of 1989? 5 Answer: Yes. 6 Question: Is it fair to say that Bill 7 Gates wasn't very happy with Microsoft's DOS at 8 that point in time? 9 Answer: Yes, that's fair to say. 10 Question: It says in the first 11 paragraph, the widespread belief that DOS 4 12 offers no real benefits, might have bugs, takes 13 more memory and breaks things like redirectors 14 has increased the piracy of DOS and made it 15 more difficult to sell. 16 People selling DOS add-ons have been 17 doing excellent business. DR-DOS is, quote, as 18 good as, closed quote, our DOS so we get into 19 purely price oriented negotiations. 20 Is that a fair summary of the state of 21 Microsoft's DOS at that time? 22 Answer: I agree. 23 Question: You have before you what's 24 been previously marked as Exhibit 7. 25 Answer: Okay. 9683 1 Question: You see in the middle of 2 the page there on the first page of Exhibit 7, 3 there's an E-mail or what appears to be an 4 E-mail from Billg, Bill Gates, and it's 5 addressed to a few people, and you were copied 6 on it; is that correct? 7 Answer: Yes, that's correct. 8 Question: And the subject is DR-DOS? 9 Answer: Yes. 10 Question: The date is September 22, 11 1988, and it says, you never sent me a response 12 on the question of what things in app would do 13 that would make it run with MS-DOS and not run 14 with DR-DOS. 15 Do you know, first of all, why you 16 would have been copied on this message? 17 Answer: Yeah, because those guys 18 worked for me. Well, Tom Lennon worked for me. 19 Question: Do you recall receiving 20 this message? 21 Answer: Vaguely, yes. 22 Question: And do you know what 23 Mr. Gates meant when he said you never -- well, 24 in that first sentence there? 25 Answer: I think it's fairly 9684 1 self-explanatory. I don't know of anything 2 beyond that. 3 Question: That when it says what 4 things in app, is that a reference to an 5 application? 6 Answer: Yes, that's right. An app is 7 typical Microsoft shorthand for application. 8 Question: And that would be a 9 reference to a Microsoft application in this 10 context? 11 Answer: I honestly don't know. 12 Question: You don't recall discussing 13 this message with anybody? 14 Answer: I don't recall. 15 Question: It's addressed Pascalm, 16 which is -- who is that? 17 Answer: Pascal Martin. 18 Question: Pascal Martin and Russ 19 Werner and Tom Lennon? 20 Answer: Correct. 21 Question: Did any of the three of 22 them come to you and discuss this message? 23 Answer: Sorry, I just don't remember 24 the circumstances. 25 Question: What about this subject 9685 1 generally, the subject of what things an 2 application would do that would make it run 3 with MS-DOS and not run with DR-DOS? 4 Do you recall discussions at Microsoft 5 about that subject? 6 Answer: I recall generally that there 7 were discussions about that, yes. 8 Question: In this time frame in 1988? 9 Answer: It could have been that time 10 frame, it could have been later. 11 Question: You can't pin it down? 12 Answer: I can't nail it down. 13 Question: What do you remember about 14 the discussions? 15 Answer: They were fairly esoteric 16 discussions of what sort of things, device 17 driver chains, and memory blocks and that sort 18 of thing, function calls, minutia of the 19 possibilities and the differences between 20 MS-DOS and DR-DOS. 21 Question: Were you looking at any 22 products other than DR-DOS when you were 23 discussing this issue of what things an 24 application would do that would make it run 25 with MS-DOS and not run with some other 9686 1 operating system? 2 Answer: Yeah, there were -- there are 3 several other operating systems and I don't 4 know if they came up in the context of this or 5 they came up in just the general context of 6 competitors, but there were other operating 7 systems clones, if you will, that were 8 discussed. 9 Question: Clones that were discussed 10 generally? 11 Answer: Generally. 12 Question: You don't recall any 13 discussions, do you, of what can we do to our 14 applications to make it so they don't work with 15 some clone other than DR-DOS? 16 Answer: Sorry. No, the answer is no. 17 Question: This is what was previously 18 marked as Exhibit 36, and specifically what I'm 19 interested in, Mr. Barrett, is the E-mail that 20 starts there on the top half of the first page. 21 Answer: Okay. 22 Question: Do you recognize that 23 E-mail message? 24 Answer: Yes, I recognize it. 25 Question: Who is it from? 9687 1 Answer: I don't recall the name of 2 the person, Percyt. Percy somebody. 3 Question: It's addressed to you, and 4 I'm reading in the center of the page there on 5 the first page of Exhibit 36. 6 It says, last Thursday you asked me 7 for a user's view of DR-DOS 5.0. 8 Do you recall asking -- 9 Answer: I don't recall asking. It's 10 a very frequent question that I ask people. If 11 somebody says they know something about a 12 competitor, I say, give me a Cordone, tell me 13 everything you know. 14 Question: And in this case, is it 15 your understanding that that's what this 16 message is from Percyt to you? 17 Answer: Yes. 18 Question: And you took the message, 19 and it appears that you forwarded it to the DOS 20 war group? 21 Answer: That's correct. 22 Question: And why did you do that? 23 Answer: Because this is information 24 that is incredibly valuable from a competitive 25 stance. 9688 1 Question: If you turn to the third 2 page, you see Percyt's conclusion. It says, 3 DR-DOS 5.0 is vastly superior to MS/PC-DOS 3.31 4 and 4.01. It is about as good as MS-DOS 5.0. 5 Would you agree with both of those 6 statements? 7 Answer: Not in tone. I'm recalling 8 this guy was a fairly wild-eyed guy. 9 Clearly 5.0, DR-DOS 5.0 was superior 10 to 3.31 and 4.0. 11 Question: So it's the vastly 12 adjective that you don't like? 13 Answer: I'm not even sure superior. 14 I would say better than. But in terms of 5.0, 15 yeah, it would probably be fair to call them a 16 wash. 17 Question: Just to make sure I 18 understand your testimony, it's that DR-DOS 5 19 was a better product than MS/PC-DOS 3.31 and 4? 20 Answer: Yes. 21 Question: And that it was a wash or 22 it was comparable to MS-DOS 5? 23 Answer: Yes. 24 Question: It did come to market about 25 a year sooner than MS-DOS 5, didn't it? 9689 1 Answer: That's my recollection. 2 Question: This is what has been 3 marked as Exhibit 1114. And I'm interested 4 only in the message that begins in the middle 5 of the first page of Exhibit 1114. It's got a 6 604 at the start, do you see that? 7 Answer: Oh, right there. Yeah, okay. 8 Question: And then it carries forward 9 to where the 605 is on the next page, and I 10 will only ask about the message. 11 Have you had a moment to read that? 12 Answer: Okay. I've read that piece 13 of E-mail. 14 Question: Can you identify that as a 15 string of E-mail between you and Brad 16 Silverberg? 17 Answer: Yes. 18 Question: Now, if I'm going to read 19 this in the order it happened, it's my 20 understanding I start at the end and work my 21 way up? 22 Answer: In this case, yes. 23 Question: So on the second page of 24 what's been marked as Exhibit 1114, the first 25 message in the string is from Brad Silverberg 9690 1 to you; correct? 2 Answer: Yes. 3 Question: And the subject is DR. Is 4 it your understanding that's Digital Research? 5 Answer: Yes. 6 Question: It says, can you tell me 7 specifically what we're going to do to bind 8 ourselves closer to MS-DOS? Since you haven't 9 been replying to my messages, I don't know how 10 to interpret your silence. 11 Answer: Interprety. 12 Question: Interprety your silence. 13 Let me emphasize the importance; IBM 14 is going to announce the DR-DOS deal at COMDEX 15 (almost 100 percent certain.) Okay? 16 Can you explain to me what he was 17 talking about there? 18 Answer: It's fairly self-explanatory. 19 He had been asking me to determine how we were 20 going to bind our products closer to DR-DOS. 21 Question: To MS-DOS? 22 Answer: I'm sorry, MS-DOS. 23 One of these days I'll get it right. 24 To MS-DOS. 25 He wanted a plan from me, and I hadn't 9691 1 gotten it back to him. 2 Question: And what does the fact that 3 IBM is going to announce the DR-DOS deal at 4 COMDEX have to do with any of this? 5 Answer: Well, I don't recall what he 6 was -- I don't recall what was being discussed 7 there on that one. I think just being able to 8 know our plans with respect to DR-DOS was 9 important. 10 Question: And part of those plans 11 involved tying your products closer to MS-DOS? 12 Answer: I believe that's what the 13 E-mail says. 14 Question: But is that your 15 recollection? 16 Answer: Yes. 17 Question: And part of the reason that 18 there was concern at this time was the concern 19 that IBM was going to announce a deal with 20 DR-DOS at COMDEX? 21 Answer: Uh-huh. 22 Question: Is that a yes? 23 Answer: Yes. Sorry. 24 Question: It's just for her, okay. 25 Now, we go to the next E-mail, working 9692 1 our way toward the front page of Exhibit 1114. 2 And that's your response to Mr. 3 Silverberg, isn't it? 4 Answer: Correct. 5 Question: Starts with an apology for 6 the silence. 7 And then it goes on to say, the 8 approach that Ralph and I have discussed is to 9 use a VXD to extent DOS by patching it. 10 Can you explain what you meant there? 11 Answer: Yeah. VXD is a virtual 12 device driver which sits inside the virtual 13 machine manager portion, the 32-bit portion of 14 the system. 15 Typically, what happens is you can 16 extend DOS by putting these little -- they're 17 called -- using break point registers or 18 illegal instructions in 8086 mode. 19 And when a program that's executing 20 along hits those, it will have -- basically get 21 control transferred down to the virtual machine 22 manager. So the virtual machine manager can do 23 something, whatever it is. 24 So this is a way of adding features, 25 this is a way of extending DOS, and I use the 9693 1 term in quotes, so that you can add new 2 features. 3 For example, and I'm only speculating 4 here, but this is how one would implement a 5 better file system, by causing an exception or 6 basically gaining control on entry to the file 7 system code of DOS. 8 So that when a program does a file 9 open instead of actually executing DOS code, it 10 will transfer control down into the virtual 11 machine manager and execute in 32-bit code and 12 go faster or whatever, whatever could be done 13 to make it better. 14 So that's the basic idea that Ralph 15 and I had discussed. 16 Question: This virtual machine 17 manager at that time was part of Windows? 18 Answer: Yes, part of Windows 3.0 and 19 to be part of Windows 3.1. 20 Question: So it was a way not only to 21 tie MS-DOS and Windows more closely together, 22 but it was also a way to improve performance? 23 Answer: Yes, that's correct. 24 Question: I'll hand you what's been 25 marked as 1115. 9694 1 I'm interested only in the message 329 2 in the middle of the page. 3 Answer: Uh-huh. 4 Question: Have you had a chance to 5 read that? 6 Answer: Yes. 7 Question: Do you recognize that as an 8 E-mail from David Cole addressed to you and to 9 Karl -- 10 Answer: Stock. 11 Question: Stock. 12 September 1991? 13 Answer: Yes. 14 Question: And at the end of the first 15 paragraph after describing -- well, I'll be 16 more accurate. 17 It says, it's pretty clear we need to 18 make sure Windows 3.1 only runs on top of 19 MS-DOS or an OEM version of it. 20 And then goes on, describes what he's 21 done, and at the end, he says, we are supposed 22 to give the user the option of continuing after 23 the warning. However, we should surely crash 24 at some point shortly later. 25 What is he talking about there? 9695 1 Answer: I think it's fairly 2 self-explana