|
IRS Systems upgrade, $8 bil. and overbudget
http://www.nytimes.com/2003/12/11/business/11irs.html
It appears that the IRS systems upgrade isn't going too well. 1 billion of an 8billion budget spent, and they're already overbudget for what they've done.
I'm not sure what to think of this. On one hand, if you're getting contracts this size, you should (hopefully) be able to estimate cost to within 100% of the final cost. But on the other hand, this is the IRS's systems, and they're not a system I would think you could easily go "oops, sorry about the bug, we'll fix it next version."
Error free software takes lots of time and money. Upgrading old complex systems (without losing that data) does as well. Sounds to me like a *very* tough thing to do correctly.
Andrew Hurst
Thursday, December 11, 2003
OK dude now seriously, everybody go check out the photograph at the top of that article. IRS data processing room photographed in 1963 and the same room in 1997. Check that out dudes. Man that is amazing. Holy cow.
They just don't build computers like they used to, man. FORTY years, man.
I bet when this new system is done it will be slower, less reliable, and have fewer capabilities than the 1963 system they are currently using.
And there is no way their new computers will still be running in 2043.
Dennis Atkins
Thursday, December 11, 2003
On the topic at hand, I am always amazed at how the lowest bidder gets the job and then there is no obligation to keep the cost within that bid. CSC wins the bid at $300 million for the FBI system (which many of us on this board could write single handedly), and then ends up charging the government $700 million and being delivered late. Probably there was some guy bidding $500 million who would have delivered in half the time and on budget but I am so sorry you do not have the lowest bid.
Really, $8 billion for some computer software for the IRS and it's 40% over budget, projected to hit $12 billion I guess. And they wasted $5 billion on an attempt that was abandoned a few years before. That's $17 billion total if it's resolved this time. For the $68 per citizen, or $204 per taxpayer this is costing, they could just buy everyone harddrives and let them keep their own tax receipts at home. Maybe package it in a little portable unit - call it "iTax". But then people might drop it or lose it. I guess it'll have to be on an implantable chip then.
You know what sort of software I could develop for $12 billion?
The IRS should outsource the whole thing to a couple of Romanian students. They could do it for $250,000 and deliver it on time.
crazy homeless person
Thursday, December 11, 2003
Crazy Homeless person,
__MASSIVE HIGH FIVE___
I like your ranting style, you sound like me.
Braid_ged.
Braid_Ged
Thursday, December 11, 2003
"You know what sort of software I could develop for $12 billion?"
I feel much the same way. Im sitting her ewondering what on earth it could be about the system thats so complex.
At base it _sounds_ incredibly straightforward....a database that contains a shitload of information, all of which needs to be added, edited and viewed.
A decent set of backup systems...they would have to have redundancy up to the wazoo but no new problems there surely.
I mean..the system would be _big_ no doubt about that...but difficult?
FullNameRequired
Thursday, December 11, 2003
That is f---ing unbelievable! I don't understand why it costs so much to upgrade their systems. What are they actually DOING with those computers?
I don't even WANT to know. I say we go for the "flat" tax -- no more loopholes. I want to see an income tax return on a postcard. Do away with the entire goddamn institution. What a waste of money.
Roose
Thursday, December 11, 2003
Not to mention that working at the IRS must be one of the more demoralizing things I can imagine. Doing away with that entire institution would put a few thousand people out of their misery.
Roose
Thursday, December 11, 2003
It's the same here in the UK. Government departments happily hand over hunderds of millions of pounds to these cowboy companies, in exchange for projects that are late and go wildly over budget.
I've always suspected it is a classic "Emporer's New Clothes" case. The problem is, little guys saying to the Government "this is all wrong" don't get listened to.
Some of the systems are no more complex than the IT systems I've built single-handedly for clients, say three man (me) years effort. For example, the recent Child Support Agency or Passport Agency systems would have been no more complex than the Insurance Underwriting, Claims Management and Document Production system that I just wrote for a $1B client.
I would love to be able to convince someone in power to let me have a go at one of these projects. I wouldn't even charge 1000 times my normal rate, to make it match the bids they're getting from the cartel of official suppliers.
Steve Jones (UK)
Thursday, December 11, 2003
The two funniest quotes in the whole article are:
"Mr. Everson said this had allowed him to put more I.R.S. executives on the troubled project, although, as a result, the agency had to set aside ancillary modernization projects."
(More executives. Yeah, that's the ticket.)
and
"Mr. Levitan said that Mr. Rossotti brought technological coherence that has averted disaster. But he also says a collapse is inevitable without a new system, because the few people who could keep the old system functioning are close to retiring."
(YOU'RE SPENDING 12 EFFING BILLION DOLLARS BECAUSE A FEW PEOPLE ARE GOING TO RETIRE! TAKE ABOUT 10 MILLION OF THAT AMOUNT AND HIRE SOME NEW PEOPLE AND TRAIN THEM! I'd be willing to take over at least one person's job for the paltry sum of $500,000/yr.)
I wonder how much money CSC makes off of clueless government execs?
bye
Thursday, December 11, 2003
At least one of the government execs was smart enough not to let the geniuses at CSC turn off the old system and then fire up the new system and load 200 million accounts onto it. Can you imagine what a nightmare that would be? It would be total ungodly chaos. We would all be able to go years without filing tax returns and the government would never know.
From the article, I gathered that the government exec eventually convinced CSC to only allow a few accounts onto the system at first and then to scale up from there. I would have thought that any project manager who had every been involved with bringing up a system of any size would have known that you don't throw everybody onto the system at once. I guess even $12 bil isn't enough to afford a decent project manager these days.
bye
Thursday, December 11, 2003
My guess is is that the IRS system is actually made up of dozens of small, complex systems that now need to either work together or be migrated to a new platform. Their system evolved piecemeal over 40 years and probably encompasses every popular business platform since then - Linux, Unix, Windows 3.1, DOS, Mac, and a few I won't even dare guess at.
We're talking about a large organization that deals with millions of people - billions of bits of data every year. Tax law changes, middle managers designing systems to get small things done that were being done by hand with no effort to make it work with any of the other systems that existed at the time.
www.MarkTAW.com
Thursday, December 11, 2003
For $12 billion they could clone just the guys who know how to maintain those old computers. At least that's what I'd do if I wanted to spend $12 billion on the project.
crazy homeless person
Thursday, December 11, 2003
Or just pay them all $10,000,000 a year not to retire.
Steve Jones (UK)
Thursday, December 11, 2003
The explanation for such immense amounts of money for such a "regular" project is simple: corruption.
Every person involved in the deal has his cut.
Those money are OUR tax money, ie everybody's money ie nobody's money: they can affor to spend it.
There is no real accountability. If shit hits the fan, a few heads fall, but that's all.
Nick
Thursday, December 11, 2003
Abolish the IRS, save $12 B (USD), support national sales tax
http://www.salestax.org/
apw
Thursday, December 11, 2003
"Abolish the IRS, save $12 B (USD), support national sales tax"
Speaking as a citizen of a country with a national sales tax, all I can say is 'here, take ours.'
Craig
Thursday, December 11, 2003
What's even more frightening is that the IRS is still recognized as the most efficient tax-collecting entity among modern nations.
I read that a few years ago in an Barron's magazine article that listed the amount of income tax revenue that various countries bring in and then compared that with the cost of running the tax collection entity.
I forget the numbers, but the IRS was waaay ahead of other modern countries. And most of us would view the IRS as the epitome of inefficiency.
Perhaps it's changed in the past few years, though..
Mark Hoffman
Thursday, December 11, 2003
"Speaking as a citizen of a country with a national sales tax...", said Craig.
Which country? Is there also an income tax? How about state/territorial sales taxes? (Just wondering)
Exception guy
Thursday, December 11, 2003
The $12B probably isn't for creating the data store, it's for migrating/reproducing 5 million "Taxpayer/Widget/Noodle per Year" reports that print on greenbar paper.
I wish the IRS would just get it over with. Just confiscate my whole paycheck, and then pay me an allowance to cover cigarettes and toiletries and the occasional copy of Maxim.
Rob VH
Thursday, December 11, 2003
>"I read that a few years ago in an Barron's magazine article that listed the amount of income tax revenue that various countries bring in and then compared that with the cost of running the tax collection entity."
And how much of that is due to the higher incomes in the US? If people earn X% more, more taxes can be collected from them without as large an increase in the cost of collection.
What would be more helpful is a comparison of the amount spent by the collection agency *per taxpayer*.
NoName
Thursday, December 11, 2003
Just a couple of notes, the IRS system is likely running on DB2 on IBM mainframes. The highest concentration of DB2 licenses in the US is just outside of Washington DC. Likely there are some large infrastructure costs associated with buying some new mainframes running zOS and DB2. Then there's all the infrastructure required to link them together seamlessly and with redundancy. Don't forget the DASD systems (likely from EDS) and all the tapes for full backups. Knowing the type of data we throw around and how large our system is, I can imagine they'll probably need 8 mainframes and at least 8 fully stocked silos.
The cost of training a new programmer in COBOL (which I was this past year) is not too terribly high, but the learning curve is significant. It does make sense for a large code base like that to at least refactor if not dramatically rewrite their code. I imagine the system was originally written for VSAM files or even plain text files on the mainframe and has since had DB2 integrated on top of that. Sorting out the one system from the other must be quite difficult.
And on systems like this, where there is a huge amount of data processing going on all with the aim of producing reports, a significant portion of programmer time is spent creating reports. That's probably the most disheartening thing about the system as it currently stands from a programming perspective.
And yes, DB2 on the mainframe, with code written in either C, C++, or a myriad of other language (although those are probably the best choices for the domian in question) is probably the best choice given the problem. Batch processing of bulk data running into millions of records isn't something that you can put on a distributed network very easily or have great speed with. Moreover they have to process all those records in a very short timeframe.
While I think the overspending is rather disheartening, I can see where the costs could be. Having seen some systems that have evolved over time from flat files to indexed files to databases I can say that unravelling the system to determine what piece of code does what is quite complicated, especially when the rules have changed year over year and the code probably has lots of dead ends.
Lou
Thursday, December 11, 2003
The IRS systems upgrade overbudget story is nothing compared to what we got in Canada... 2 words:
firearms registry
A google search will turn up a few gems on this one.
Thursday, December 11, 2003
They have guns in Canada?
'eh?
Thursday, December 11, 2003
The problem is the outsourcer, and these types of problems occur much more often than ever makes it to the papers.
Outsourcers are classic cases of management not being developers or understanding development. They implement rigid hierarchical systems and tend to hire low-cost developers.
When they arrive, they install an army of project managers and account managers who write things, then pass it to developers to build. It's no wonder their work is almost always poor and is always very expensive.
If they let the IRS build the system itself, it probably would have been finished in three months. Instead the shonks line their pockets.
echidna
Thursday, December 11, 2003
That Canadian gun registry is getting more press for sure. More of teh boondoggle is the entire program to begin with since gun control laws increase violent crime everywhere they've been implemented:
http://www.bcwf.bc.ca/s=122/bcw1069915015293/
>November 27, 2003
> Gun laws do not reduce criminal violence according to new study
...
>Disarming the public has not reduced criminal violence in any country examined in this study. In all these cases, disarming the public has been ineffective, expensive, and often counter productive. In all cases, the effort meant setting up expensive bureaucracies that produce no noticeable improvement to public safety or have made the situation worse.
Dennis Atkins
Thursday, December 11, 2003
I can't believe you're slamming outsourcing echidna. It's the wave of the future don't you know!
Dennis Atkins
Thursday, December 11, 2003
" gun control laws increase violent crime everywhere they've been implemented"
come on dennis. _thats_ just stupid.
Violent crime rates in the USA are the _highest in the world_ per head of population.
(with the possible exception of places like..he he...iraq, and afghanistan where there have been wars recently...)
So...tell me _exactly_ how taking away weapons will make those rates any worse?
More than 10,000 people are murdered _with guns_ every year.
Now, take away the guns and its going to be a _lot_ harder for murder to be committed...thats just common sense.
FullNameRequired
Thursday, December 11, 2003
"Violent crime rates in the USA are the _highest in the world_ per head of population."
Yawn.... care to back it up with recent stats?
Dennis Atkins
Thursday, December 11, 2003
"Now, take away the guns and its going to be a _lot_ harder for murder to be committed...thats just common sense. "
What an idea!! Hey! Let's make drugs illegal too! That will solve the drug problem, right? Oh..we did that and drugs are still a problem. Hmmm.....
Even if guns were even more highly regulated and controlled, criminal elements will still have them. And likely use them. After all, gun regulation really only affects law abiding citizens. NYC has some of the toughest gun laws in the country and they are awash in illegal guns. Any crackhead with $200 in his pocket can get one on virtually any street corner.
FWIW, I do support gun control and regulation, but it's intellectually dishonest to suggest that making guns illegal will make them go away. Only in Utopia, bud.
Bang Bang.
Thursday, December 11, 2003
FullName,
Since you are lacking facts and data, allow me to supply them for you. Let's take the most recent year for which we have data.
Homicide Rates, 2001
----------------
US - 6 per 1,000,000 (source: FBI)
England - 10 per 1,000,000 (source: Home Office)
Violent Crime Rates, 2000
-------------------
US - 500 per 100,000 (source: FBI)
England & Wales - 1,400 per 100,000 (source: Home Office)
Property Crime rates
----------------
US - 4,000 per 100,000 (source: FBI)
England & Wales - 8,000 per 100,000 (source: Home Office)
Thus, I have completely discredited your claims as ignorant and entirely nonfactual. Thanks for participating.
Dennis Atkins
Thursday, December 11, 2003
I distrust any statistics on gun control, regardless of which side of the argument they favor.
Pro gun-control bandy England and Japan as great examples of countries with low gun fatalities.
Opponents of gun-control use Switzerland as an example of how peaceful a citizenry can be when they are all armed with fully automatic weapons.
Statistics are easily, easily manipulated to reflect what other side wants.
Bang Bang.
Thursday, December 11, 2003
hi Dennis,
do you have any links?
Ive found this:
"And in all of Britain in 1999-2000, there were only 62 firearm-related murders. By comparison, in the USA, 7,950 homicides were committed with guns in 1999. (The U.S. population is about 4 1/2 times Britain's.) Forty-two of the British murders were committed with handguns obtained illegally. Armed robberies, also with handguns, have increased dramatically."
at
http://www.usatoday.com/news/world/2001/08/07/guns-usat.htm
this also looks interesting (although maybe somewhat partisan)
http://goodsforguns.org/nationalfacts/
<g> theres a bunch more but it doesn't take more than a basic google search to case a little doubt on your figures...
FullNameRequired
Thursday, December 11, 2003
"US - 6 per 1,000,000 (source: FBI)"
this one here I _really_ have problems with.
whats the population of the USA? Im guessing somewhere around 250 million?
6 * 250 = 1500
so that would make the total # of peopel killed by gunds each year
<= 1500
which sounds incredibly low....dont you think it sounds a little low dennis?
FullNameRequired
Thursday, December 11, 2003
people killed by _guns_ I mean... <g>
FullNameRequired
Thursday, December 11, 2003
Dennis: please cite your sources more specifically. I went to verify your numbers and was unable to.
According to Home Office, the average murder rate between 1998 and 2000 in England is 1.5 per 100,000; in the US, 5.87.
http://www.homeoffice.gov.uk/rds/pdfs2/hosb502.pdf (page 10).
Alyosha`
Thursday, December 11, 2003
OK, you guys are right, I've been made a fool of by the disreputable antics of the anti-gun-control lobby and I apologize.
My source for the statistics quoted, which 'claimed' to give FBI sources but 'conveniently' misplaced a decimal point to give a rate of 6 per million rather than the real rate of 60 per million:
http://www.sfu.ca/~mauser/papers/failed/FailedExperiment.pdf
It's rate for England matches yours and I quoted correctly there.
The official US government statistics that the previous so-called 'study' supposedly quoted which matches your guys' figures:
http://www.ojp.usdoj.gov/bjs/homicide/hmrt.htm
Damn. Looks like you're right on the overall numbers.
That study above has been reported extensively in the news recently. I give you permission and this ammunition to sock it to those jackasses who made a fool of me.
Dennis Atkins
Thursday, December 11, 2003
Hi Dennis,
its interesting...you are blaming the gun industry for misrepresenting the facts, but in fact a few moments of thought would have made you seriously doubt them. (it certainly did me).
Statistics are very easy to come by these days, its more important than ever to both double check them (particularly when they come from such an obviously partisan location) and apply the 'common sense' test...ie, how do the implications of these statistics work with what I had previously thought to be reasonable.
Merely a few seconds spent on _either_ approach would have likely saved you in this case...the stats you presented as facts are so easily shown to be doubtful.
ah well...
FullNameRequired
Thursday, December 11, 2003
FullName,
I hear ya. When that study came out a couple weeks ago, I was pretty shocked and surprised by the numbers, but the guy was presented as a reputable professor doing a simple analysis and facts don't lie. He has charts and tables in his report and he specifically says over and over that the US homicide rate, which used to be high, had drapped at a remarkable rate over the last few years, in part due to harsher sentencing laws. Well, it all sounds very credible when presented that way and I fell for it - especially the bit where he gives the FBI as a source for his data! I didn't think to double check his FBI statistics -- since this study was being reported extensively in the popular media and not just some fringe crank study, I assumed that reporters commenting on it would have checked the basic facts. I was wrong. It's may be an honest mistake to misplace a decimal point, or it may be something else designed to look like an honest mistake. Probably we will never know. This information about US homocide rates falling and homocides due to knifings and being hit on the head with beer mugs going up in Britain is very popular in the progun lobby right now, but it does not appear to be what they are saying it is. As it is, I support the right to bear arms, and the very last thing I need is studies like this that are distorting the facts because it completely destroys the credibility of the position.
So thanks for correcting me, I do appreciate it.
Dennis Atkins
Thursday, December 11, 2003
Dennis -
I must say the first graph on page 11 of the pdf is amazingly misleading in its use of 2 scales, especially since none of the subsequent ones do.
Devil's Advocate
Friday, December 12, 2003
Dennis, thanks for making this a quite non-typical gun control discussion by your attitude. I was reading it tonight expecting it to go the way of the open source discussions lately, and was quite surprised.
Also thanks to everyone else who didn't make this a flame fest. Quite a refreshing change after mucking through the rest of the internet.
Andrew Hurst
Friday, December 12, 2003
Devil's Advocate,
Whoa nelly! Didn't you make a good catch.
Boy was I fooled. That is like, the most misleading thing I have ever seen, short of outright bald faced lying. I can't recall ever having seen such a bizarrely misrepresentational graph, and quite obviously it was deliberately constructed that way.
Dang.
I have to give it to bang bang -- he was positively psychic when he said "I don't trust any statistics associated with gun control." I was totally skeptical of him, and yet he nailed it. bang bang, you get like a total high five. Very impressive.
Mark Twain was right. There are...
Dennis Atkins
Friday, December 12, 2003
Andrew,
Thanks. I'm passionate about many things and perhaps even a little harsh in my readiness to argue a point, but I'm also a believer in the truth and wiling to admit when I am absolutely wrong about something.
Dennis Atkins
Friday, December 12, 2003
Stop hijacking this thread with your guns. This is about the IRS and large overbudget systems.
NoName
Friday, December 12, 2003
Yeah but we may need the former if the latter gets out of control :)
Dan Maas
Friday, December 12, 2003
Whether US crime has gone down because of harsher sentencing and "zero tolerance" is also a moot point.
I did read that the much vaunted decline in crime in New York was more or less the same as that in Chicago, where a very different policy was applied.
And the increase in the prison population in the States has created a second version of the military/industrial complex, where so many communities are financially dependent on the prison industry that it will cause considerable disruption, and voter revolt, to try and lower it.
Stephen Jones
Friday, December 12, 2003
"Yeah but we may need the former if the latter gets out of control :) "
LOL!!!!
Bang Bang.
Friday, December 12, 2003
 Recent Topics
Fog Creek Home
| 
