If you have additional ways to avoid mistakes, please share them.
In item 2: 'Do your research first' I note that the trial and error method of learning was more prevalent in the past. I always enjoyed solving problems this way but it sure took a lot of time and there were the inevitable mistakes along the way.
The Internet has become such a wonderful source of information that even inexperienced developers can solve complex problems simply by learning how to best use search engines.
Fortunately not all solutions can be found on the Internet. I say fortunately because good problem solving skills are still required - and problem solving is one challenge that every good developer truly enjoys.
Discussion on:
View:
Show:
You forgot the most important. Do your design first. In our agile world so many forget that. It is possible to design excellent software.
You, as most everyone also over emphasize testing. I have started to say "don't test - design". Of course that is a little rhetoric, but there is a lot of truth in it.
You should write correct code that does what it shall do all the time. You do that by designing first, creating a simple design document that will be thrown away as soon as you have finished the first release version. Then you write a few lines of code, execute, a few more lines, execute, and you know it's correct. Of course this is testing
But the aim is to understand exactly what you write and that it's correct according to a design.
Then don't listen to users!
They are not software designers, and can not design the code, not even the user interface for you. There are a few exceptions for user interfaces with exceptional users, but otherwise you have to do the design.
To do this you have to listen to your users!
But not to do what they say they want, but to learn what they really need and then deliver something better than they expect. The wow effect is close, but it does require a lot of work.
To be a professional developer you have to know better than your users their requirements and needs. You can only do that by learning from lots of different users by communicating with them and studying their profession more then they do.
Lots of work, but your aim is excellence and nothing less is possible.
You, as most everyone also over emphasize testing. I have started to say "don't test - design". Of course that is a little rhetoric, but there is a lot of truth in it.
You should write correct code that does what it shall do all the time. You do that by designing first, creating a simple design document that will be thrown away as soon as you have finished the first release version. Then you write a few lines of code, execute, a few more lines, execute, and you know it's correct. Of course this is testing
But the aim is to understand exactly what you write and that it's correct according to a design.
Then don't listen to users!
They are not software designers, and can not design the code, not even the user interface for you. There are a few exceptions for user interfaces with exceptional users, but otherwise you have to do the design.
To do this you have to listen to your users!
But not to do what they say they want, but to learn what they really need and then deliver something better than they expect. The wow effect is close, but it does require a lot of work.
To be a professional developer you have to know better than your users their requirements and needs. You can only do that by learning from lots of different users by communicating with them and studying their profession more then they do.
Lots of work, but your aim is excellence and nothing less is possible.
That's been an adage of ours for all our working life. What it means is that what users see on the surface has to be a good match for what they say they are looking for, but what exists underneath has to be well engineered, future-proof, and covering all the cases the users have not even though about. The "mistake" is implementing what the user asks for, without engaging your brain.
Can't do 1,they read your final word and won't tell me. 
Accepted wisdowm is good, but don't take it at face value. It was right then, for them, with those. The one constant is change, don't ignore it to avoid thinking.
Have a plan, yeah right. Remember the plan was a hope, a wish, a fiction to provide an illusion of control. If the illusion breaks down, your wish is not granted, your hope shattered on harsh reality. Change the damn plan, and plan better netx time. If the plan wasn't followed there was a reason, and it could be that the plan was crap. Just because your boss made it, doesn't mean it's right, in fact if you were not allowed to contribute to it, it probably guarantees it was complete bollocks.
Create and adapt standards and templates to meet your needs then follow them, until they don't help, then adapt them again. Poor choices are as bad as none at all.
Revsion control is a mechanism for sharing already written code. The key communication is why it was written and why like it is. Group think sessions are my favourite tool in a team, get a consensus in approach, document it, institute a frame work based on it, use it specify the immplementation detail. Revision control is bog all use until you know which piece of code you need to look at. It's major purpose is an audit trail, a back up (assumimg you remembered to back up the host!) and a way of rolling back in the event of a bad idea.
Trial and error, or more likely suck it and see is a perfectly acceptable way to learn. Even if research turns up some "answer", you still have to find out if you asked the right question. Identify unknowns, work on knowing them. Then expect what were unknown unknowns to now be unknowns. Iterate towards comprehension, instead of waiting for it.
Allow enough time. You must be new. Time is set by the business, they want it yesterday, that's how long you have. Don't whine about it being impossible they don't give a crap.
Reuse abuse is one the biggest failings in software design. Remember as soon as you change it, it isn't proven anymore. Remember common code solves a class of problem where required and could cause one everywhere it is. It's not a free lunch.
Checklist are just another standard or template. A decent one particulary for admin style tasks. Keep them up to date though. If the step has gone remove it, if there is another add it. If in certain scenarios a step is ambiguous, document it.
Plan so you can test right from the get go. Every confirmation you are on track has value. Test after you've gone too far down it, is a lot of luck and bugger all judgement. Identify your deliverables, define key properties of them. Plan a test, do the work, do the test, then take you next step.
Unit test, autmation test, sanity checks, peer review, dedicated testers, a demo version, a key customer who might trial it in the real world, damn straight, do all of them if you can.
I've made lots of mistakes during my career. being creative requires mistakes to be made. Fear of error means you never get better, which by definition in a constantly changing environment means you are getting worse. Sometimes they can turn out to be more useful than successes.
Don't plan to make no mistakes, plan to minimise the consequences WHEN they happen. If you aren't making mistakes you aren't trying.
I give the way to make less mistakes doing anything.
Identify assumptions, and review them constantly.
Accepted wisdowm is good, but don't take it at face value. It was right then, for them, with those. The one constant is change, don't ignore it to avoid thinking.
Have a plan, yeah right. Remember the plan was a hope, a wish, a fiction to provide an illusion of control. If the illusion breaks down, your wish is not granted, your hope shattered on harsh reality. Change the damn plan, and plan better netx time. If the plan wasn't followed there was a reason, and it could be that the plan was crap. Just because your boss made it, doesn't mean it's right, in fact if you were not allowed to contribute to it, it probably guarantees it was complete bollocks.
Create and adapt standards and templates to meet your needs then follow them, until they don't help, then adapt them again. Poor choices are as bad as none at all.
Revsion control is a mechanism for sharing already written code. The key communication is why it was written and why like it is. Group think sessions are my favourite tool in a team, get a consensus in approach, document it, institute a frame work based on it, use it specify the immplementation detail. Revision control is bog all use until you know which piece of code you need to look at. It's major purpose is an audit trail, a back up (assumimg you remembered to back up the host!) and a way of rolling back in the event of a bad idea.
Trial and error, or more likely suck it and see is a perfectly acceptable way to learn. Even if research turns up some "answer", you still have to find out if you asked the right question. Identify unknowns, work on knowing them. Then expect what were unknown unknowns to now be unknowns. Iterate towards comprehension, instead of waiting for it.
Allow enough time. You must be new. Time is set by the business, they want it yesterday, that's how long you have. Don't whine about it being impossible they don't give a crap.
Reuse abuse is one the biggest failings in software design. Remember as soon as you change it, it isn't proven anymore. Remember common code solves a class of problem where required and could cause one everywhere it is. It's not a free lunch.
Checklist are just another standard or template. A decent one particulary for admin style tasks. Keep them up to date though. If the step has gone remove it, if there is another add it. If in certain scenarios a step is ambiguous, document it.
Plan so you can test right from the get go. Every confirmation you are on track has value. Test after you've gone too far down it, is a lot of luck and bugger all judgement. Identify your deliverables, define key properties of them. Plan a test, do the work, do the test, then take you next step.
Unit test, autmation test, sanity checks, peer review, dedicated testers, a demo version, a key customer who might trial it in the real world, damn straight, do all of them if you can.
I've made lots of mistakes during my career. being creative requires mistakes to be made. Fear of error means you never get better, which by definition in a constantly changing environment means you are getting worse. Sometimes they can turn out to be more useful than successes.
Don't plan to make no mistakes, plan to minimise the consequences WHEN they happen. If you aren't making mistakes you aren't trying.
I give the way to make less mistakes doing anything.
Identify assumptions, and review them constantly.
Checklists are a must. We all need to recognize we are busy and often interrupted. Let the checklist be our memory.
Ha ha! Very clever of you!
Normally I don't put the '10 things' items in any particular order. For this article it made sense to order the items to match as best as possible the Project Life Cycle. Maybe I should have put item 2 first!
Normally I don't put the '10 things' items in any particular order. For this article it made sense to order the items to match as best as possible the Project Life Cycle. Maybe I should have put item 2 first!
No matter how stupid a mistake may seem, it has a lesson to be learned behind it. Furthermore, is a questions of honesty and integrity for owning up on ones actions. I am a PMO manager for a large multinational mining organisation and would like to sleep easier knowing that our team will homour the code of conduct and professional responsibilities. Sometimes "stupid mistakes" can cost an organisation millions- chances are, mitigation actions can be drawn up sooner if more info. about a problem is known upfront. Just think what would now happen to the "google buzz" test team if they had known about the privacy issues and "bugs" in buzz and not reported it. Bigger picture - consequence such as those sighted in the conclusion of the article can be attributed to environments with low maturity and/or emotional intelligence.
I agree. Unless Alan meant that the "mistake" would have no harm done, which is very rare.
Alan - that was a clever ruse to get us writing replies! I had posted the following before reading the further article and I will willingly admit my error!
I couldn't agree more. Hiding mistakes is a disastrous idea with potentially disastrous consequences. A project leader needs to foster a culture where people will admit their errors, no matter how stupid, and learn from them. In the example how could the writer be 100% sure that the new files he copied were exactly the same as the ones he'd deleted?
I couldn't agree more. Hiding mistakes is a disastrous idea with potentially disastrous consequences. A project leader needs to foster a culture where people will admit their errors, no matter how stupid, and learn from them. In the example how could the writer be 100% sure that the new files he copied were exactly the same as the ones he'd deleted?
I agree, I'd much rather have people realize a "stupid mistake", own up to it, but most importantly LEARN from it. Then I know there is growth and integrity!
Even you're talking about a product (project development, may be a system), you've described the ITIL V3 Foundation core (IT Service Management): V model, communication policy, Catalogue, RFC approve after CAB, etc etc. That's great, because ITIL has bourned as good practices after a long research about IT best practices inside various world businesses. You can also express in this way: even you're an IT professional don't think to be the only in the universe fighting against a big project; be honest, take your time, share ALL in a trasparent way, learn from and collaborate with others (not only IT); in this way the goal will be yours. Isn't it?
You had me going until the "Final Word". While you should immediately resolve your mistake, in the efforts of fostering your very first point, you should be willing to relay and even laugh at your stupid mistake. How wrong would it be if a colleague made the same mistake you did because you were too embarrassed to relay your mistake to them????
Finally, I've always said:
"If you are an active developer (and haven't moved on to other responsibilities), you should be able to look back at the work you did a year ago and see ways you could have done it better, even if what you did a year ago was quite good. Otherwise, what have you learned in the last year?"
Finally, I've always said:
"If you are an active developer (and haven't moved on to other responsibilities), you should be able to look back at the work you did a year ago and see ways you could have done it better, even if what you did a year ago was quite good. Otherwise, what have you learned in the last year?"
If you watch the TV show "House", you know that many of his miraculous cures come when he discovers the client's lie. Same is true in IT Projects. Few clients know the real issues, so they tell you only part of the truth. Your challenge is to uncover all the issues and build them into the plan.
A sophisticated manager understands that people make mistakes. We're all human.
Now, two things have to happen :
1. The mistake has to be fixed.
2. The process that allowed the mistake to occur has to be fixed. And communicated to all. Chances are if you made that mistake, others could have as well. So, the experienced manager calls a meeting and explains it to all.
Now, two things have to happen :
1. The mistake has to be fixed.
2. The process that allowed the mistake to occur has to be fixed. And communicated to all. Chances are if you made that mistake, others could have as well. So, the experienced manager calls a meeting and explains it to all.
... credited to Henny Youngman - http://www.funny2.com/henny.htm
Patient: "Doctor, it hurts when I do this."
Doctor: "Then don't do that!"
And then there is the saying in basketball, "No harm, no foul."
I see that there are a lot of you who think I should have shared my mistake with others.
The two words "really stupid" and the fact that the mistake was fixed before it could be noticed are the reasons why I felt no need to share my mistake. I could be wrong, but I could see nothing to be learned by others:
Alan: "I deleted some system files on a client's PC."
Boss: "Don't do that!"
I agree 100% that if others can learn from your mistakes you should consider sharing it.
I would classify how to handle mistakes into three categories:
1. The mistake becomes known to others and fixed - Read and follow Calvin Sun's advice at http://blogs.techrepublic.com.com/10things/?p=270. Take responsibility for your error and try to make something positive out of a bad situation by helping others learn from your mistake.
2. The mistake is not known to others, fixed and others can learn from your mistake - Consider telling others so they can learn from your mistake. It's your choice. You have to weigh any benefit gained against any personal damage. Sharing too many mistakes can be item one on your next performance review.
3. The mistake is not known to others, fixed but there is nothing to be learned from your mistake - It might make for some light humor (he did what?!) but what else is there to be gained? It is even more unwise to share your 'mistake' with the client. No harm, no foul.
We all make dozens of 'mistakes' a week that we find and correct. I used to have to participate in a teleconference call at one of my jobs. It usually lasted 45 minutes to an hour. I can't begin to imagine the litany of mea culpas if everyone on the team decided to share every little mistake. In fact, I don't remember anyone volunteering to share any of their errors, big or small.
I do however remember the big mistake I made and I am happy to share it with you here. I submitted a request to reindex one of our database's indexes. The request was queued and never ran because, unknown to me, the reindexing could only be completed when there was no user activity. By the morning, the entire system was locked up and yours truly got the opportunity to step forward, take responsibility and explain what I had done to everyone on the conference call.
I am encouraged to hear so many of you say that you should take responsibility for your mistakes. I totally agree. I just don't think that it is wise to share every mistake and, in good conscience, I can't recommend that you share a mistake made in scenario two or three since it could damage your career with little or no gain. That has to be your choice.
Patient: "Doctor, it hurts when I do this."
Doctor: "Then don't do that!"
And then there is the saying in basketball, "No harm, no foul."
I see that there are a lot of you who think I should have shared my mistake with others.
The two words "really stupid" and the fact that the mistake was fixed before it could be noticed are the reasons why I felt no need to share my mistake. I could be wrong, but I could see nothing to be learned by others:
Alan: "I deleted some system files on a client's PC."
Boss: "Don't do that!"
I agree 100% that if others can learn from your mistakes you should consider sharing it.
I would classify how to handle mistakes into three categories:
1. The mistake becomes known to others and fixed - Read and follow Calvin Sun's advice at http://blogs.techrepublic.com.com/10things/?p=270. Take responsibility for your error and try to make something positive out of a bad situation by helping others learn from your mistake.
2. The mistake is not known to others, fixed and others can learn from your mistake - Consider telling others so they can learn from your mistake. It's your choice. You have to weigh any benefit gained against any personal damage. Sharing too many mistakes can be item one on your next performance review.
3. The mistake is not known to others, fixed but there is nothing to be learned from your mistake - It might make for some light humor (he did what?!) but what else is there to be gained? It is even more unwise to share your 'mistake' with the client. No harm, no foul.
We all make dozens of 'mistakes' a week that we find and correct. I used to have to participate in a teleconference call at one of my jobs. It usually lasted 45 minutes to an hour. I can't begin to imagine the litany of mea culpas if everyone on the team decided to share every little mistake. In fact, I don't remember anyone volunteering to share any of their errors, big or small.
I do however remember the big mistake I made and I am happy to share it with you here. I submitted a request to reindex one of our database's indexes. The request was queued and never ran because, unknown to me, the reindexing could only be completed when there was no user activity. By the morning, the entire system was locked up and yours truly got the opportunity to step forward, take responsibility and explain what I had done to everyone on the conference call.
I am encouraged to hear so many of you say that you should take responsibility for your mistakes. I totally agree. I just don't think that it is wise to share every mistake and, in good conscience, I can't recommend that you share a mistake made in scenario two or three since it could damage your career with little or no gain. That has to be your choice.
So I got out the backup...
Oh there was no back up, or it woudn't restore, or those files weren't on it, or you couldn't extract just those files...
Well there's the real coke up then!
The fact that you manually deleted the files is irrelevant.
A programming error could have done it,
A power cut, a lightning strike, pour your coffee down the cooling vents. A hard drive failure or corruption.
Your 'mistake' was fortuitous!
Learning it from it, is not "don't do that again".
It's that your client was a clock cycle from disaster, is a clock cycle apparently, since you didn't ameliorate the risk in any effective way.
Oh there was no back up, or it woudn't restore, or those files weren't on it, or you couldn't extract just those files...
Well there's the real coke up then!
The fact that you manually deleted the files is irrelevant.
A programming error could have done it,
A power cut, a lightning strike, pour your coffee down the cooling vents. A hard drive failure or corruption.
Your 'mistake' was fortuitous!
Learning it from it, is not "don't do that again".
It's that your client was a clock cycle from disaster, is a clock cycle apparently, since you didn't ameliorate the risk in any effective way.
http://www.cbc.ca/health/story/2010/02/24/surgical-checklist.htm
Above is a recent article where a breast was removed by mistake. For the 2nd time by this surgeon. The 1st time it was hushed up. Now, infer what you will.
Above is a recent article where a breast was removed by mistake. For the 2nd time by this surgeon. The 1st time it was hushed up. Now, infer what you will.
Regarding item #1
Why would I want to learn from mistakes - I would rather learn from successes - I see a lot of Project Groups, Sales Team, and corporations analyzing what went wrong instead of what went well.
Why would I want to learn from mistakes - I would rather learn from successes - I see a lot of Project Groups, Sales Team, and corporations analyzing what went wrong instead of what went well.
Shawn,
Why would anyone want to know history?
http://thinkexist.com/quotation/those_who_don-t_know_history_are_destined_to/346796.html
I agree with you but you would be missing a big part of the picture.
Why would anyone want to know history?
http://thinkexist.com/quotation/those_who_don-t_know_history_are_destined_to/346796.html
I agree with you but you would be missing a big part of the picture.
Good article. All 10 ways mentioned are very effective and practicle. Specially proper and thorough testing with good collaboration among team members helps a lot to deliver good product.
- Keyboard Shortcuts:
- Prev
- Next
- Toggle
