Wednesday, September 01, 2010

Reactions to the MIDAS Project-I (Old History)

I've listened to the first hour or so of the first day's program introducing the MIDAS program for FSA.  I've some reactions to it, which may spread over multiple posts.

One of the speakers, either the Administrator or Mr. Hanley, said MIDAS is the first time FSA has looked in depth at its operations since 1985.  One of the problems with government bureaucracies is, because of the turnover of political appointees there's a loss of historical memory. ASCS/FSA has a long history of looking at itself, either singly or in conjunction with other agencies.  The successes are few and far between.  Let me reminisce:

The first project I ran into after joining ASCS was a data project, led by Alan Morrison.  Didn't go anywhere.

Then in 1970-71 we had the Management Analysis Project, which looked at all ASCS operations across the board.  I got roped in pushing papers around and managing the library of documentation.  I'm sure some changes were made in some processes because of the studies, but not much. It was notable for establishing task forces on various areas without any plan for how to handle their recommendations.

Next, a couple years later, came Bill Ruble's project to automate county offices by putting dumb terminals in county offices, hooked to the mainframes in Kansas City.  A county office in Mississippi got some terminals and started data loading before it blew up, partially because of cost, partially because people were concerned about privacy of data.  (The Privacy Act was passed in 1974, I think.) This was back in the day when the disk drives were the size of an air conditioning unit and held 7.5 megs.

The next project I remember was Jim Dimwiddie's successful IBM System-36 project, which started putting computers in county offices in 1985, based on work done in 1984.  I was on one of the task forces.  You've got to give Jim a little credit for putting over the project; you've got to give county offices a lot of credit for surviving the hell we put them through.  (Lesson learned: never do a big automation project at the same time you're implementing a new farm bill, particularly when Gramm-Rudman-Hollings reductions come into play.)

Once we had System/36's in the offices, we almost immediately started running out of space. I think the first ones had a hard drive of around 200 megs for the smallest one.  Big concerns about our outgrowing the System/36 led to the Trail Boss effort of Chris Niedermayer.  ("Trail Boss" was a GSA concept, with OMB and GAO approval, for handling big automation projects.)  He started work around 1989 with a big team.  The methodology was "information engineering", as embodied by James Martin.  It was an elaborate, well planned effort, which took so long it produced little, because when the Dems came into office Chris had antagonized some, and his patron had antagonized most.  We did get some mailing software done, which was useful.  (Lesson: big projects tend to run out of impetus and support before they produce results.) (Chris bounced up to USDA It shop, and most recently is deputy CIO for HUD.)

Secretary Madigan in mid -1991 initiated the Department's Info-Share project.  Part of the effort was to consolidate county offices and co-locate the USDA offices in each county (SCS, ASCS, FmHA, Extension). The other parts of the effort were trying to share data among the agencies and to make them work together.  I was involved in pilots with county offices in Kansas and in Mississippi where the different agencies shared PC's hooked to a Sun server and provided on-line access to the handful of farmers with PC's. (Note that Madigan's effort was happening at the same time as the Niedermayer effort.  Madigan got the bee in his bonnet without knowing what was happening in ASCS, or in SCS for that matter.  Lesson: in a place as big as USDA it's hard to do change rationally.)

Info Share stumbled in the summer and fall of 1992, then went into limbo as the new administration took over. After a while it resumed for a while, but without much success.  By this time the new buzzword in IT systems was "business process reengineering".  Rather than looking at the data and getting it rationally organized, you looked at processes, figuring the data would handle itself. (Lesson: big IT projects tend to fall victim to the latest style being pushed by the private consultants hoping to make money off the government.  As ex-Sen Simpson might say: to suck the government tit.)

In 1994 Congress reorganized the agencies, changing SCS to NRCS, splitting FmHA between ASCS, now called FSA, and the new Rural Development, into which the Rural Electrification Administration was folded.

After Secretary Espy left and Glickman came in, Greg Carnill became leader of  what was now called the "service center" effort.  This evolved in part to an attempt was to combine the administrative support for NRCS, FSA, and RD into one organization. Unfortunately this was killed in Congress due to opposition from partisans of the different agencies. (Lesson: any attempt to rationalize USDA organization must win the support of the appropriate Congressional bigshots and the special interests who whisper in their ears.

Before I retired we  did get started on some changes to the FSA name and address system which were worthwhile.  (Lesson: if a bureaucrat leaves, the successors will have their own ideas.) And the people who worked on GIS for FSA/NRCS did implement the common land unit layer.  Whether what was accomplished actually helped the county offices and the farmers they serve is an open question.

No comments: