Speed matters: Why working quickly is more important than it seems (2015)
jsomers.net
jsomers.net
I love this presentation because it adds strong economic arguments to a lot of practices that are quite common in software engineering but typically lack good argumentation as to why they are good. He covers a lot more ground in this presentation. Well worth your time if you are involved in any way in planning software development. I love his notions of option value and cost of delay. Cost of delay is exactly the economic argument why working quickly is important. Once you get it, is obvious that you should do agile, release early, and avoid embarking on long/uncertain refactoring/redesign efforts unless you can justify the payoff.
In short, any development that ships early maximizes the useful economic life on the market. If you ship something new earlier than your competitors, you get to charge a premium until they catch up. At some point your new thing becomes a commonality and the value drops. In some case, OSS just gobbles it up and the value becomes 0$.
So shipping 3 months earlier means you get to earn revenue for 3 months extra when it is still most lucrative. If you instead delay for 3 months because you are doing some thing that make things better in some way, you have to consider the cost of doing that AND the cost of missing out on that early revenue.
Don Reinertsen suggests you simply do the math consider the cost of delay when e.g. deciding on a big refactor vs. a lets get this to market ASAP type strategy. You don't even have to do a particularly good job at doing the math to out-compete gut feelings here.
And in the realm of programming, it’s often easy to take shortcuts even though later on they wind up costing more time for you or others involved: omitting documentation, not making something reproducible, poor code quality, etc.
Having spent most of my career in academia I see a lot of the work makes this trade off: efficiency is sacrificed for a sloppy, immediate “speedy” solution. You need to find a balance and think through problems, considering not only what is immediately in front of you but also what you’ll be doing in a month or year. Otherwise you’ll wind up with a project or codebase that begins to drag your productivity and output with it.
At least according to my experience and everyone I know in the industry.
I also thing it is vastly overstated how difficult it is to incrementally refactor a project which is "fucked". It's always possible to break it into smaller pieces, and gradually improve the codebase until it is manageable and easy to work with.
Yes, it's possible, but at what cost? E.g. long hours, unhappy devs, implementation taking too long.
It takes a bit more nuance to get a team to work that way rather than just enforcing Best Practice(tm) always, but in my experience it can be much more productive overall.
The point of code quality is to enable the company to change its mind or add new features. Over-perfectionism that gets in the way of this is going too far. At one company I worked for, the policy was not 100% strict DRY. Instead, we waited for 3 uses of the same idiom before we put it in the library. This prevented the collection of too many trivial methods, while still giving the benefits of DRY.
What if there's no automated testing worth a damn, and the penalty for a bug is very high? Been there, done that, and I have to say, you don't ever want to stay there!
You don't see good looking legacy, because it didn't survive.
Technical debt manifests in a code base, most generally as various forms of pain that occur when trying to add features. That's how the company "pays interest" on the technical debt. Coding so that you can always change your mind, however that is accomplished, is the key there. (This is one place where the "Law of Demeter" really shines!)
It's definitely a balancing act.
One extreme is always going for the instant gratification, paying no attention to how things fit into the overall design.
The other extreme is trying to design everything up front with no feedback and getting more and more divorced from the practical reality as time without implementation goes on.
At this mode, my mind uses all available time slots to make the best of them. For example, by the time my program compiles, my brain visualizes possible breakage scenarios and possible solutions, in case the compilation fails.
I could be on confirmation bias here, but this seems to be a neat trick to get more things done.
Surprisingly, at the end of the day (typically after two/three of these zones) I feel more fresh, satisfied and productive. But that's just me.
Which is that all accomplishments are meaningless in the end. And we are just getting older. Once a person dies they aren’t experiencing any stuff that came as a result of their effort. They may as well have just messed around and had a family sooner, or traveled, or not. It’s all meaningless anyway.
I can’t shake it. Anyone felt this way all the time? Any advice?
As for me, Yes, I might pick traveling and having fun over creating software. But I'm not from a wealthy background and my need for a better quality of life is satisfied by selling my skill for money. In that sense, I picked long-term quality of life over short bursts of fun and happiness.
Also, I do have fun coding so it's kind of alright for me.
So, consider your heirs, and work long enough to build a nest egg cushion the devote time to having a wonderful spouse, making babies and enjoying them.
Think of yourself as a part of humankind - you can contribute big or small to the goal of continuing the existence of humankind and its further expansion - with your work, raising children, other.
Ultimately, there is no meaning given to us from any source (given that you are not religious), and you search and choose your own.
I think starting your own project or startup can fall into that category, too. It can be very intrinsically fun and rewarding, and the idea is to do stuff that's intrinsically fun and rewarding while we're still alive, I think. Writing 2.0 of BigCorp's latest TPS management app carries none of that, and so it seems like a complete waste in the grand scheme of life.
Another option is to do things that really will have a significant positive impact on people's lives post-death, and despite the fact that you won't be able to see, feel, or appreciate the impact you caused after you die, it will still feel very worthwhile for the entire duration that you're still alive. Like helping fight global hunger or disease, for example. You'll forget that TPS 2.0 app a few days after you've quit that job probably, but if you quit a non-profit helping fight hunger and you made something that actually tangibly helped, you'll probably think about that your whole life.
Where anxiety and stress can come in for me is letting unsolvable, abstract-ish things hound me for a long time: "I have made a big mistake in life N years ago", etc.
Where I think urgency is hurting is breadth: I focus only on the easiest solution and long term view goes out the window so I need to intersperse walking or similar to let my mind wander to strange, stupid or potentially good ideas. Just my 2c, YMMV.
It obviously doesn’t sustainable though, both your work and health. It get a job done, but I know that it lack finesse quality, also all that stress definitely shorten my lifespan.
I disregard urgency when it exists for urgency's sake, coming out of business despair or a gut feeling. I will gladly work for 2 - 3 months on hyper mode, giving away my mental and physical health for a feature that will actually provide business value and for which the business can make a good case for why it's needed yesterday.
Working with multiple companies on executive and / or consulting levels, I feel like a lot of business rush is done because business 'needs it now' in the hope that it will improve the bottom line because it's easier to expect a miracle than it is to take a good look at your fundamentals.
The businesses with the most solid foundation I find are the ones that rarely need to rely on speed and instead rely on quality and produced value... once in awhile, a feature idea comes up that can propel the company - especially in the face of competitors - and it needs to be done as fast so that the producer can get as much value out of it before the competition catches up. That's a good reason to expect speed.
Today's society expects hyper growth, hyper speed, hyper scale, but that is because this is what is mostly advertised and we as humans have a tendency to expect the world to work as is advertised, when in reality, there are a lot of smaller companies that don't create anxiety filled working days for their employees.
I feel sorry for you.
As some who's been around the block a few times, I will never, EVER, give away my mental or physical health for someone's business.
My health is more important than their quarterly targets.
Slavery here really hyperbolic. We're talking about hard work that is both interesting and rewarding.
Your hard work is way more rewarding to your masters than to you, though.
There are plenty of people who do though I imagine.
I would say it differently though - do I have +-guaranteed massive return that over-compensates my extra effort and loss? If I get something meaningful, ie paid next 5 amazing vacations, can retire 1 year earlier, pay off massive part of mortgage or something on similar level, then I can at least consider it.
Then some kind of politics held the project up (likely for years), and they had me work a different project, so I quit, 'cause it was the success of that particular project that had me coming to work (it was gonna prevent traffic accidents, and it was fun to work on).
I have a healthier work/life balance now, but I don't care about what I'm working on. If they could get the wheels moving on that project again I'd switch back in a heartbeat.
I'm gonna die either way. If I have to put in a bit more work in order for it all to have meant something more than pointless meetings and products that nobody wants, then that's work I'm willing to do.
I've since left that institution precisely for what you mention, politics, and exactly like you I something think about working on that project again. I don't think any project I'll ever work on will be as satisfying from a societal perspective, as that one.
I will always look after my own health and will always try to maintain control, but each of us has his own context. "No product is worth that" is a general statement that might not apply to everything and everyone.
Fear and pity them all you want, but your health is not theirs. Don't give away the only thing you truly have, and the only thing you can never really get back.
I think it's also important to note the context of who is doing this kind of sprints. Does he have a family? Does he have other engagements? Does he have other health issues?
It's not for everyone, but I believe in my case, timing them properly, recovering properly afterwards and reaping the benefits has worked in my favor.
I might have given the impression that I am advocating for regular such sprints when in fact I have done this rarely, at pivotal moments, with years in between them.
Why don't you find personal avenues in which to exert that creative energy for you, your family, or society?
It's called a work/life balance. Work, and the related money are a means to being able to live in this society. And how we choose to use the money and power, is how we affect society. But running to work after everyone else "goes to sleep" doesn't appear as a healthy relationship.
And I know of nobody who said when they got old, "I wish I spent more time at work".
My current employer is shrewd in that they have aligned their interests to mine -- they have moved away most of the BS typical of engineering jobs. They have other people do the parts I don't enjoy doing, and they have given me wide latitude in Product and Product Engineering decisions. They have also given me wide latitude on task selection -- which tasks I take and which get assigned to others on my team (or others on other teams.) This isn't a constant state, it happens on several sprints a year.
On a side note, yes, I exert lots of creative energy in other domains. I'm on the board of a non-profit mentoring org, I'm an advisor to a nature conservancy organization, and I spend a huge amount of time with my family and children.
When I was single, I was part of a community food cooperative and several educational mentoring orgs. Now, the lions share of free time goes to children.
Side note -- if/when my employer turns into Initech and I'm filling out TPS Reports, you can be sure I'll be clocking out at 5pm.
I'm doing O.K., thank you for the concern.
> As some who's been around the block a few times, I will never, EVER, give away my mental or physical health for someone's business.
> My health is more important than their quarterly targets.
I wholeheartedly agree, however I believe there are some situations where such a choice is not so bad. If you're young, have no obligations, believe in what you do and will end up a lot better after the sprint, I think it's not such a bad thing. The most important thing I believe is being able to take the time off after said sprint to recover and get things back to normal. It shouldn't be a normality in your life and you should be in total control of the situation.
Also such sprints should be seriously spaced.
How will YOU end up better after the sprint if you're not working on your own product but on your company's? That will benefit your CEO and shareholders for sure but how does it benefit YOU exactly?
Keep in mind, you have one health in your life and when it's gone it's gone for good whereas there'll be plenty of world-changing unicorns where you can work yourself to exhaustion if that's what you wish. Nobody has ever regretted on their deathbed not having worked more for someone else.
Looking back over the years of software development I've been involved in (jeez, 27) and remembering the things that were developed with all haste, not one of them, not a single one, met with success. There were some that were developed quickly, usually to meet contractual deadlines, that were good and gave the customer what they needed, but even then, if they could have just waited one more month, I think they would have gotten something better and been more successful.
If I may add this, I believe two forces are particularly strong when leading us to deliver rushed solutions. One is this expectancy of the market to ever grow, faster, stronger, harder. All. The. Time. The second is stakeholder focus on the short term and using estimates as deadlines to 3rd parties.
These forces get translated by business people into the classic 9 women can give birth to a child in 1 month situation which we can actually deliver because most of them can actually get even more value from being able to sell maintenance.
Hey, as long as lives don't get lost from shoddy implementations all is well; oh wait...
I felt the same way until I caused permanent damage to my eyesight from stress. It isn't worth it.
This seems like a good idea for a browser extension. Maybe sometimes you want to undo the work that the website did making things go faster?
Artificially slowing down those time waster websites instead might do the trick. Not so slow it becomes unusable (or I'll turn it off) , but slow enough that over time I start associating it with a bad experience.
It's been a lifesaver in tackling my internet addiction. It has the useful 'delay option' and other nifty features that helped wean me off social media sites.
Personally, I have it installed on all my desktop browsers and Firefox on Android. I give myself half an hour quota per day for the time waste websites (Youtube, Twitter, FB, Insta, etc) after which a 60 second delay rule kicks in. I've found that my monkey brain almost never bothers waiting for a minute for the page to load unless it's absolutely necessary.
One goal is to quit something completely. For that, you probably want delay across the board so that it eventually saps the joy out of the experience.
Another goal is practicing moderation. With social media as an example, maybe you want to stay up to date on something, so 20 minutes a day is fine, but you don't want to fall into the trap of killing 3 hours on there. So, you could make it full speed for 20 minutes of daily usage, then gradually degrading from there. By the time the speed degrades, you've probably done all the truly appealing parts anyway and it's easier to walk away when nudged.
Obviously this only works over some kind of messaging service; face-to-face conversations have significantly different expectations around long pauses.
They both have the habit of very frequently interrupting me to ask me questions they either know or can find the answer to easily, or asking me to do things they can do for themselves.
I try to avoid just giving a blunt "no" because I'm also trying to instill in them the idea that family is about helping each other. So I try to say, "Yes, but I'm in the middle of ___. Give me a few minutes then I'll help you."
More often than not, as if by magic, it turns out they have the capacity to do it themselves all of a sudden.
Say you want to become a better writer. That's a broad goal and there are several approaches you could take towards that goal. On one extreme, you could read a lot about writing, write little, try to perfect things, and get something out once or twice a month. On the other extreme, you may write every single day and get it out in front of some people, get consistent feedback, learn, and improve. Both of these approaches will take time… the second approach, with the shorter feedback loop, can probably get you there faster though.
Going fast doesn't allow for secondary concerns and produces a single simple solution. Of course you can now do everything else it needs that you think of but code written this way with a clear core path is so refreshing from bottom-up generation where every detail has equal importance.
I even recall a time where I was performing a cascade of rebases and merges across three git branches and 4 or more submodule branches. Each feature update/sync took so many tries to get right. I eventually got fed up and just brute-powered through it without prethinking it. Because it was a short time from start to finish I was able to clearly see each commit in each tree I was working on without a pause wondering where I was or doing next. It was also less error prone than going slow. It fact I just did the whole process twice in a fraction of the careful time and compared results. No diffs were taken as correct and any diffs were examined and the correct variant used.
That doesn't sound like going fast, so much as "the simplest thing that could possibly work." Once you have that running, you can see where the shortcomings are and start to fix those. That will often proceed faster than trying to envision the whole complex thing up-front. (And that often results in a complete, but illusory vision, which ignores something which only becomes apparent once it's realized.)
Developing on a live fintech system is a different kettle of fish
Banks keep ledgers of transactions instead of just the balances, so that if an account balance is off, it can be recalculated from the transactions.
Transactions can be cancelled in case there was an error.
Invoices are often checked by humans before they are sent to customers, introducing an option to catch errors. If the customer receives a faulty invoice, they can contact whoever wrote it and ask for a correction. And so on.
Of course, there isn't room for error in every corner -- having a flaw in an authentication or authorization protocol could be very costly to recover from.
But in general, the assumption that in finance in general, there is no room for mistakes is simply false.
Then I can assume that all of your financial software has verification proofs and/or is fully model checked? If not, then clearly there is room for mistakes even in finance.
In software the material cost is not the cost of a concrete material such as wood (like you mentioned) but developer time
The other thing to bear in mind is that there is always more work to be done than time to do it. You have to factor in opportunity cost. As many other commenters have said, ultimately it's a balance which depends on your company's particular situation.
Thanks to version-control software it's also incredibly cheap to experiment freely with an existing code-base in a completely non-destructive fashion. I see no reason to treat existing code as fragile or precious.
I really disagree with this. Aiming for something you actually think is unhealthy means you either think you should be working unhealthily ( :( ) or you don't trust your measure of healthy (in which case recalibrate that).
So to become fast, learn to write code which is very easy to verify that it works. If you get surprised when things work the first time you try then you are not fast.
Fast iteration isn't about doing things faster. It's about requesting feedback more frequently. Which often means doing less, not more, and then evaluating the effectiveness of that small amount of work. Then adjusting your plans.
In writing, doing an entire book before your editor can tell you you're accidentally stealing a storyline from another author is bad. His analogies seem to be equating to this sort of behavior but missing the cause.
And I thought they wanted argue for faster replies...
But I think there is some truth to that. If a task seems overwhelming, it can sometimes help to look back at something you already completed.
I don't have that impression for programming at work, since I have a set schedule anyway, but for projects outside of that the kick-off seems to be the hardest part in awe of all the work before you. But then there is all the projects you already completed, which were mostly labor intensive as well.
It's a cute way to think in opposites, but unfortunately as punching advice, it's untrue. Punch force comes from body mass x acceleration, not arm mass x acceleration.
Responding fast to emails is just crazy. I lost count on how many urgent problems got resolved automatically. "Hey can you help me with this". 15 minutes later "Nevermind, found it".
I'm in the "Slow is smooth, smooth is fast" camp. https://www.lesswrong.com/posts/4FZfzqMtwQZES3eqN/slow-is-sm...
Seems to me, that the recent spate of online coding interview techniques are basically looking for people who type fast. I suspect that they're finding people who are actually "10X" coders, and people who have memorized the typing out of the solution.
(EDIT: A problem with such 3rd party companies, is that the incentives are subtly misaligned, as they are for any recruiter. If they can present a good image, and a steady stream of candidates who produce videos where they're going claka-claka-claka on the keyboard at a good clip, like a hollywood hacker (which runs and compiles) then they're golden. It doesn't matter to them, if they're missing out on good, thoughtful candidates. False negatives are the order of the day.)
I think the bigger misalignment is the time horizons for success (placement vs successful performance), but that's something that'll get figured out as the interview companies build up their track records and keep track of how long their placements stay in their new jobs.
I think waterfall style development works for one person inside one's own mind, but not beyond that.
A lot of the stuff I see from 20 year olds these days is monstrously over-engineered and Rube-goldberg like. Many keystrokes, sure, but less thought. This also yields bloatware that hogs resources and is hard to install and maintain.
http://www.ariel.com.au/jokes/The_Evolution_of_a_Programmer....
More thought and less typing young padawan.
But there are four big constraints on waterfall:
- it takes considerably longer
- it produces no visible results until right at the end
- all but the smallest changes require restarting the clock and budget
- it requires a small team at the start and a big one at the end (an observation dating back to Brooks)
Also, disclaimer, I didnt actually read the article :o
Trouble is that this approach rarely works beyond a very small team.
Fully agree, including some of the sibling comments.
Absolutely. Not saying speed is never an important metric, but virtually all complicated engineering and construction projects are difficult to schedule and exceed their time budgets. If the end result is worth it those birthing pains are quickly forgotten. But a poor output will never be forgotten.
Sure, agreed, which is not to say that time budgets don't exist, and aren't exceeded :).
edit: I'm 42, if the matters.
What they mean is that you can't be fast (on a track) without being smooth, and that typically beginner-to-intermediate drivers/riders are trying to be fast in ways that make them less smooth, and ultimately (and counter-intuitively) makes them slower. So the advice is to forget about being fast and concentrate on being smooth. The speed will come.
I do think this has value as an analogy for software, but it is in concentrating on your tooling and methodology - smoother operating will deliver real value faster than just trying to type fast.
In software smooth is fast because you didn't have to bug squash and refactor. But even if it isn't faster, the grandparent comments point was that it doesn't matter as much as being correct anyway.
Most problems in software engineering are caused by only optimizing for speed.
"A stitch in time saves nine."
"Haste makes waste."
Speed helps you deliver more.
Delivering something after it is needed is useless. Chronically delivering things the day after requirements change due to the world turning, is uselees.
Musicians and other technical performers practice both slow&carefully and quickly, so that over time they can be more correct and more quick (as quick as the task can appreciate)
> That doesn’t mean be sloppy. But it does mean, push yourself to go faster than you think is healthy. That’s because the task will come to cost less in your mind; it’ll have a lower activation energy. So you’ll do it more. And as you do it more (as long as you’re doing it deliberately), you’ll get better. Eventually you’ll be both fast and good.
Take the time now to do it as well as you know how to do; but you're also correct, and taking too long to do the thing is useless, as well as overbuilding.
It might in the short run.
Take coding hacks for example. It is fast in the short term, but might cost you double in the long term.
Not saying that quick hacks can never have their place, but saying that being fast might cost you time in the long run.
So faster is not always faster, and does not always help you deliver more, depending on the timeframe.
I would a always tell my residents that to start a medical practice, the 3 "A"s are important, and in this order. Availability, affabilty, and then ability.
People have a problem, and they want it taken care of. If you can be that guy, or company, that's there for them, then you'll get there business in the future, as long as you do a good job.
Some choice quotes from the post:
> It is a truism, too, in workplaces, that faster employees get assigned more work.
> Whereas the fast teammate—well, their time feels cheap, in the sense that you can give them something and know they’ll be available again soon.
> push yourself to go faster than you think is healthy.
Dude that’s not a feature. It’s a bug.
Working fast can be good, solving the right problem at any speed is better.
Personally I have not found my speed matters so much as my mental state. When all the things come together, I know what problem I'm solving, I know how to solve it, I have time to focus on it for many hours... That is when I can enter a state of flow and build a great head of steam productivity-wise.
Finally I think continued long-term attending to something can trump speed and overcome all kinds of obstacles that prevent progress at all let alone progress at some arbitrary sense of fast.
As Carlin said, any person driving faster than you is a maniac. Anyone driving slower is a moron.
quality = w_0 * time_to_market + w_1 * correctness + w_2 * performance + ...
The weights w_n vary across domains. As with most things, it's about finding the right balance.
1. Suppose you want to draw a line N pixels long. It costs X to draw a pixel, Y to copy and Z to paste. What is the cheapest (fastest) way to build the line as a function of N?
2. So I think in general, you may start slow, but them exponentially speed up, overtaking the spaghetti architecture guy.
Working quickly helps you validate your idea fast and iterate.
Working quickly can also mean not putting enough thoughts into planning and researching, so you just waste your money, your time, or your credibility faster.
Working quickly can get you promoted.
It can also ruin your health and personal life.
"premature optimisation is the root of all evil" is a saying that came from 1974 when computers were slower, languages were less effective and development processes immature. Today you should make your software fast.
At some undefined point you may realize that you are trying to solve an undecideable problem. You can spend an infinite amount of time, and not get an answer.
You cannot build an algorithm to determine the amount of time that an arbitrary problem can be solved in, or if it's solvable in the first place. You must make a choice based on imperfect knowledge.
When you build a framework to support a feature you decide to never implement, that's premature optimization.
I'm not concerned with optimizing code, generally. I'm concerned with optimizing developers. Don't waste time on unnecessary work.