Should we add the ChryslerComprehensiveCompensation project to SuccessfulXpProjects? Is it considered a success? Using what criteria? --JasonNocks
A project can't be a success unless the people who started the project end up using it and are moderately happy with it.
CCC did not go into production in its first month. Hence its rollout issues were NotAnXpProblem, because it dropped the FrequentReleases practice. What worked was the WholeTeam, RefactorMercilessly, PairProgramming, and TestDrivenDevelopment.
My understanding is that the initial 9000 monthly salaries were and continue to be paid using the system. That was a successful completion of the first rollout. The next rollout to the remainder of something like 27000 employees was cancelled. Any further comments, info? --JasonNocks
This is contradicted by ChetHendrickson on CthreeProjectTerminated:
"At last report, the software is still running. It is not paying as many people as it did at its zenith. In fact it is paying a very small group of former employees who are eligible for extended layoff benefits. Not what we had in mind when we started, but it is doing a job that no other available piece of software can do.
still wants to replace all of their payrolls systems with a single solution and are gearing up to spend about 4 times C3's total cost to do it. This will be DaimlerChrysler
's fifth attempt to do this. C3 is the only one to go live, in fact, at one point it was paying employees from what were three separate payroll systems."
KentBeck stated at the very start of ChryslerComprehensiveCompensation that he was "involved in a GemstoneProject at ChryslerCorporation to replace their many payroll applications with a single application." In the end, it didn't happen. If they tried three times before and are about to try again, that kind of suggests that it was a CriticalSuccessFactor for the project...
Why are you yelling? This discussion was started with a question. There is much conflicting information on the WikiPages. Part of the problem is that this is all second-hand. Do you have first-hand knowledge of the project or subsequent ones?
I wasn't yelling, merely the third voice. I agree that there is a problem with sources, which is why I was referring to authoritative sources who did have first hand knowledge. If these sources wish to repudiate their earlier remarks, I am sure they will feel free to do so.
To me, bolding a response that consists of several sentences comes across as yelling.
Then that's not success, is it?
There's a quote about halfway down the ChryslerComprehensiveCompensation page that indicates "The launch was considered a success by everyone." --JasonNocks
Most marriages last more than one month. Success means working through the
hard issues of real deployment and maintenance. The project from a consultants
point of view is one where you got to learn a lot but it is canceled so
it never has to really work.
Except if it is still in use by 9000 people, then it works for them, doesn't it? --JasonNocks
No. That wasn't the goal of the project. The goal was to handle it all.
If you don't get to meet that goal, possibly through no fault of your
own, the project is not a success. Scale and change are critical parts
of any system. Those did not get to be evaluated.
I don't know what the goal of the project was. I wasn't on the project, were you? From what I've read on the ChryslerComprehensiveCompensation page, it sounds like the goal was initially to get the software released and in use by the set of 9000 users. It sounds like a *new* goal was introduced at that point, and that goal was questioned, leading to the cancelling of the second rollout. Does that change the success of meeting the first goal? Again, I wasn't involved in the project. Anyone want to comment who was?" --JasonNocks
If the rollout was to be for 27000 users then that was clearly the goal.
It doesn't sound like a new goal. Projects are usually released in phases
where you have to pass the first phase to move on to the next. Sorry if
you want this project to be a success. But they canceled it and it never
got to run in its full configuration. That's not success from an objective
I don't want to treat the project one way or another. That's why I asked the question? :) I'm just trying to get an accurate account and put together an accurate assessment. Your questions have helped me formulate an opinion. Unfortunately my opinion seems to differ from yours. Again, please read the ChryslerComprehensiveCompensation page where it indicates "The launch was considered a success by everyone." That page seems to indicate that there is a definite question as to wheter the project was supposed to be launched beyond the initial 9000 employees.
"After termination, the Customer (the second one the project had had) was quoted as saying that he never thought the idea was really to launch the subsequent populations."
Please use your bullshit filter. People don't spend gazillions of dollars
to make 1/3 of a system. But they do try to cover things up afterward.
See also CthreeProjectTerminated
. By all standards, something that is cancelled by the customer after four years, while benefitting by the contribution of top Xpers like RonJeffries
, should be considered a serious failure. The funny thing is that XPers are quick to blame the customer on this one. If there was a success then the project would not be terminated
Well, if you are claiming that anyone is covering up the termination of the ChryslerComprehensiveCompensation, why does the CthreeProjectTerminated WikiPage exist on this wiki at all? And, please don't forget that the ChryslerComprehensiveCompensation was pretty much the very first project using ExtremeProgramming.
If the project was really a successful launch, combined with a follow-on project that was terminated, then the initial project was a success. That seems to be what I see when I check out the documentation on this wiki. Do you have some other information available?
I was not on the project, were you? I don't know the specifics involved at the time, do you?
No, but i don't toss away my common sense from being on 20+ projects, have you?
Why does it matter if it was the first XP project? People fund projects to solve
a problem. It's not charity. If the next phase was 27000 and they didn't do it
then the project didn't complete and it didn't succeed. I have been on several
projects like that. I would not count them a success. If you do then you
would count planning the wedding ceremony a successfull marriage.
However, the law considers a wedding a marriage after consummation. One might anticipate more from a marriage but just one is enough. If the deliverable functioned as specified and was nonetheless terminated that might be the result of a marketing or political failure but not an engineering one.
Software projects and marriages are totally different things. The analogy is completely inapplicable and off-base. Software projects should have endpoints defined in advance. These endpoints should be something other than death ('til death do us part). If you regularly get into arrangements where the customer can end the relationship, good for you.
I have been in a situation where the original requirements were met and working tested software delivered, the relationship continues, new requirements are added and delivered, and this continues until the customer decides that there are no more requirements to add that are worth the time and dollars. I would consider that a success and don't feel that I am stretching the stated definition of success in any way. I've been brought back in later by that same customer to do more work. I would consider this a very successful relationship.
In a marriage, these periods of inactivity in a relationship would be considered periods of separation. The marriage would be considereed "on the rocks". Again, your analogy is completely off-base.
As for "Why does it matter if it was the first XP project?". We're humans, we learn through experience. We have the potential to do a better job the next time. You brought up that the project benefitted "by the contribution of top Xpers like RonJeffries and KentBeck". My point is that if this was the first XP project, what experience can these "Xpers" possibly be offering above and beyond anyone else? They have no past experience with the XP practices, because previously there was no such thing as XP practices!
Many of the practices of XP were common. Some were not. Some were novel. From a customer perspective it doesn't
matter what methodology someone uses. Nobody cares. I'm sure they just wanted their payroll
system. This could not have been a requirement change. They knew
they had 27000 people to do payroll for. Not doing the entire payroll doesn't make any sense. So i can't buy
I believe that you are mistaken whey you say "Many of the practices of XP were common"? I believe that there was no official XP process at the start of the ChryslerComprehensiveCompensation project. And, XP practices are still somewhat uncommon. If I'm mistaken, please feel free to add projects successfully practicing XP to this page.
Customers do care what methodology developers use. Particularly, when past software development projects have historically failed something like 75% of the time. Particularly, when they can see a productivity boost of 42% (that means they potentially spend 42% less). Particularly when they see 1/5 the number of defects. However, customers don't want to be the first to receive the result of a new methodology. That is the whole point of this page.
Not doing the entire payroll of a company as diverse as DaimlerChrysler may make a great deal of dollars and cents. It might be significantly cheaper to have a couple of different payroll systems covering different employee bases (by country, type of employee, etc.). You don't need to "buy" anything on this page, the question at hand is whether there was successful application of XP at DaimlerChrysler. There seem to be two answers to this question, yes during rollout to 9000 people. And, no, to rollout to 27000 people. Would you be happier with moving the ChryslerComprehensiveCompensation project to a page listing PartiallySuccessfulXpProjects??
Any team that adds more testing and increases communication should
do better. There's no reason not to have successful XP projects.
Yes, so there should be plenty of XP projects that can be briefly documented. This should be an easy list to put together. This doesn't change the fact that some more conservative BusinessPeople? want to see a list like this. --JasonNocks
From my understanding, it was an business success, because the adoption of XP made clear the true cost of the project. That the marginal cost outweighed the marginal benefit of continuing is beside the point; from a business stand point having that information to make that decision is inherently valuable.
A good way to determine whether the stakeholders considered the project a success: What performance reviews were given to the design team? Often, being a key participant in a project held by management to be a "failure" results in poor performance reviews or other negative consequences for the key personnel on the project--up to and including sacking.
Note that a cancelled project does not equal a failure. I've been on one project that shipped four months late and well over-budget; management considers that one a "failure" despite the fact that the product has been a good-performer in the marketplace. I've been on other projects that were stopped in their infancy when it was clear there was no money to be made; that was nonetheless considered a success. Guess which one negatively affected my next review, and likely cost some others on the team their jobs? The one that shipped.
And, XP practices are still somewhat uncommon. If I'm misstaken, please feel free to add projects successfully practicing XP to this page.
To me the novel XP practice was pair programming. I had never heard about it or seen it
done. Writing tests first was new as well. I don't agree with it, and find when listening
to the code i know the order, but it was original to my ears.
All other XP practices were out in the wild in one form or another.
I had never heard of PlanningGame, ReFactoring, UserStory, etc. There may have been practices similar, or origins in other practices, but that's not the same thing. And, certainly as a collective, and mutually supporting whole, the C3 project was the first XP project as far as I understand.
Certainly most people do want their religions to have sprung whole with no nasty
origins in other practices, but that's not usually the case. Your need for the practices
to be new, in an industry with many years of practice and many skilled developers,
doing many different things is telling. I don't know if C3 was the first XP project,
but that wasn't the point, nor is the project being XP relevant to the success
discussion at all.
Did you not read the sentence "may have been practices similar
, or origins in other practices." ? Who's claiming new with absolutely no "nasty origins"? I said "as a collective, and mutually supporting whole".
You took it back by saying 'but that's not the same thing'
Who's talking about religion? Who's trying to convert anyone else to try anything new in this discussion in any way? Since when did asking if something was successful indicate that I am trying to convert that person to doing the same thing? Can't there be an attempt at intelligent discussion?
Stating that these practices are new is telling of what? That I enjoy XP practices and find them refreshingly different than previous experience? Or, that maybe I think that the industry needs some new practices when previously something like 75% of all software projects were considerd failure? That maybe I think that some of those projects might be more successful? The majority couldn't be a whole lot worse! So, yes, I'm guilty of wanting to try something new. So what?
The C3 project being the first XP project is relevant because it was claimed that people with significant XP experience were involved in the project, when there was previously no experience to be had. This is like when employers require FiveYearsExperienceInNewTechnology
. And, this discussion is about XP projects being successful, currently one in particular. Or, at least it was before it started to degenerate.
The C3 project being the first XP project is relevant because it was claimed that people with significant XP experience were involved in the project.
Good experienced people were in the project. That is a huge predictor of success.
A project is only a success in the eyes of who paid for it. The process used
or the feelings of the developers about its success aren't important. In a project
retrospective what would you bring up as the criteria for success? It lasted
longer than we wanted and cost more, but they used XP so that's ok? I don't think
so. When i buy a nail, as long is works as expected, how it got that why
isn't important to me. Software is the same.
See also: WhatIsSuccess