Slashdot Mirror
California Cancels $208 Million IT Overhaul Halfway Through
g01d4 writes "According to the LA Times, 'California's computer problems, which have already cost taxpayers hundreds of millions of dollars, have mounted as state officials cut short work on a $208-million DMV technology overhaul that is only half done. The state has spent $135 million total on the overhaul so far. The state's contractor, HP Enterprise Services, has received nearly $50 million of the money spent on the project. Botello said the company will not receive the remaining $26 million in its contract. ... Last week, the controller's office fired the contractor responsible for a $371-million upgrade to the state's payroll system, citing a trial run filled with mishaps. More than $254 million has already been spent.' It's hard not to feel like the Tokyo man in the street watching the latest round of Godzilla the state vs. Rodan the big contractor."
I'm glad to see that they didn't fall prey too badly to the fallacy of sunk costs. Too many places wouldn't realize they've already lost the money they threw at the project, and no amount of extra spending in the hopes that it will eventually succeed will get that back.
In fact, this very scenario has happened a decade before, albet with Oracle instead of HP.
There's no -1 for "I don't get it."
you would think there would be a lot of good expertise in the computing arena for the state to tap in to.
Ahahaha, with our government? If they even had the slightest idea of how important it was to stay out of the fucking 1980's with IT equipment that serves critical functions for the state and its citizens, they wouldn't have waited for the problem to "cost taxpayers hundreds of millions of dollars" to do anything about it.
If they can't get that much straight, how can they possibly hope to know what technical criterion to look for when hiring contractors?
HP screwed over Vermont: http://governor.vermont.gov/newsroom-Vermont-HP-reach-DMV-settlement-gov-shumlin in its attempt to redo the VT DMV.
Of course, we end up paying for the incompetence that drives the grossly misnamed Department of Information and Innovation...
It's not about the government. It's all about the useless IT consulting companies. Pretty much every single flashy consulting company billboard/AD that you see at an airport is just a way to milk the gullible and not deliver. This is an across-the-board problem. Nobody wants to fucking do their jobs. The government thinks they don't need the right people to do it, so they hire a contractor. The contractor doesn't want to do the job either, it's not their core competency (nobody knows what it is anyway), so they hire subcontractors. Subcontractors have very little vested interest in anything, and they maybe deliver, maybe not, but due to multiple layers of clueless management, it's of no use anyway. So there you go.
A successful API design takes a mixture of software design and pedagogy.
We see this all the time in the military. A low estimate is given on a minimally speced out project. Then as the project money is spent, the agencies go back to the congress and ask for more money, saying we already spent this money, and it won' really work the way we need it to. Instead of firing the con artists, and suing the contractors, and accepting the money as lost, we fund it more thus encouraging the fraudulent behavior.
"She's a scientist and a lesbian. She's not going to let it slide." Orphan Black
"The contract was awarded in 2007 to the Texas-based Electronic Data Systems. The company was later bought by Hewlett-Packard and renamed HP Enterprise Services. Hewlett-Packard is now run by Meg Whitman, who during her failed campaign for governor in 2010 promised to save California money with better computer technology."
I smell something going on here. I'm thinking this may have been a bit too convenient.
I thought it was an "emoticon" of a large breasted woman giving the finger.
W..w..W - Willy Waterloo washes Warren Wiggins who is washing Waldo Woo.
This is by far the best line of the article....
"Hewlett-Packard is now run by Meg Whitman, who during her failed campaign for governor in 2010 promised to save California money with better computer technology."
This is entirely normal when you take a government that chronically under-staffs on IT and relies on consultants. They go and try to do something big, and they don't have the expertise in house to deal with it. Enter more consultants, particularly of the variety that like to write a lot of powerpoint presentations and bill a lot of hours but never actually deliver a bloody thing. Of course, since the government doesn't have enough IT expertise to actually figure that out, the high level senior managers that love powerpoint and high-level mumbo jumbo MBA talk think everything is going well.
And then, scope creep happens. It follows one of three lines:
1. Election happens. New government comes in, with new priorities and a new way they want to do things. This is obviously bad for a huge project in progress.
2. The existing project has a new department join in, which means new managers and thus a new set of demands. Instead of starting up a new project, they try to shoehorn those into the current project to satisfy management's desire for design by a giant committee of managers.
3. Someone realizes that the project didn't actually have all the requirements properly captured in the first place, which is pretty much inevitable in my experience.
You'd think at some point the government would learn that they can't manage projects in this way and rely on consultants to sort it out, but they never do. Of course, in the case of #1 or #2 even in house IT doesn't really save you, but in my experience they tend to be more flexible than a giant Enterprise consulting outfit (mostly because there's no contract they can hide behind to deliver X, even if X doesn't actually solve the problem that prompted the project in the first place).
The whole process is a giant mess.
-- "So they told me that using the download page to download something was not something they anticipated." - Bill Gates
how important it was to stay out of the fucking 1980's with IT equipment that serves critical functions
Talk about blanket statements. I suspect that there is quite a bit of 1980s IT equipment in your life that you are not even aware of.
The problem is not what decade the equipment comes from, it is whether or not the equipment meets its requirements. If equipment from the 1980s is continuing to meet the requirements that governments face today, then there is no reason to spend enormous amounts of tax money to replace that equipment unless doing so will pay for itself before the next upgrade. Unfortunately, there are few cases where such upgrades actually do pay for themselves, so in terms of what is best to do with tax dollars, upgrading old equipment that continues to function as needed is questionable.
Now, if the equipment is not working, then it is time to replace it. The real problem is that government contracts are not typically given to companies deemed best for the job, and so these situations arise. Contracts are awarded to companies that bid low and to companies that are well-connected, even when better companies are available.
Palm trees and 8
>"The DMV project began in 2006, according to the California Technology Agency. Instead of using 40-year-old, "dangerously antiquated technology," DMV staffers were supposed to get a modern, user-friendly system that minimized the risk of "catastrophic failure," according to a DMV report on the project."
This encapsulates solving multiple problems at the same time. This cannot be done. You update large systems by plotting a path through incremental improvements that get deployed, tested and fixed before the next increment, so that get you to where you need to be. It might not seem like the optimal path, but anything involving a switch over of technology, UI, back end, infastructure and buckets of code all at the same time is simply never going to work.
In the case of the DMV, it might involve unifying disjoint databases pair by pair until you have only one, while maintaining the same interface to the heterogeneous clients. Then one by one converting the heterogeneous clients to a standard back end interface. Then one by one adding the features of a client to the grand unified client and switching over that system, until the GUC has all the features for all the clients are new client. Then one by one, updating the organizational procedures to make them better, and updating the GUC while doing so. You can make these changes one by one. You can roll back one step if it doesn't work right the first time. You can measure progress by the number of working updates, not in how much less non-working the global-replace-systems is today.
I should use this sig to advertise my book ISBN-13 : 978-1501515132.
Having worked on govt projects before, it's all spent on :
a) Management. Lots of it. About 5 times as many managers/sub-contract managers/advisors etc than there will be coders. Because the more management a project has, the harder it is to blame any one person.
b) Paper. Lots of paper. The amount of pages generated on specifications, revisions, reports, recommendations will be able 10 times the number of _lines_ of code created. All to show that no taxpayers money was wasted.
c) Tendering. It costs a lot to tender a bid, which reduces the competition to only the big ones who can afford to throw a million at a 1in5 chance. Whereas, if they were allowed to go to a small consultancy who only has 30 employees, they'd be able to get a much better price.
d) Changes. The requirements are often so written in very complex language that noone really understands it, and then they come along with changes every 2 months which require 3 months of recoding because they didn't fully understand what they were asking for to start with.
e) User acceptance. Don't underestimate the ability of a low level govt employee to refuse to use the new system because 'I've done it this way for 30 years and it worked just fine! This doesn't work like the old one did.'
Oh good grief what a bunch of bunk. Do you know why big contracting companies exist? It's because companies have lost faith in their internal organizations to deliver. Plain and simple. I've seen it in dozens of organizations in my career where an entrenched group builds a castle of "can't." So, the execs hire an outside firm not only to get what they want but to also force these little castles to actually deliver something. Lazy contractors? Please, how about lazy employees who feel that they hold the keys and as long as they keep pushing back and feel empowered so nobody is going to mess with their careers even though they may be doing such innovative things like writing System 360 Assembly Language. It's everywhere and it's not just in IT, people in this country have become lazy and foolish relying on attitude rather than customer service and trying to do a good job.
Every large organization has this problem and IT isn't just one of the areas where dead space can occur. So, the big companies come in, push change, make big promises that sometimes are overblown. In the case of California I could probably guess that the specifications of what were required were done by bureaucrats who have no clue on how to spec out requirements or were based on something that wasn't possible to build. Are there bad firms? yes, but are there bad customers? hell yes and they can make it absolutely impossible to deliver anything because the same people who have to approve or test anything are usually overworked, or not committed to the project leaving the contractor and subsequently the whole project in limbo. That usually leaves to failure despite the best efforts of all parties involved.
So before you blame consulting companies for this failure, remember they wouldn't exist if people were doing their job or came to the realization that their skills and abilities are out of alignment with what their management expects.
Harrison's Postulate - "For every action there is an equal and opposite criticism"
I'll never look at that word the same way again :)
Therefore, by the (faulty) logic you're using, you're just a cow with a keyboard - osu-neko (2604)
Could it be that the way the government contracts are structured and micromanaged by government agencies is the problem and not the contractor or their programmers? I work for a company that provides government services under contract to the State of California and the government agency that oversees us micromanages us so much that it is often impossible to to develop systems properly. The 4 biggest problems I see are 1)constantly changing requirements that are written by government employees with little or no IT/web knowledge 2) contracts secured by being the lowest bidder which do not allow us to have the resources to properly design or test the system we are building 3) forcing us to work with other contractors including non-profit ones that are "donating" their services (very strange to me really) and that provide inferior IT systems we must use or integrate. 4) Requirements, features and design being dictated by government agencies or advocacy groups with little knoweldge of system design & development. For example, we are currently forced to support an application written by one of these "non-profits" that uses ASP classic and violates every current IT standard. My company has the IT staff & talent to completely rewrite the application but we are not allowed to and must instead support and integrate the badly written one that was donated to the state. It is unclear why this non-profit is allowed to force the agency & us to use their product, but it seems they have political connections that make it so. I believe also that government contracts almost always go to the lowest bidder and not the company with the most expertise. Often a contractor is the lowest bidder because they plan to cut corners and not follow good IT practices, or have not estimated costs correctly. Also as a web developer for a company that works under government contracts, I cannot count the number of times we have received requirements for a website from people that have little or no computer skills, let alone web skills or experience. You would think in this day and age that the government employees providing requirements for government IT systems would have at least basic IT knowledge, but this is often not the case. I am not exaggerating that I have received requirements from people that have no Excel, Word or even email skills and have obviously barely even used the Internet. Many people in the top levels of government management are older (baby boomers) or were promoted for reasons other than great IT skills. They often have no professional experience with developing IT Systems, ADA or other required standards and yet they are the one writing the criteria for the contracts and the system requirements. State agencies also often demand that large amounts of money be spent on "usability studies" or other commitees where a lot of people discuss and dictate what the IT contractor should do in building the new system. The people running these studies often have very poor IT skills themselves and have little experience designing IT systems, but they often have an enormous say in how the system is designed. By the time the IT contractor's development staff is involved in the project, everything is already specified by non-IT government people and between that and the contractor management trying to save every dime (therefore not providing resource for testing), it is not really possible to build a quality system. I say all of this inspite of the fact that the State of California actually has a good Department of Technology Services that provides great ADA compliant web templates. The California State government is so large that even with a good DTS department, the management and staff at specific agencies providing the requirements for a new system may have no knoweldge or interaction with that department and never involve them in creating the contract or project requirements. I think the solution to this is the state should be involving its DTS department in creating all contracts and requirements for new systems projects and ind
There isn't a DMV per municipality. There are about a hundred and sixty scattered around the state.
This has been a complaint too. For a state with a population of 38 million is (figure half of them drive) it's about 119,000 per DMV office. At 250 working days a year, that's 475 per office per day, or about 59 people per hour. That shouldn't be that hard, but the lines there are typically 1-3 hours long. They have a reservation system where you can make an appointment in advance. But the last time my registration came up for renewal, there was a problem which required me to visit the DMV instead of renew by mail. I tried to use their reservation system, only to discover that even though I was trying to make an appointment the day after I got the mailed notice, all the nearby DMV offices were booked solid until 3 weeks after the renewal deadline. I ended up making a reservation at some DMV office in the desert 70 miles away (still had to wait in line 45 min), and used the trip as an excuse to do some sightseeing and visit some friends in the area.
If you have a AAA membership, that's by far the best way to get your DMV stuff done in California (if it's a service they can do - they don't do driving tests and a few other things). I've never had to wait more than 30 min there, and usually they get to me within 5 minutes. They charge a few dollars more, but it's worth it compared to wasting several hours at the DMV.
In contrast, the RMV in Massachusetts and the DMV in Washington had wait times very similar to the AAA. Massachusetts even puts offices in the mall so parking is convenient and you can drop by while getting other shopping done. So I dunno what California is doing wrong, but whatever it is they're doing it very, very wrong.
There are some cases also where this misalignment of skills and management expectations is more of a management deficiency. Many organizations have technical people who are quite willing and capable, but they have been pigeon-holed and beaten down by policies which incentivize apathy. I have worked with long-time developers in quasi-government jobs who have skills only on legacy systems, and I have had the pleasure of helping them participate in the development of modern SOA interfaces. Most of these people just need an opportunity to learn, grow, and feel like their contributions will be meaningful. And it is not expensive, in fact if it looks expensive you are doing it wrong! You don't need or want conventional "training" for them, and if it is done right, it can cost little or nothing in extra time. These people have a goldmine of lost productivity to tap into -- productivity that poor management has beaten out of them and that good management can cash in on.
Strange things are afoot at the Circle-K.