» Politics and other things

  1. Vitalogy

    Oregon sues Oracle Corp. over state's troubled Cover Oregon health insurance exchange,

    This should be a pretty simple outcome. The vendor failed to deliver what they promised. Oracle should lose and I hope Larry Ellison gets to clean out some pit toilets.

    Posted on August 22, 2014 - 07:47 PM #
  2. skeptical

    Even if Oregon doesn't win, Oracle has a black eye for being a half-ass provider -- willing to do shoddy work if the state doesn't supervise.

    Posted on August 22, 2014 - 08:20 PM #
  3. Scott Young

    I can't help wondering if there isn't more to this than meets the eye. I'm reminded of a fantastic book about the Apollo program, "Angle of Attack: Harrison Storms and the Race to the Moon." North American Aviation had the contract to design and build the Saturn V rocket. They suffered greatly and were constantly behind schedule, largely because the payload kept changing, which was beyond North American's control. Is it possible something similar happened to Oracle? I have no idea, but I can't help wondering...

    Posted on August 22, 2014 - 08:34 PM #
  4. missing_kskd

    It is hard to know, but totally valid question.

    I really don't think much of Oracle, for a lot of reasons related to painful projects I have been on, but the higher end people should have nailed this cold.

    So would have our tech startup crowd here, a growing thing and capable like Silicon Valley people are.

    Next time, Oregon should open these things up more, consider open source tools and fund real geeks to do it. I think we would get more for our money personally.

    That said? I don't think we know enough. This mess will have to play out in the courts.

    Posted on August 22, 2014 - 08:49 PM #
  5. skeptical

    We weren't going to the moon here. We were just trying to sign people up. How hard can it be? Clearly it was too much for Oracle.

    Sadly there's no smoking gun. The state is instead going to have to make do with 102 million used shells.

    Posted on August 23, 2014 - 12:41 AM #
  6. missing_kskd

    It's not. Oracle could nail something like this. There were definitely communication problems and priority problems.

    I want to understand the project management and lines of communication and responsibility before making any determination.

    Back in the early 2000's, I was part of an effort to get the State of Oregon to consider open source software for stuff like this. Unlike commercial software, OSS doesn't have funded advocates, generally is shut out of expensive to comply with State RFP / RFQ (request for quote / proposal) policy, and the vast majority of our legislators do not have the technical education needed to understand how and why that matters and what the economic value proposition really is.

    A few do though. I was surprised.

    From time to time, these proposals see traction, and so far, expensive closed software lobby efforts have shut them down.

    The example I used back then was the large amounts Portland spent a couple times on water system billing. One of those was something like 40 million dollars!! I knocked out a plan, and let me contrast that to what happened:

    Portland entered into a contract, and at the end of the contract, was on the hook for ongoing improvement, maintenance, bug fixes, and didn't get the source code or any real ownership of the system. Essentially, somebody wrote themselves a golden check and got Portland to agree to it in return for improving water billing.

    The truth is billing sucked really hard, and it was painful, and there were, and sometimes to this day, continue to be very expensive errors.

    My plan was to take that amount of money, employ open software to avoid nearly all required licenses, and fund a group technically competent people for two years to do what the contract firm did.

    Their charter was a bit different. Rather than simply have the goal of correctly billing for water, use the funding and OSS to create something that is more sustainable, employ people from Oregon, and get it done.

    Once done, the project would staff down to an appropriate salary level so that it could be maintained, improved, and most importantly, exported to other cities. Or, the group could begin to tackle other projects too.

    Portland could just generalize water billing, and share that project with other cities, who could employ somebody, or simply let PDX do the hard work, in return for some compensation.

    Projects like this, done as cities need them, would result in a fairly large and capable body of software created by and for municipalities, which would generate employment, build real public infrastructure useful for future projects, and save States, cities, etc... an awful lot of money, while leaving them in control over their own future.

    It remains cheaper to do that, than it very frequently does Oracle or other big software company fashion. Others advanced similar ideas.

    House speaker at the time, Minnis, shut this down and would not bring it to a vote. I saw the guys in very expensive suits, who looked at us with great disdain, who ended up making some promises, threats and buying a few great meals.

    When Minnis shut it down, we had bi-partisan support, and that bill would have very easily passed the Oregon House. We had most of the Senate, but did not see an opportunity to get the work done there either. Again, despite solid bi-partisan support.

    Once the legislators were well educated and advised, they understood how things could go and were motivated to do a project or two. It's a win all around.

    When I see this today, I can't help but think perhaps it's time for a similar effort. Insurance communication, billing, qualification, etc... software as a service should be mostly open, easy to adopt, and maintained in the fashion I mentioned above. No brainer.

    Doing that would work, and it would open the door to the Silicon Valley types, who can and do use OSS well, and who have the skills to nail something like this cold.

    Posted on August 23, 2014 - 08:16 AM #
  7. Vitalogy

    If I hire and pay a builder to build a home for me, I expect the finished product to be what I paid for. Same goes for Oracle. The state of OR is the customer, Oracle is in charge of building it. How Oracle can lay blame on the customer is beyond me.

    Posted on August 23, 2014 - 11:56 AM #
  8. missing_kskd

    Here's how it can happen:

    You spec a building or home. They do their planning, secure materials, allocate time and dollars. Some of those might be your dollars, required up front, depending.

    Construction starts. Many elements, such as the foundation, really do need to be established prior to finished work, and the work in progress, as well as many specified features have dependencies.

    Say the project is 3 months.

    One month in, after many dependent features are done, you redesign the spec, and there are costs. Both parties can agree to modify costs, but there are now unknowns, unless sufficient planning time is taken to understand what needs to be undone, what can be kept, what has to be modified, and what needs to be new are all made clear.

    Say the 3 month date has a little flexibility, but not a lot. Say you plan on moving in, or perhaps launching a business to which you've committed time and material investments that will prove expensive, if the business isn't equipped with a building to receive them.

    That planning time might, in and of itself, prove deadly to the project.

    At that critical point, a decision needs to be made:

    1. Roll back the changes and continue with original spec, perhaps firing off a secondary project to meet new requirements. They would not fail in this case, unless it really is their fault.

    2. Perform rough planning, which will seriously increase risk, and let cost run free in order for the contractor to meet the date. They could now fail and it could be according to mutually agreed upon terms, which the project manager would need to account for. Both parties could fail, and both could be at fault.

    3. Terminate the project, cutting losses. This is a fail, but a cheaper one.

    4. Do the planning anyway. There is some failure, but the whole thing is "pot committed" and whatever costs are associated with moving the date and changing requirements are going to be accepted and paid regardless. In this mode, both parties could succeed, but will have to account for cost overruns and time.

    Etc... There are other choices here, but that shows how there can be a failure on one or both sides.

    A secondary complexity is whether or not the project as defined will actually meet requirements. Oracle needed to advise on this, and potentially just didn't do their due diligence, accepting the project, which may well have been doomed to failure anyway. Their fault.

    (this is one I check religiously BTW It is extremely painful when it happens. And it's painful in that, "but you got what you paid for" way, where the customer really will say, "but it doesn't meet requirements, why didn't you advise me, or turn the project down?" All hell, all the way down when that happens!)

    This is why I really want to understand the project management, lines of communication, change protocols, research, requirements specification, etc... before making any judgement. There could be fault on one or both sides here, and I suspect it's both.

    My gut says Oracle saw a nice project and it would make their services numbers, and it's from a State, so there is "slush room", and that typically means you bid like crazy to get it, promise stuff, and then work it all out along the way. (high risk of failure, but good revenue --extremely tempting to the services sales team, who should deny a deal like this, but won't)

    And the other half of that would be Oregon not actually thinking requirements through, making changes and managing their vendor poorly.

    Both sides are very highly likely suffering from a lack of planning and established communication protocols.

    So that's my guess. Who knows until discovery digs the goods up for us?

    Want some popcorn? It's going to take a while.

    Posted on August 23, 2014 - 12:27 PM #
  9. Andrew

    Vitalogy: If I hire and pay a builder to build a home for me, I expect the finished product to be what I paid for. Same goes for Oracle. The state of OR is the customer, Oracle is in charge of building it. How Oracle can lay blame on the customer is beyond me.

    What if you change the specs on the house as they are building it? After they have put in the foundation, you tell them it needs to be 200 sq. ft. larger? Oh, and you need to add another story?

    I'm not saying I know Oregon did this a lot and is 100% to blame, but I suspect there was some of this: moving target specs or specs simply not defined until late in the process. I am fairly sure this is some blame on both sides. Is it 50-50? 90-10? Not sure, except that it is probably "he said, she said" from here on out, and we will each have to choose which side to believe.

    Posted on August 23, 2014 - 12:49 PM #
  10. missing_kskd

    Posted on August 24, 2014 - 02:45 PM #