Enable JavaScript.

1988: DDM & Frayed DRDA Tempers

Rick Sanders and I groaned when Roger Reinsch indicated that the next item on the agenda of our meeting would be a presentation by Mel Zimowski on how Systems Network Architecture (SNA) Advanced Program to Program Communications (APPC) could be used by the

(DRDA) that we were designing. We had repeatedly tried telling Mel that this protocol work was already covered in great detail in Rick had written those sections of DDM with his typical zeal for dotting every "i" and crossing every "t", a labor of many months. And further, these protocols had already been reviewed, approved, and implemented by four IBM products, some of which would also be implementing DRDA.

Mel had a copy of the DDM specifications, but here we were, months into the DRDA project, and he still hadn't read them. He had this unshakable notion that DDM architecture only dealt with message formats. And this was at least the fourth meeting where we'd sat through his fumblings with SNA APPC.

Somewhat impatiently, I blurted out "For heaven's sake, Mel, how many times do we have to go through this? This work is already done and written up in the DDM specifications."

Mel scowled at me, but before he had a chance to say anything, a reaction to my comment came from an unexpected source. Bruce Lindsay, from IBM's Almaden Research Lab jumped out of his seat across from me, grabbed a white-board marker from the table, and threw it at me, shouting "And I'm sick and tired of you DDM guys trying to stuff all of that file stuff down our throats. Relational databases are different from what you did for distributed files, and it just doesn't apply." Fortunately, the conference room tables had been arranged in a large U shape and he was nearly eight feet away from me. The marker missed me, but I really don't think he was aiming right at me. He just needed to throw something.

With his shoulder length red hair and full beard Bruce looked like a mad hippie who'd just escaped from a 60's commune. In fact, this was partly true. He had been a member of a commune in the 60's, and was still an advisor to one, but he was also a highly respected database researcher with a doctorate from Stanford University. He had been involved in System/R, the original IBM prototype of relational databases, and then later he had worked on System/R*, a prototype of relational databases executing SQL statements on multiple computers.

But now, he was getting more and more excited as he shouted at me. Suddenly, Bruce turned and started to run around the table toward me. It wasn't very clear to me what was going to happen, but Bob Jackson and Sam Resch grabbed him, calmed him down, and brought him back to his chair. As quietly as possible, I demanded an apology for the outburst, saying I would report the incident to his management otherwise. Pat Selinger, his manager, was sitting right next to me. This caused several people to chuckle, but Bruce could see that I was serious. He immediately apologized.

I turned to Roger and said, "Look Rog, Rick and I have been over Mel's APPC materials several times now. There just isn't anything that he's brought up that isn't already covered in the DDM spec. Can we at least get you and Mel to read the DDM material? We'll make whatever changes to it that you think are needed for DRDA. That's what we're here for. But let's move on to other material."

Roger agreed to look at the DDM communication specifications, but asked that we extract the pertinent sections from the DDM specifications and distribute them to the entire group. This we quickly agreed to do. Part of our problem at this point was that the DRDA group really didn't know what was in DDM architecture. Because of the file orientation of the first two levels of DDM, they hadn't been interested in wading through 800 pages of detailed technical specifications. The attempts Rick and I had made to educate them about DDM as a general framework for distributed processing hadn't been very successful.

After all of the commotion, I thought we would be going on to the next topic on the agenda, but Mel spoke up and insisted that he be allowed to continue with his presentation. Roger tried to turn him off, but Mel just plowed on into his presentation. Roger quickly gave up and let him continue for his allotted hour. I just gritted my teeth and decided to shut up. My point had been made. We'd finally gotten their attention.

After returning to Rochester, Rick and I carefully went through the DDM specifications and extracted the parts that pertained to DRDA. We formatted these parts as a separate document, still several hundred pages long, but containing none of the file protocols that the DRDA group had found so distracting. Later, as we defined new DDM messages strictly for DRDA use, we added them to the overall architecture document, but we always gave the DRDA group only the parts related to relational databases.

At the next meeting, in Austin, Texas, the DDM APPC Communications Manager was approved by the group. While Bruce continued to push his own agenda for DRDA, he stopped being an opponent to the involvement of the DDM architects. In part, I hope, this was because of the quality of our work, but it was also due to the fact that we were willing to do the tedious work of documenting all the details of DRDA. In some discussions, he was even an ally.

1988-1990: DDM & Les tempéraments DRDA effiloché

Désolé, pas encore traduit.