Forwarded from: matthew patton <pattonme@private> > "It's hard to run multiyear projects when money is doled out year to > year," Balutis explains. "The biggest difficulty is that you do a > plan and then all of a sudden you're $50 million short." I was intimately involved in the annual DoD budgeting process as a USAF officer at the Pentagon and have to agree that maybe it's time to put the budget on a 3-year cycle instead of an annual one. Congress turns over staff every 2 years, the Prez every 4 so maybe 3 years makes sense? Obviously when things are as politicized as they are and Washington has millions of self-serving beaurocrats and contractors feeding at the gov't trough without regard for the fact that it's the people's money they are playing with, fraud and waste are ubiquitous. Almost never do the best and brightest fill gov't posts so the quality of management is always of poor quality - the ability to kiss the right asses is what makes for a successful administrator. Having to fight political wars to keep the money rolling in for a long-term project distracts from the management thereof and lowers the worker-bees' interests in doing a decent job in the first place. It's all fine and good for the FBI to lay some or a large share of the blame at the ex-CIOs' doors and indeed the FBI was grossly lacking in basic project management skill. I worked for SAIC on Virtual Case File too. Thing is, VCF was a multi-year project and it was funded as such. No, the ink wasn't technically dry on every year's congress-critter allocation but there was almost no doubt about it being funded year over year. As convenient as it is to blame the FBI for VCF's failure, the blame more squarely belongs on SAIC's shoulders. Even if the FBI had the best project managers the world has to offer, bad design, poor programming skill, and an attitude of "make-work" on the part of SAIC is why VCF was such a boondoggle. Good FBI project managers can not eliminate the problem with SAIC's failure to manage their own people. VCF didn't fail lack for specifications. I've personally read all 3+ inches of program specifications that the FBI and SAIC signed off on. Unfortunately, the people who wrote the specs on both sides and those who read and blessed them weren't very smart nor frankly very good at their jobs. Page after page of stupid and inane things were specified which would only hamper and interfere with the product. Like other naive specification documents that plague IT efforts, it frequently tried to dictate the 'how' instead of the 'what'. SAIC failed to examine and study how the field agents actually worked in real life and take into consideration how much VCF deviated from that daily practice. FBI agents aren't geeks. Yet geeks design things only geeks can love and then wonder why the rest of the world thinks they're nuts. SAIC's data-analysis team was poor too, making all kinds of mistakes in entity relationships and failing to think thru the product enough to spot some of the traps they were setting for themselves. I plastered their data-diagram with stickies pointing out their errors. When a contract operates on a cost and materials basis which is what VCF was, then it's open season on the budget and accountability goes out the window unless you've got some SERIOUSLY good managers on the gov't side. The contractor has absolutely no economic incentive to do well or act responsibly. When I was on the project SAIC had 200+ people, most of them programmers doing practically no work. There was a lot of water-cooler angst over the C programmers getting let go in favor of the Java ones because maybe management had changed their mind about which language to use. There was a whole pizza party/pep rally one day to settle the nerves. Programmers are not cheap, and idle ones less so. Yet the FBI was paying probably at least 1.5x their salary (the general DC cost multiplier) to produce nothing. And this is a full year into VCF! Given the immature status of VCF in August of 2002, the SAIC team should have been about 2 dozen people at the most. A dozen bright engineers of varying disciplines needed to get locked in a room, slide in the coke and pizza, until they figured out all or at least most of the angles before the minions are recruited to sling code as needed. SAIC didn't have 2 dozen bright engineers and they hired the minions many, many months before the project was even sketched out. Instead they were trying out different GUI's and button colors, icon screen placement and trying to get the FBI to sign off on it without having any notion of what they were supposed to accomplish. IT systems in general and in particular of the scale and varied clientel that represents the FBI, require many iterations before getting reasonably close to a workable model. Iterations are cheap when it's pretty much all on paper and only costing the salaries of 20-odd people. But those kinds of numbers don't impress superiors who are looking for profit. Superiors want to see head-count. They want to see lots of zero's in a row on the monthly invoice. Afterall, if there is 50million in the pot they damn well want every last panny. 20 guys spending weeks or months laying and relaying the groundwork isn't likely to suck up even a tenth of that. And what of the FBI who asked for 50mil and so far has only spent 10? Congress is going to come right back at them the following year and say, C: "well, you only spent 10 last year and you want 50 this year again like you asked for last year? Hell no, you get 5." F: "But we're starting implementation!" C: "Use the 40m in the bank and get lost." F: "But we're going to need the 40 and then some" C: "like we care" Congressional budgeting is a disaster and will likely remain so. Any entity that doesn't burn every last penny every year will have it's budget summarily sliced. Extenuating circumstances? One-time or recurring cost reductions? Not on your life. Gov't doesn't reward thrift or wisdom. Never has and never will. Instead it encourages waste, neigh mandates it and penalizes those who don't. Afterall, it's somebody else's money so what do they care. So why should contractors behave any different? VCF should have been a fixed cost contract with rewards for quality, thrift, and achievement but congress-critters don't tolerate that kind of discretion or innovation and they don't even begin to know how to handle agencies having money left over. Not to mention a pissed-off contractor can trivially file a law suit and try to get a court to give them what they think they deserve even if they don't. Whatever the case, the FBI desperately needs to find a project manager with some clue and hefty clout. Frankly Congress and the FBI, or better yet the GAO should fine SAIC a 100 million. Afterall, the GAO has been on SAIC's case about VCF for several years running. But when "accountability" is defined as making the statement "I am accountable" yet failing to resign or appologize, or biting a quivering lip in a TV interview and "feeling your pain" how are things going to change? Congress has never been about having the balls to do what's right. It's far more lucrative and expedient to coddle incompetence, accept donations from grateful contractors to better cement one's power and status, and perpetuate the corrupt and unaccountable system. Those of us who care either get co-opted by the system, give up and leave, soldier on and try to ignore the corruption, or get booted out the door by daring to question and confront the powers on high. The VCF trainwreck could have been halted in the fall of 2002 if anybody cared to listen to those who said it already was a mess. Competent management by both the FBI and SAIC could have backed the problem up another 6 months if not prevented it in the first place. Alas, nobody will ever learn. The faces on the congressional panel will change, the faces of the accused will change but nothing short of a free market or the elimination of free money will actually improve the situation in Washington. _________________________________________ Bellua Cyber Security Asia 2005 - http://www.bellua.com/bcs2005
This archive was generated by hypermail 2.1.3 : Thu Mar 10 2005 - 01:55:13 PST