Closed. This question is opinion-based. It is not currently accepting answers.
Want to improve this question? Update the question so it can be answered with facts and citations by editing this post.
Closed 3 years ago.
Improve this question
We're generally familiar with code smells here, but just as damaging if not more so are when the business side of things - as much as it falls within our domain - is going wrong.
As examples, the inverse of anything on the Joel test would be considered a major process smell (i.e. no source control, no testers) but those are obvious ones and the point of "smells" is that they're subtle and build into something destructive. I'm looking for granularity here.
To start off with here's a couple (which can be turned into a list as the answers come in)
Writing code before you have a signed contract with the client
Being asked for on-the-fly estimates ("just a rough one will do") for anything which will take more than a day (a few hours?)
Ancient cargo-cult wisdom prevails (personal example - VisStudio sourcesafe integration is banned)
You've stopped having non-project specific group meetings (or lack any similar forum for discussion)
So what are some other process smells, and just how bad are they?
The book "Antipatterns" by William J. Brown et. al. has a bunch of project-related smells. They aren't always disasters in progress; mitigating circumstances exist for just about any smell.
The Portland Pattern Repository also has a page on Antipatterns, covering many of the same topics as the "Antipatterns" book. Visit http://c2.com/cgi/wiki?AntiPatternsCatalog and scroll down to "Management Antipatterns." A few examples:
Analysis Paralysis - a team of otherwise intelligent and well-meaning analysts enter into a phase of analysis that only ends when the project is cancelled.
Give Me Estimates Now - a client (or PointyHairedBoss) demands estimates before you have enough data to deliver it. You accept the "challenge" and give out of the head estimates (i.e. guesses). The client/boss then treats the estimate as an iron-clad commitment.
Ground Hog Day Project - meetings are held which seem to discuss the same things over and over and over again. At the end of said meetings, decisions are made that "something must be done."
Design By Committee - Given a political environment in which no one person has enough clout to present a design for a system and get it approved, how do you get a design done? Put together a big committee to solve the problem. Let them battle it out amongst themselves and finally take whatever comes out the end.
Collect them all! :-)
Back Dating - being given an end date and then told what needs to get done
Inverse QA Coverage - QA focuses on the non-essential items (because that's all they know how to test)
Environmental Alignment Issues - the various environments (Dev, Test, Staging, Production) are not in sync for code and data - therefore any testing prior to production is invalid
Delivery Date Detachment - no one believes in the end date because: it was made up to begin with and 100% of prior projects never met their delivery dates
Old Grumpy Code - old code is feared because there's no desire to refactor
the evil pm triangle (scope, cost, resources and/or quality) - to adjust the project you need to add people, reduce quality, reduce scope, etc....once a project is in motion, most changes (even reduction in scope) will increase time and cost and reduce quality..once the train tracks are down, it's tough just hanging a left turn
One smell I have a real problem with (because I work with it): Not ditching tools, dev software, methodologies, or anything else that doesn't work.
Many times, there is one (or more than one) piece of software that clearly, blatantly, doesn't work and likely interferes with the development process, but which a project manager simply refuses to replace/upgrade "because it would cost too much {time, money, whatever} to replace."
Edit: This also extends to machines and other infrastructure too (examples: a build server that takes an hour to do a two-minute build, or a version control system - ahem CVS - that takes 15 minutes to find out whether there have been any updates on a 50MB source tree).
Shipping a prototype - "we'll productize it later"
I suggesting checking out the Organizational section of the wikipedia page on Anti-Patterns. The I've had to deal with are 'Crisis mode' and the 'on-the-fly estimates' you mentioned.
You haven't had a post project review....when the project ended 6 months ago.
Some smells I have seen:
Optimistic management, but they can't pay your salary this month. This is real bad. I left the company in time but it died a few months later.
Extreme fanatic team building sessions. Focussing on how great the company is. But in the end it all goes down.
Good new people are laid of because they tried to change the process. Real shame. I have seen some people that really tried to improve the company, but old habits never dies so it often ends in a big desillusion.
The boss is always right mentality...
There is more but I won't spoil the fun for others.
Changes to process are made with no thought to timing or current deliverables, then immediately reversed when deliverables turn up late due to instituting new process.
Someone goes on medical leave and the team as a result is behind trying to pick up that person's work as well as their own and when the managers or clients or client sales reps are told things will be delayed as a result, they are only concerned about when things will happen and can you work nights and weekends in the meantime and never even ask about the person with the emergency and how he or she is doing.
When overtime for low level people is expected but the people who want this urgently leave on time and are not availble to answer questions. Or when they make you work overtime to be ready by 8 am and then don't look at it on QA for three more days. Hello I could have done it by then in regular hours.
Delivery of needed files (for data import for example) or information minutes before the due date and then blaming developers when due date is not met.
What I call: NIH (Process edition), a.k.a. Choose your own adventure.
Evidence of this:
you spend endless meetings debugging the process. And refactoring it.
nothing really gets done, because no one knows what they should be doing.
I guess this is an antipattern, rather than a smell.
Interesting question and even more interesting answers. Thanks for those.
I have been in almost all roles of software development (Developer, QA, Tech Lead, Project Manager - even client) and I can safely list the following smells
How quickly does the team react to new inputs (and how accepting are they of change)
How many layers does it pass through to get things done (beaurucracy)
How clear are the features/tasking - and are the goals SMART and do we have any KPIs.
How serious is the team working on the project about it
Is the team meeting regularly (read daily) to discuss achievements, goals and issues.
Most important, however, and the most evident (to a good nose) is the hygiene level of the project management tool being used (excel sheet, piece of paper, agile tools, email, whatever in whichever methodology you use). That is the first thing I notice while evaluating projects.
Do I know where the project stands at the moment?
Can I tell (Without asking the team) what needs to be done next?
Can I tell what the team is working on right now?
Can I tell when the next release is and if its achieveable?
Can I tell if client is getting regular updates?
Can I tell if client is giving approvals and if his feedback is taken care of in time?
Can I tell just from looking the load distribution of the project on the engineers?
Obviously, all this is well covered if you pick any modern Agile methodology, but depending on the market and kind of work, the mileage may vary. So keeping myself methodology agnostic, this is the bare minimum set of smells that should be rid of.
Related
Closed. This question is opinion-based. It is not currently accepting answers.
Want to improve this question? Update the question so it can be answered with facts and citations by editing this post.
Closed 3 years ago.
Improve this question
I need a methodology to organize my team for a university assignment.
I am a university student and already have some programming experience. My experience of working in teams of more then 2 people on a relatively large project is that everything usually gets done very quickly and badly in last week or so because of planning, organization and communication issues. In January I will have to do a (quite challenging for my skill) programming project (Java application on Oracle) in a team of 6. I already know my team members and was elected the project leader. It won't be realistic to expect people to get together for any meaningful-sized amount of time - everyone is free at different time and probably nothing more then weekly 1-hour meetings is realistic. People are hard-working and devoted to success, but everyone has their own circumstances. Mostly distributed work is the likely way to go.
I've looked through XP, Scrum, but they all require to sit together (unlikely), aimed at full-time development of a project (people have other assignments and part-time jobs) and customer involvement (we will have written specs and as far as experience goes, emails from tutor will be answered at best in 2-3 days).
Any suggestions how to organize people and divide the work? I am serious on researching this topic, because there will be much more of this kind of works later on.
Any help appreciated.
While not a programming project, I am currently in a position similar to you - elected group leader in a 5-person group assignment. We will be finished two weeks early, so I would say that the group has been a success. A few things that I did, as group leader, to ensure that this happened (I am also drawing on experience of managing projects in a 'real world' situation, so perhaps have a little more advantage):
Set ground rules from the offset. Make sure that everyone knows what is expected of them: Meeting attendance, work deadlines, what to do if they are struggling. We have a system at our university where people can be removed from groups if they are not 'pulling their weight.' If your uni has a similar thing, the steps taken for you to initiate that procedure should also be outlined. For example, miss two meetings and you will be 'yellow carded'. Miss another, then it's a 'red card'.
Break the assignment into manageable chunks and then decide on a time line. If you already know what should be done each week, it will be much easier to assign tasks each week. Or, of course, whatever time frame you decide on.
Meetings. There should be an initial meeting where everything above is decided on, probably a rather long meeting (our first meeting took 90 minutes). In that meeting, set out tasks to be completed before the next meeting. Then, in each subsequent meeting, validate the work that has been done by each person, ensuring it is both complete and correct. And then, of course, delegate tasks to be completed for the next meeting. And so on...
Each person, pair, or whatever, should go away and do the work themselves, independently. Because meeting times are so short (as were ours), they should be about making sure everything is completed, plans are made and tasks are delegated.
Communication. I set up a forum for the members of my group to communicate about the work they are doing and to upload completed material. I also have a specific sub-forum where people can post up any problems they might have - with a rule that they should do so within plenty of time of the next meeting - so that others can help, or tasks can be re-assigned. It's important that each meeting has minutes taken and an agenda set for the next meeting. I get these uploaded to the forum so that anyone who did miss a meeting isn't left in the dark and know exactly what they need to do.
Because your project is a programming project, the following may well help the group stay organised and cohesive:
Early on - preferably in the first meeting - break your program into 'modules' or classes/procedures/whatever; basically, manageable, stand-alone chunks of code to be written. These can then be assigned to someone each week. To make sure that time won't be wasted on changing code needlessly later on, decide on your global variables, if any. It might be a good idea to also decide on the methods and properties of each class before any coding starts.
As sgolodetz suggested, you might also want to use version control to keep track of code as it's being written. Again, if you do this, make sure that rules/guidelines are in place for when/how often to update.
Perhaps arrange that people working on related tasks meet up between meetings to make sure their code integrates well. This way, they can work together to make sure each person's code 'cooperates' with the other's.
I think the key here is to be organised. Have a strict, detailed plan and stick to it as much as possible.
In university, at least, I think it's important to have a certain amount of ruthlessness - remember that if people don't pull their weight it will affect your grade as well. It might just be me, but once the rules are set, they should be abided by and breaking them (ie: Not completing work or not letting people know you're struggling, not turning up for meetings, etc) should always result in the 'card' system - as would have been agreed at the start. It's your future on the line, so don't let people who don't care jeopardise that.
I was one of the leaders of a team project (6 people) last year at university. Based on my experience, I have a couple of key points.
While forums, online chat, email, SVN commit messages etc. are all very good communication mechanisms (and I do recommend them), nothing is more important than weekly meetings. These should involve delegating/discussing/monitoring tasks, talking about general issues and the bigger picture, making group decisions, and simply using your app and identifying issues with it as a group. By doing this, it becomes instantly obvious where people are at with their tasks, where they might be confused, and whether they understand what needs to be done. This is crucial because the feeling of being 'lost' and having no idea where to start is often the biggest obstacle to making progress, rather than laziness. And besides, group decisions are best made in person where you can bounce ideas off each other... and this all helps to establish a sense of shared ownership.
There is no one best way to formulate, delegate and monitor tasks. But it is important that you can divide and parallelise (as much as possible) the workload into sensibly sized tasks and identify dependencies between them. Our group maintained a TODO file in our repository which consisted of a list of tasks, dates, and bugs. We prioritised tasks according to factors such as whether they were mandatory requirements or optional niceties (think MoSCoW), whether they were on the critical path, how much risk/uncertainty was attached to them etc. We allocated one or two people to most tasks based on their strengths and whether they actually wanted to do the task, but we also left some tasks open to whoever wanted them (this is potentially dangerous too). A couple of us stronger programmers also played 'floating roles' meaning we could help out with any task. A key point is that the task descriptions should contain anything discussed in the meeting... the what, the why and a couple of 'how' sentences as a kickstart (i.e. "to accomplish this, you might want to read up on blablabla")... anything which makes it easy for your teammates to slip into a task.
A few thoughts:
Communication - perhaps use IRC or a wiki
Assigning people to tasks - you need to figure out where people's strengths and weaknesses lie, break down the project into chunks and assign them accordingly
Monitoring - a big problem with projects where people are working remotely is that they may not do the work on time; you need to think about how to keep on top of this (e.g. use version control, and look at the check-ins) - to make this work, everyone needs to buy into it up-front
Strategic reserve - it might be worth having one person (perhaps you if you're the strongest programmer) who takes on tasks across the project that other people are finding difficult to prevent people getting bogged down
First, set up a version control system for the project. This serves the purpose of the code that has been checked in is available to all the team, you can revert to earlier versions if someone breaks something, and all decent professional shops use source control, so it is a job skill it is good to be in the habit of using.
Now you have a way to check on progress. If Joe hasn't cheked anything in, you know he is not doing the work. IF Sally quits in mid_term, you have a record of everything she already did.
Another thing you want to so is code review. Everybody's code gets reviewed by one or more other people. You will all learn a lot from this and this too is a necessary skill for the professional developer.
Plan what has to be done in what order. Make sure the people developing the stuff that everything else depends on are the most reliable devs on your team and that they will meet their early deadlines or nobody else will be able to do their parts.
Weekly progress meetings and don't be afraid to call someone out for not doing his or her part. The sooner you identify a person not pulling his weight, the sooner you can address the issue.
Closed. This question is opinion-based. It is not currently accepting answers.
Want to improve this question? Update the question so it can be answered with facts and citations by editing this post.
Closed 5 years ago.
Improve this question
I posted this question on Reddit Programming and did not get a single response. So I am hoping that Stack Overflow community will have an opinion.
Have any of you ever been on a software project that had fallen behind, where 'Crashing' or 'Fast-Tracking' the project schedule actually brought the project schedule back on track? I have never seen either of these project management techniques actually work. And all the articles on software development that I have read all state that these 2 techniques do not work and actually pushing the project further behind (for example literature on the Mythical Man Month). So who has seen it work?
Thanks Bill.
I have only ever seen it work once. It was a three or four month long project that was projected to run an extra two months over the original delivery date. The project got fast-tracked and things ended up getting back on track for the release.
...keep in mind though, that was only once. I've been on many more projects where the PM tried to use one of those two methods and they failed miserably and dragged the project out for months beyond already extended date.
It can work. But there's a price to be paid: lower quality (more bugs, less testing) and turnover of burned-out programmers.
And in many cases, a fast-tracked project will both fail to deliver on time and will still pay the full negative price, for the reasons stated in Mythical man-month.
I've seen it work but it's not the norm.
Things I'd want to see before I thought it might be feasible:
1) Staff available with suitable skills and approach. By that I don't mean ".NET programmer", I mean detailed technical skills, business domain skills (so they understand the problem), personality fit and understand the tools and the approach (source control, methodology and so on). This can happen in large companies where there are common tools, standards and knowledge but you need to be sure that they're ticking pretty much all the boxes.
2) Tasks must be nicely divisible. The best situation is where there are whole modules, applications or tasks unstarted and you can put new people on that. It minimises upskilling, additional communication and so on. If you can't separate out what the new people will do you're likely to majorly disrupt the existing team.
3) The whole team must have bought into the approach. If the existing team don't agree that bringing people on board will be right they'll likely fight it and you're doomed.
4) You need to be sure you've addressed why it was running late in the first place. If it was just bad estimates then are you confident the new estimates are good? If it was scope creep have you got the scope and change control in hand now? If it was because the deadline moved, are you sure it won't move again?
If you can't tick all four of those off, it isn't going to work.
Crashing and Fast-Tracking are two very different things...
Fast Tracking is where you take something (tasks or work packages) out of sequence and do it early. This may because of hardware delivery lead times, availability of resources, risk or whatever. So you might do things in parallel where originally you had planned to do it sequentially. I've fast tracked a lot of projects.. and yes it works.
Crashing a project is different in that you typically throw more resources at a problem to get it done quicker... this can be tricky. If it's done as a crisis response it can be painful adding extra people as you are already under the pump. In some situations you just add more problems.
Another alternative to crashing is to reduce scope. This is not always possible, but it should be considered.
With fast tracking or crashing... the sooner you know when you need to make a schedule change the easier to manage. This is why early deadlines are so important, they indicate how the rest of the project will go.
Both of these project management techniques work well to maintain a schedule, but they should be used intelligently by judiciously analyzing the network diagram:
study the variance,
study lead and lags;
decide what suits to your project: ‘Crashing’ or ‘Fast-Tracking’.
There is a software management principle that says adding manpower to a late project makes it later.
That said, as long as the measures taken are sensible it should be ok. Don't expect too much of your staff and provide reasonable incentives and don't take short cuts. It won't make miracles happen but if you're practical and want to push things just that little bit faster it can definitely be done.
When people have a stake in the potential success of something it's amazing how much more effort they're willing to put in.
It depends on what you mean by "work". I don't think I've ever seen it make a way late project deliver on time, if that's what you are asking.
However, I have seen it make way late projects deliver only a bit late. From the fuzzy perspective of management, that might be called "working". I've also seen it significantly lower the customer-based pressure on the company. Some might also call that "working".
Of course the price is rather high. Employees burn out, develop health problems or big problems in their neglected personal lives, etc. All of that has large financial repurcussions to the company. So I doubt the company comes out ahead in the long run. Is that "working"?
Closed. This question is off-topic. It is not currently accepting answers.
Want to improve this question? Update the question so it's on-topic for Stack Overflow.
Closed 10 years ago.
Improve this question
I have been in the IT industry for 10 years now but have worked in "traditionally" managed project teams (both well managed and badly managed ones).
I have heard of the "new" scrum or XP type of project management and yearned to be part of one (as s/w folks we always like anything new I guess) but have not got an opportunity.
My question is this - what are your experiences in moving to the "new" way - was it significantly better or worse or not any different? Has there been any project success rate improvement when using XP way of development or it is same as any well managed traditional projects?
This should not be a political question but just your experiences as you have moved to the new world or experienced at least once and back.
Thanks in advance
Before I ever heard of XP, I had a really good manager (Mike) at an early job I had. He was used to managing engineers and transitioned to managing software. After a few bad working experiences I looked back at his style versus typical project management I had before and after working with him.
Met with everyone at least once a day but gave us space to work
Used a whiteboard with two columns, people working and what they are working on anyone could look at that board and see if something had been done or was being done
Had everyone cross-train. I learned rcs and then cvs there and how to use make files
Ran productive "post mortum" when a task was completed. He would ask question like "would it have helped if X?" or "next time, can we try to..."
Kept everyone working on short tasks and managed our time so we always working on something but never had a ton of stuff piled up
Mike did everything on paper. He would keep notebooks and index cards with him. He insisted that anything asked of him by management be converted into manageable tasks, often written on note cards. He refused to have anyone work on anything that couldn't be clearly explained or had a clear objective. He would ask the VPs "what do you mean by faster?" "What kinds of metrics are the reports meant to show?" "Why should this be a priority?" He seemed to have near infinite patience in writing out what needed to be done and what was meant by "done"
When I first read the XP book, I was amazed by how much was familiar as "the way Mike worked"
It seems that Agile is just about implementing a set of best practices and evaluating how they work in your environment. When they don't work, change them. When they do work, stick to them.
I think the real problem with traditional project management is that more often than not, it doesn't really exist. I'm amazed by how many shops claim to use RUP or Code Complete or even Agile and don't actually have anything recognizable as project management. Sure, there are meetings. And people called project managers. But ask a simple question like "what has been done on project X" or "what is left to do on project Y" and no one has an answer. They have to dig though emails or point to a comically inaccurate MS project file.
If a person claimed to be on a diet and couldn't answer questions on what they were eating or how they were exercising; would you accept that they were really on a diet?
You take your old baggage with you when you go. Meaning that any project management bad practices you had before will still linger.
However, I will say that things improved greatly when we began to close the loop between us and the customer. Greater and more frequent feedback and prototyping with the customer means far fewer moments of the customer saying, "This is not what I wanted."
I've used (a slightly modified) Scrum before at work and here are my thoughts:
The daily meetings and burn-down provided motivation to make progress on tasks.
Our manager could talk to colleagues across the pond and show them "this is what we're working on this month."
You knew exactly what tasks you needed to get done, and had already estimated the time required to complete.
When priorities changed (new tasks, important bugs added), there was a well-defined process to handle adding them to the sprint or simply pushing them to the backlog.
These are lovely answers, but I think everyone's confusing project management with development/design methodologies.
I'm on a team that started Scrum a few months ago and we seem to be getting things done faster and with much less "waste" (projects that are scrapped). Just my observations from our small team (4 devs).
I've found the overall move to Agile/XP practices very positive, in many ways it front loads quality into the project/development process. You'll need buy-in from management and from the team to really see success...a few suggestions:
trial any change with a small project (2-3 people)
understand what areas your current team can most improve (quality? productivity? time-to-market?) and incorporate a few Agile/XP/Scrum (what ever) processes in...don't incorporate them all in at the same time and understand which processes address which issues prior to any change
if possible - track those areas you're looking to change and compare to another project running at the same time (the mere focus of improving something often is enough to improve it ,there's a study/term for this, but I forget what it is)
sometimes you'll see a dip in performance as you begin a new process, this is part of the learning curve
never assume that a good change today will remain a good change tomorrow, always review your project areas and be ready to change any process at any time
no change remains good forever, just like refactoring code, refactor your processes
ensure you get buy in from the team and management, you can't force success
I like some of the things the agile approaches do, but I also value some of the things traditional approaches do.
Both can work, as can a mixture of the two, which is what I find works best for my team now. I have implemented incremental development and it really helps us; iterative development is a little harder and we're still working on that. However, we have a variety of constituents, and many of our stakeholders (and PMs) prefer traditional artifacts and milestones. So we have to keep finding the right balance.
I have also found that even more important than the methodology is the people implementing it. Good people find a way to do good work and get things done regardless of the methodology, although certainly the methodology can have effects on efficiency (and morale :) ). Poorly aligned resources, however, can use the finest methodology and find ways to deliver poor results.
For developers, the great lessons of XP & Co. are shorter release cycles, and a more evolutionary approach - in the sense that change of requirements is accepted as a natural part of any project. Also, Customers suggest solutions, but designers and developers need to understand the problems.
Lessons for managers: Developers are not exchangable spec-to-code-converters, their individual strengths and weaknesses can make a productivity difference of 10 or more for a given topic. Knowledge and experience are the most valuable skills in your team, and developers can teach each oterh. Managers need not understand what developers do in order to enforce desired results.
XP & Co. are usually mixing solutions to these with the problem to make a company change. The heroic XP consultant singlehandledly saving a doomed, delayed and derailed project acts as large part as a buffer between development and management. But if you are looking at what to learn, you have to separate these aspects.
What I learnt in the recent years is that bugs aren't a personality fault, and that the sky doesn't fall when specs change. I've learnt that while design errors are still the most expensive to make, there isn't a single "perfect" design. Instead of getting one thing right we need to implement safeguards that of all the many details none goes wrong - and I've learnt to use the leeway between "right" and "not wrong" to our advantage.
My experience has been that I prefer to use Scrum over traditional approaches as it hasn't happened often that requirements could stay unchanged for the length of a project where usually projects seem to run at least 6 months to my current one that is over a year.
There can also be the case where there isn't any project management and everyone just scrambles to "make it work" so having some formal structure is good over nothing. There is something to the question of how well does the team come together and egos rarely appear as it isn't someone's code but rather the code of the team and there is a kind of group think where while each person has their view, no one tries to make everyone else see things that way.
At times it seems to me that some Scrum and Agile approaches I've used end up being like rapids instead of a big waterfall. What I mean is that the cycle of gather requirements - Analyse and Design - Implement - Test - Deploy and get updated requirements seems to be repeated over and over so that what comes out in the end would be extremely hard to state at the beginning of the project unless the project sponsor could give very detailed requirements that would never change.
As it currently stands, this question is not a good fit for our Q&A format. We expect answers to be supported by facts, references, or expertise, but this question will likely solicit debate, arguments, polling, or extended discussion. If you feel that this question can be improved and possibly reopened, visit the help center for guidance.
Closed 11 years ago.
As long as there are software projects, the world is wondering why they fail so often.
I would like to know if there is a list or something equivalent which shows how many software projects fail today. Would be nice if there would be a comparison over the last 20 - 30 years.
You can also add your top reason why a software project fails. Mine is "Requirements are poor or not even existing." which includes also "No (real) customer / user involved".
EDIT: It is nearly impossible to clearly define the term "fail". Let's say that fail means: The project was more than 10% over budget and time.
In my opinion the 10% + / - is a good range for an offer / tender.
EDIT: Until now (Feb 11) it seems that most posters agree that a fail of the project is basically a failure of the project management (whatever fail means). But IMHO it comes out, that most developers are not happy with this situation. Perhaps because not the manager get penalized when a project was not successful, but the lazy, incompetent developer teams?
When I read the posts I can also hear-out that there is a big "gap" between the developer side and the managment side. The expectations (perhaps also the requirements) seem to be so different, that a project cannot be successful in the end (over time / budget; users are not happy; not all first-prio features implemented; too many bugs because developers were forced to implement in too short timeframes ...)
I',m asking myself: How can we improve it? Or do we have the possibility to improve it? Everybody seems to be unsatisfied with the way it goes now. Can we close the gap between these two worlds? Should we (the developers) go on strike and fight for "high quality reqiurements" and "realistic / iteration based time shedules"?
EDIT: Ralph Westphal and Stefan Lieser have founded a new "community" called: Clean-Code-Developer. The aim of the group is to bring more professionalism into software engineering. Independently if a developer has a degree or tons of years of experience you can be part of this movement.
Clean Code Developers live principles
like SOLID every day. A professional
developer is the biggest reviewer of
his own work. And he has an internal
value system which helps him to improve and become better.
Check it out on: Clean Code Developer
EDIT: Our company is doing at the moment a thing called "Application Development and Maintenance Benchmarking". This is a service offered by IBM to get a feedback from someone external on your software engineering process quality etc. When we get the results, I will tell you more about it.
Bad management.
Projects are not successes or failures based on some underlying feature of the project, but on whether they fulfill the needs of the users. (They can fail altogether, in which case there was a gross misstatement of what was possible.) It is mostly in the process of evaluating the feasibility and cost-benefit ratio of the project, and establishing goals, that software projects tend to fail or succeed.
There's a disconnect between people who deal with facts and things (like programmers) and people who deal with other people (like sales types and managers). To a programmer, the facts are the facts, and have to be dealt with. To a sales person, the facts are what other people think, and are changeable.
There's also differences between tangible and intangible facts. Nobody thinks that workers could build a large bridge in a month if they were really motivated; they can see all the steel and concrete and other stuff that has to be moved and fixed into position. Software is much less tangible, and lacks the physical restrictions: while it is not even theoretically possible to build the bridge within a month, it is conceivable that a team could create a large project within a month, as "all" they have to do is get everything right the first time. It is physically possible to type thousands of lines of code a day; it's just that the chance that they're usable as is is so close to zero it doesn't matter. The actual productivity of a top developer is actually pretty unimpressive in word count, compared to (say) the productivity of a journalist.
Therefore, those who are used to flexible facts don't have the imposing physical limits to remind them that things can be pushed only so far, no appreciation for what programming actually requires, and no good feel for how much productivity is realistically possible. Moreover, they know how to get their way in negotiations, much more than the average developer, so in negotiations about what's possible they tend to assume more than they can, in the real world, get.
In the meantime, software development is inherently fuzzy, because producing the physical product is trivial. I can produce a copy of software quickly and cheaply, once it's been developed. Software development is design work, pure and simple. Anything corresponding to manufacturing is ruthlessly eliminated with such things as compilers and wizards and code generation. The developer, faced with the manager who wants the impossible, finds it hard to say the impossible is actually impossible, because there's no way to say it's actually impossible. Given facts that are unknown enough to feel flexible, the person with strong negotiating skills and determination will typically get the answer he or she wants.
Given this disconnect, one might ask whose responsibility it is to bridge it. The answer is, in my opinion, clear. The responsibility for understanding how different people think belongs to the people who specialize in dealing with other people. The responsibility for coordinating different types of people belongs to the people whose job it is to coordinate these things. Therefore, managers.
Managers who do understand software development and developers, and can deal well with other managers, will do well, and their projects will generally succeed. There are still far too many of the other type in the world.
Not a direct answer, but I found the Virtual Case File to be a fascinating case study on how a big government-backed well-funded project can still tank.
You can also add your top reason why a
software project fails.
Another IEEE Spectrum Online article "Why Software Fails" examines this very question. It summarizes the major points as follows:
Unrealistic or unarticulated project goals
Inaccurate estimates of needed resources
Badly defined system requirements
Poor reporting of the project's status
Unmanaged risks
Poor communication among customers, developers, and users
Use of immature technology
Inability to handle the project's complexity
Sloppy development practices
Poor project management
Stakeholder politics
Commercial pressures
Poor planning.
Hofstadter's Law
It always takes longer than you expect, even when you take Hofstadter's Law into account.
Mismanagement.
SW project get started by throwing developers against a perceived problem. Business requirements crystallize as the project progresses. New functionality gets added while deadlines stay put. More developers are thrown in. Original project members quit or get fired. By this point too much time, money and resources is invested in the project so it cannot be canceled. As the deadline passes the project is declared finished and successful despite the obvious lack of finished product.
Come to think of it - I've jet to see a SW project fail...
Honestly, I think its because most programmers are not very good at what they do(and I don't mean just cranking out code). People on stackoverflow are probably the exception. I don't know about the rest of you but as a consultant/contract programmer I have worked in or around many places, and the ratio of mediocre or poor programmers to good ones is about 10 to 1.
One of my strengths has always been estimating accurately and then delivering on time and on or under budget - I always aim for coming in 10% under cost and on time. Then I like to tell my client that because I got things done for less $$ than expected, which of the "extras" would you like to add in?
Even a perfectly functioning product that is late and/or over budget will be considered a failure by many business managers. Programmers often focus on just the technical aspects of what they do, with little regard for the cost or deadline. You really need to do all three well for it to be deemed successful project. There are many other programmers that could code circles around me without a doubt, but for the person paying for the project, that is rarely enough.
It is because no-one seems to read anymore.
Every single reason why projects fails has been analyzed time and time again.
You only have to read three books to know why 80% of projects fail:
The Deadline: A Novel About Project Management (Tom Demarco, published 1997)
It's a great introduction and it's pretty entertaining.
Peopleware : Productive Projects and Teams (Tom Demarco, published 1987)
The Mythical Man-Month: Essays on Software Engineering (Fred Brooks, published 1975)
We as a profession simply seem to forget everything every 3-5 years (see "centralised computing is inefficient; let the clients handle it" vs cloud computing).
(From a programmers point of view - I'm not a project managemer, but I've often been involved in the process).
A number of people have mentioned that bad programmers are endemic. But I think this is true in another sense as well - we're all bad programmers in that we find it difficult to anticipate complexity, an unavoidable issue that 50 years of magic bullet estimation and planning schemes have failed to solve.
Anticipating the side effects of large projects gets exponentially more difficult as projects grow. This is a dull truism, for sure, but for me it means that on any project I've worked on where I've been involved in the estimating process I've run into some case where there's an unanticipated consequence of a design decision that causes everything to come to a grinding halt, or at least a few days of bugfixing - just something that nobody foresaw, not any sort of malpractice or stupidity. It's just the nature of a complex enough system.
Aside from the built-in uncertainty, there's also a tendency to underestimate things whose outline is known, because the fact that they have less uncertainty makes them seem simpler to implement.
So the uncertain stuff gets magnified, the clear stuff gets minimized, and what really kills you is the thing that you didn't think would be uncertain.
The number one reason: a failure of project management.
A PM's raison d'etre is to make a project succeed, ergo a project failure is their failure. Certainly there are factors beyond their control, but it's still within the PM's job description to manage that risk, and the only get out clauses should be someone higher up the food chain taking decision control (which is a terrible thing to do to a PM) or acts of god.
In my experience failures mostly occur when PM work has been fast and loose or non-existant, including when decisions start to flow from sales people and when the client starts decreeing change control. A good PM is priceless.
Failure is a judgement -- more of an accusation, really.
"The project was more than 10% over budget and time."
Which budget? Which schedule?
6 months ago, I wrote a plan saying it would take 6 months.
3 months ago, the users asked for more stuff. I gave them a plan that said it would take 9 more months.
Last month I was told that the project was 6 months over budget and therefore a "failure".
But wait. It delivered what the users wanted. It was over the "original" estimate. It was under the revised estimate. The users want more. IT wants less.
I'll approach it from a different aspect than most the rest here.
I've noticed a project slowly fail over a period of time. Sure, it's gotten better in that time too--but it still isn't profitable. In this market profitability, and being in the black, means success.
Why is it failing? I think it's simple: you get what you pay for.
Software is like a bank account, not primordial ooze. If you don't put resources into it (time, money, focus, effort) then you won't get anything out of it except failure and cost. So you must invest things into your project, and sometimes the earliest work sets the stage for years to come. You can't throw mud at your computer and expect a new mouse in two years and $10 million dollars later, so likewise there must be effort expended.
One of the biggest problems today are "budget developers" in a third-world country. I don't begrudge them their part of the market, but for a well-funded Silicon Valley startup to seek them out and get a budget foundation (framework or even prototype) is to make a poor investment in the future. This very same budget framework is what is causing my friends so much of a hassle today. It works now; it worked when it was written, but it wasn't written well and few even take the time to maintain it. Were the company to stop and rewrite the software the way it should have been written in the first place they wouldn't have all this trouble. Can they afford the time? Nope. They have to make it profitable before they can even thing of it.
As the saying goes, "I can make it: cheap, fast or good. Now, pick any two of those." Everyone wants all three, myself included. But if you don't invest the time, planning, and work required to make your project a success from the start ... then don't expect anything you can be proud of later. It'll feel like a forged Mona Lisa where you, and every other engineer like you, can see a defect here and there that shouldn't have been there from the start.
So:
Don't undertake what you cannot afford in: time, money, effort, focus, etc.
Don't skip planning!
Don't be afraid to rewrite early when it counts the most. (Later it'll be worse than a trip to the dentist, believe me.)
Don't underestimate the power of bureaucracy to prevent you from doing it right.
And don't be cheap where you should spend the most of your time. It will cost you later, guaranteed. And if not you, then someone else will take the bullet for you.
One common mistake is that sales people and technical people do not communicate sufficiently. So the salespeople sell things that are technically not feasable within budget. (And then they run with their bonus :) )
Being over budget and time is not a good definition of failure (and actually being in budget and time doesn't always mean success). Take the following examples provided by Hugh Woodward, PMP, in Expert Project Management - Project Success: Looking Beyond Traditional Project Metrics:
Sydney Opera House: With its graceful sails dominating Sydney Harbor, the Sydney Opera House is arguably one of the most recognized buildings in the world. Yet, from a project management perspective, it was a spectacular failure. When construction started in 1959, it was estimated to cost $7 million, and take four years to build. It was finally completed in 1973 for over $100 million.
[...]
Project Orion: This massive effort to develop Kodak's new Advantix photographic system was reputedly very well managed from a project management perspective. PMI recognized it as the 1997 International Project of the Year, and Business Week selected the system as one of the best new products of 1996 (Adams, 1998). But Kodak's stock price has fallen 67% since the introduction of the Advantix system, in part because it failed to anticipate the accelerating switch to digital photography.
Corporate Intranet: Finch describes a project that involved the implementation of a corporate intranet to globalize and improve communications. From a traditional project perspective, it failed to meet its success criteria, but not significantly. It was one month late and believed to have been accomplished with a small budget overrun. But both the project manager and senior management viewed the project as successful. The hardware and software had been installed successfully with a minimum of disruption, thereby providing all staff members with access to the corporate intranet. Following implementation, however, employees made only limited use of the intranet facilities. The main objective of the project was therefore not achieved. In this case, both the project manager and senior management focused on an objective that was too narrow.
[...]
Manufacturing Plant Optimization: A paper manufacturing company with five plants across North America decided to increase its manufacturing capacity by embarking on a de-bottlenecking program. A project team was formed to install the necessary equipment, and charged with completing the work in 18 months at a cost of $26 million. But almost immediately, the project team was asked to defer major expenditures until an unrelated cash flow problem was resolved. Rather than stop work completely, the team adopted a strategy of prototyping the technologies on which the de-bottlenecking program was based, and actually developed some cheaper and more effective solutions. Even when the project was authorized to proceed, the team continued this same approach. The project eventually spanned five years, but the resulting capacity increase was three times the initial commitment. Not surprisingly, the company immediately appropriated another $40 million to continue the program.
[...]
So in these examples, which ones were truly successful? Examples like the 2002 Winter Olympics and the Batu Hijau Copper Concentrator would suggest that these are truly successful because they not only met the traditional project managers' definition of success, but also met the projects sponsors' perception of success.
As we start to look at the examples
like Project Orion, the Corporate
Intranet and the Laptop Upgrade, we
notice that the traditional metrics
start to fail. These projects are
considered successes in project
managers' definition of success, but
failed at meeting the sponsors'
success criteria. The project Orion
example is quite astounding as this
project was recognized by PMI (Project
Management International) in 1997 as
the International project of the year.
Yet it did not increase Kodak's
revenue, because they did not foresee
the adoption of digital cameras.
Most interesting are the examples of
the Manufacturing Plant Optimization
and the Sydney Opera House. They both
failed to meet the traditional project
managers' success metrics but were in
fact considered successes. This is
particularly shocking when you see
that the Sydney Opera House had a
"cost overrun of 1300%" and a
"schedule overrun of 250%".
Once we realize that projects can fail
to meet the traditional metrics of
success, but still be successful to
the stakeholders, this creates a
quandary for the project manager. How
does one really define success? Is it
possible that a "Challenged" project
could be canceled that would have met
the sponsors' needs? Is it also
possible to identify a project that
should be canceled that is currently
on time, on budget and meeting the
defined needs?
What do you think of that conclusion? How do you really define success?
My answer is rather unusual from the rest of this, but quite common around here: killer bugs. I had a project go an extra two weeks (50% extra time) because of one switch in source that I didn't know about until I dug through the source code (there was nothing documented in help or on the web).
People/companies do not proudly shout about their failures, so many cases wont be ever heard.
Poor use of practices and software development methods. In my experience, one of the big reasons a project failed its that the development team use a wrong method to face the software development process. Choosing a methodology without having a good understanding of how it works and what it takes, can bring a time consuming issues to a project, like poor planning.
Also a common problem is also the use of technologies without a previous evaluation of it to understand how it can be applied, and if it brings any value to the project.
There have been some good studies done on this. I recommend this link from the Construx website (Steve McConnells company).
The Construx link above is real good!
Many projects are managed on a rosy picture of reality. Managers tend to power talk developers into optimistic estimates. But say a 20 week projects gets "talked down" to 10 weeks. The requirements phase will now be 1 week instead of 2. After 1 week, the requirements aren't finished, but the project moves on. Now you're working on shaky ground-- and on a stretched schedule!
This can be funny. Once there was this old guy in a room opposite mine. His job title was system adminstrator. But the system he was supposed to adminsiter wasn't there. It had never been finished, although management thought it had been. The guy played games for about a year before he got bored and moved on.
The funniest part? Management put up a new job opening after he left!
IT Project Failures is a blog about project failures that may have a few answers here if one wants to read about it.
Myself, I think a large part of this lands on the question of being able to state exactly what is expected in x months at $y when really the answer is much more open-ended. For example, if a company is replacing an ERP or CRM system, there is a good chance that one isn't going to get all the requirements exactly right and thus there will be some changes, bug fixes and extras that come from taking on such a large project. For an analogy consider how many students entering high school or university could state their precise schedule for all 4 years without taking any classes and actually stick to that in the end. My guess would be very few do that as some people change majors or courses they wanted to take aren't offered or other things happen that change what the expected result is but how is this captured in a project plan that we started here and while we thought we were going there, we ended up way over in spot number three.
The last statistic that I heard was that 90% of projects are either over time or over budget. So if you consider that failing, quit a bit.
The reason why it fails mainly lies on process. We as software engineers don't do a good job of gather requirements, and controlling the customer. Because building software is a "mysterious" job to people outside of IT, it is difficult for them to understand why last minute changes are difficult. It's not like building a house and clearly showing them why it isn't possible to quickly add another door to the backside of the house made of brick.
Not only software projects go over budget, or take more than scheduled time to complete. Just open the newspaper and look at public projects like bridges.
But for a software project it is far more easy to cancel everything. If a bridge or building is half finished, there is no way back. Half the infrastructure is in place. It is very visible and it takes money to remove it.
For a software project you can press Shift-Delete and nobody notices.
Its just a very hard job to do an accurate cost analysis.
Using the FBI's Virtual Case File system it comes down to poor program management. The program had pre-9/11 requirements and post-9/11 expectations. Had government management done their job there would have been contract mods.
http://government.zdnet.com/?p=2518
"Because of an open-ended contract with few safeguards, SAIC reaped more than $100 million as the project became bigger and more complicated, even though its software never worked properly. The company continued to meet the bureau’s requests, accepting payments despite clear signs that the FBI’s approach to the project was badly flawed, according to people who were involved in the project or later reviewed it for the government."
Although $105,000,000 for 700,000 lines of code comes to $150 per line of code. Not too bad.
To truly gage whether a project is really successful, the original estimate / budget is probably a poor yardstick. Most engineers tend to underestimate how much time it will take because of political pressure, and they don't know how to estimate anyway. Many managers are poor planners because they want unrealistic deadlines to please their boss, they often don't understand what's involved, and plus it's something they can look at and understand and use as a stick in meetings, as opposed to actually helping solve problems. Practically, it helps businesses make rough projections of expense for budgeting purposes, at least it gives them something the go by.
A better measurement of project success would be - are the users happy? Does it help the business make money? How fast will the money gained help recover the cost of the system? These are harder to gauge than a simple plan, though.
In the end, I've find you're better off delivering on deadline, even if it means working some overtime. It sucks, but that is the reality of it.
As stated previously the vast majority of people involved in software development to not actually understand how
ask the right questions to learn about the problem
appreciate the users goal and judge expectations
understand technology available and the related Eco structure
most of team directly/indirectly involved are poorly skilled.
do not appreciate or know when they are wrong do they can take action.
Even with perfect requirements and related definitions too many developers don't know what they are doing.
Just look at the types of questions asked here. Would you go to a doctor that asked the same equivalent question. The scarey thing is that they ask and don't know how to learn or reason.
Different agendas
Management do not really understand what a developer does, how they produce code and the difficulties encountered. All they care about is the final product delivered on time. But for a developer they care about the technical aspects, well produced code in a solution they our proud of.
Deadlines
I've often heard developers say they wish they could produce better code, and that deadlines often push them into producing something that just works, rather than good code. When these deadlines are unreallistic the problems are exacerbated.
I think this thread managed to gather the biggest group of tallented unhappy software developers, engineers, project managers, etc. that it is possible to gather in one place.
I share point of view with most of the posts stuck here and I think they come out of a lot of suffering through seeing co-workers not doing a good job when job (programming) and success is the most important part of our lives.
http://www.clean-code-developer.de/
They have an incredible cause! and their philosophy, if taken ahead, could manage to create a new layer of heroes, as the main stream of developers and IT professionals is so ROT these days.
I'm working on something similar here in Brazil, 'cos I love our profession of bringing software alive both as PM and software developer (.NET) and I can't stand people who face programming as nothing else but they way out to make big money with almost no effort.
... sure I don't consider going overnight in front of the computer geniality. but the few who matter, have done it more than once.
Closed. This question does not meet Stack Overflow guidelines. It is not currently accepting answers.
This question does not appear to be about programming within the scope defined in the help center.
Closed 5 years ago.
Improve this question
The company I work for has historically had very little process as far as software development. Currently we don't really follow any specific method. The problem is of course it makes it difficult to plan, successfully have a decent release or even attract good software developers.
I think I may be able to convince them to do some sort of Scrum process. Key however is getting management/owner buy-in. The idea of locking into specific features for any period of time I think scares them off.
Does anyone have any suggestions on how I can make my case?
So far I plan to:
Give presentation on how Scrum works. how I see it working with the people we currently have and how it will benefit the business.
Ask for training for specific people so we aren't "making it up as we go along"
Set a date to implement, there is some planning and loose ends I probably have to tie up to start a process fresh.
If your projects are like the standard / typical IT projects, then chances are your projects have failed, or been buggy, or cost too much, or didn't do what the customer (internal or external) needed, or took too long to develop.
If you are going to advocate a process, it needs to be shown that you will not lose flexibility just to have structure.
Points to make to decision makers:
Having a Scrum-like process will improve how much information that management has at its fingertips, and allow them to make decisions more quickly. Consider the scenario that you have a 6 month project. Well, with no processes, how do you know how much work is done until it is released? With Burndown charts, you can track how much time is left in a visible way. If you couple that with TDD, where you define say 100 test cases, they can see that 50% of the test cases are left to get working, but from the burndown rate there is only enough time to do 25% (remember Managers like it simple, so this isn't a perfect state of the project, but it is an easy to understand one that was better than what they had before). .e.g. they will feel more in control because the projects have better visibility.
Having process allows you to improve quality, which long term will result in less bugs, less time spent on bugs, more knowledge transfer (what happens if your star developer is hit by a bus), and all this means that the company will get developers focused on a better product than on continuously fixing bugs. e.g. this will save them money
A small set of changes will be implemented first. This will be a proof of concept, and safe and easy to back out of if needbe. e.g. this shows that you are mitigating perceived risk . And you need to mitigate perceived risk because that is what they'll be focusing on. That said, you will want to gather some data before you even make the proposal. Why? Good question: you need a baseline for 2 reasons:
You'll want to know how much the changes have helped. So you can propose more changes.
You'll likely have a manager complain about a problem while the proof of concept is going on. You'll want evidence that shows that problems in a chaotic process free environment are the norm, and this is not a worsening of the state, and perhaps a slight improvement. You can bet on something going wrong in a process-free environment. And you can bet that the proof of concept process changes will be blamed. So be ready for it.
In my experience it's easier to sell management on a design methodology or practice after it's been piloted once. I would cherry-pick a small project, usually internally facing if possible, and ask to "pilot" your new scrum process. Generally it's a lot easier to get people to buy into a pilot because they only have to commit on a limited basis.
As your new scrumified pilot project moves along, be sure to document (post-its, notepads, Word doc, whatever) how scrum is making your project more or less successful than the previous (lack of) method. Be brutally honest here, and try to quantify things in real terms whenever possible.
After the project completes, compile your notes and present to management your findings using the completed project as evidence. Use findings such as:
"product backlog provided users with real sense of progress on featureset X"
"pigs/chickens meetings style saved X man/hours a week by keeping meetings in control"
"sprints allowed developers to work more closely together and resulted in X% less buggy code"
Generally, if you can bring leaders to a spot where they can draw dollars-and-cents conclusions, they will go for a new product or methodology. Also, and this is important as well, be prepared to walk away from your original process ideas if you find them not bearing out during the pilot.
Good luck and happy productivity!
You can sell Scrum as a "No Lose" proposition. Look at what happens when you use Scrum:
All development work is always focused on the highest priority tasks.
Progress is 100% open, and inspected daily.
Users/customers get to examine the progress at the end of every iteration.
Shifting requirements are handled automatically.
The only reasonable objection that I've ever seen to Scrum is that it isn't really possible to predict how much a project will cost, or how long it will take. This is because Scrum acknowledges that everyone will learn as the project commences, and the requirements will change. Waterfall pretends to be able to do this, but we all know how well this works.
Run the Joel Test to determine how much work you have to do. If you are having trouble estimating release dates, look into Evidence Based Scheduling.
Provide some sort of argument that shows how Scrum will address past pain points experienced by the key decision maker. Extra points if you can also provide evidence that demonstrates this.
Keep in mind that it is also possible that you don't have a process because the management doesn't know and doesn't care about it. If your managers have no interest or no understanding of a process, such a process could also be started by getting all the programmers to agree to it (or at least team leaders) and telling new employees, "this is how things are done." Of course, it is necessary that you pick a process that is compatible with your manager's requirements if you do this (e.g. if your managers ask for daily updates on milestones, don't pick a process that has no coding for the first two weeks).
This is really only appropriate if you have a discussion with a manager and their basic reaction is "It doesn't matter, as long as you keep writing code." If you present a process as being a means to redistribute order of work done rather than as one which adds new work, you're more likely to succeed in such an approach.