Diary

TUE, FEB 3 '87 9:00 AT WORK, READ THROUGH UPPER CODE [diagram] WORD WPDISP WPSCROLL

fast scroll up/down

Thurs Feb 5 '87

Sunny mostly

7:15 Get up. Breakfast. Go to gym

8:00 At gym. Quiet.

9:45 Go home, shower, etc. Call mom. No dinner tonite.

10:45 Go get cash & bananas

11:30 Call Dr Hills office, after deciding to get work done.

11:45 Lucn hat Pagliaccis. Go to work

1:00 Resync, remake. Finish reading PC Word 3.0 modules for making a schedule, and go thru Works spec to see what other features are needed after primary level is done. Locker MP 2.0 code too.

10:30 Go home, to bed

Fri, Feb 6 '87

7:00 Got up. Good Sydney, Australia stuff on the Today Show

8:30 Do yesterday's mail

8:45 Go to work

9:15 Misc

9:30 Type in schedule

11:15 Give one to Tony. Misc. Ann Thronson calls. Robertson house available.

11:45 Go over schedule w/Tony. He issued email that sets goal for date for full functionality at June 1.

12:30 Discuss Lasbels with Ed

12:45 Lunch, gasoline, car wash

2:30 Misc office work. CR@sync, etc, email, list email. Finish PC Mag on Integrated Apps making list of features that we don't have.

5:45 Resync wwspec. Diff & study

6:30 Micronews. Misc

7:15 Legue. Make & eat dinner. Start Laundry

8:00 Watch Moneyline most of USW.

9:15 Read WSJ, Michael Milkon of Drexel Burnham involved w/Boesky.

C O B

9:30 Read old IBM Cobol Manual. Thought about how much excess Cobol provides beyond what people need for most business DP, which is just a lot of

record input

selection

sorting of selected info

reporting on selected records with breaks, summaries, averages, etc

also some relational operations.

And Cobol is just a record keeping language for businesses. That is what it is good at, but not it can be surpassed at those few things listed above, by other so-called "DBMs" which have less flexibility, but instead focus on what most people need to do most often, for business record keeping:

payroll

inventory

billing

accounting

Yet old languages like Fortran and Pascal survive because they are meant for solving problem which demands a flexible, powerful language. In other words, the classes of problems that these solve cannot be split into other subclasses that are big enough for other languages to focus on.

10:00 Look at Microsoft Employee Price List to see what Cobol is available now at its 2.10. I wonder what that has over 1.00 released by my group. Is that Cobol in C? Or just with Debug? Can't tell from Micronews but I see there that Xenix version came out in May 1985. Is this Cobol in C?

See Bob Zawalich's name on list of developers and since he's working in Apps now I'll ask him.

Also go to library to see it, and maybe buy a copy of 2.10.

During this hour, more laundry, boil eggs

11:00 To bed

Sat, Feb 7 '87

7:00 Get up. Misc.

8:45 Go to work stopping at to vacuum car since B&D won't suck.

9:30 Resync, etc

10:00 Look at keyboard spec (new) and old keyboard code

11:45 Lunch

12:30 Go to library for Cobol 2.10 manual but they don't keep it on shelves, so stop by Steve Snydov's a bit. (Also pickup early 1986 brochure on Microsoft Lanuages from library.) Back in office find original press release on Cobol 2.0, Jan '85. So I know the Cobol I saw to market was mainstay for 83 & 84! And the team I put together in 1982 built the new Cobol 2.0. (I never really thought much about this.) Also added list of other systems achievements since I left in 1983. And last I cut out Cobol review from Nov '86 Computer Language.

2:15 Resync wwspec & diff word. Not much. Resync ww again, nothing of interest.

2:30 Back to keyboard spec & code. Looks much better! Great getting rid of End-Mode, and aligning with WP. Next make small change to ward.c.checkin. Ask Jabe questions about spec.

5:00 Ask Ed, and eventually Jabe, about feature to turn off background pagination. Code changes for Ctrl-Pg Up/Down & Arrow.

6:45 Resync DOSENV & WINEVV

7:00 Go home. Buy groceries. Make dinner.

8:00 Eat during "Glenn Miller Story." Good move. After first intermission, I left message for Mark Thompson. Watch rest of movie. But third quarter was just music, so I reread that Cobol review a bit.

10:30 After the movie, I finished review. No compiler product seems to stand out.

11:00 Reviewed the Cobol materials from 1982 and early 193.

12:30 No change in my opnion: I did good manging cleanup, building new team, and passing on the ball after I lost interest.

12:45 AM And why did I lose interest? After getting the 8080 & 8086 code cleaned up, and delivered, the things left to do were

1) Reimplement for big memory, and this meant recoding compiler.

2) Adding more features to make Cobol full "high ANSI"

No. I didn't appeal because that would just be another Cobol, and I wanted to make something new and more special. But what? The key is that to do Cobol one better, one can't just add features. One has to understand the needs of the market which is business data processing:

inventory

payroll

billing

accounting

customers

market research

=> DBMS

And that's transactiosn processing, forms design, data entry, sorting, querying (selection), reporting. None of which turns me on. Perhaps the solution would and implementing it may, but designing it doesn't. I'm not talking about designing one application. Clearly Pers acct fascinated me. I'm talking about designing a DBMS system.