Cost of Making Software - Quantitative Model

Discussion in 'True Programming Stories' started by Dan Allen, Sep 2, 2017.

  1. Dan Allen

    Dan Allen Administrator Founder Not Banned Radio Button Problem - Leader

    Quantitative eEnough for purposes at hand.

    While we have a moment without the thunder of trains, here is what I have. I have not put these numbers to Arabic symbols, because all of them are orders of magnitude greater than anything we can manage. I have seen numbers put together exactly with these rules of thumb and some of the work that followed went smoothly in accordance with these guidelines. I have seen others run years and millions of dollars over budget.

    All the guidelines end up in spreadsheet, with tasks, people assigned, checkpoints scheduled and costs tabulated. There are various kinds of programmers and various kinds of managers. A lot of yakking on software philosophy is normal, because humans do not have a good handle on how to do this. Among the crucial numbers that cannot be left out, in some ways the only ones that matter are FTEs (full time equivalents matching specific tasks to specific people with the skills to do them). Those FTEs generate 80% or more of the cost. A lot of that is salary, a lot of it office space (including maintenance, electricity, etc), equipment, taxes and health insurance. In 1990, an FTE cost business about $100,ooo per year on average. Business are comfortable with that number. Businesses unable to afford that number look for ways to do it for less and endure agony. Businesses with big money don't bother trying to cut from the $100,000 benchmark, because they figure it's cheaper to pay it an reduce hassles. Managed well, those numbers can get it done.

    The only way this applies to us is it tells us we are on a hard road by a huge margin. I think our experience matches that part of the model.

    Another Angle on the What's Going On
    I wanted to understand this better. I started seeing how software development works. These are the my numbers, my original work, you won't see this anywhere else. To me, this just puts into words what I think we both can see. I think you will see we do not need Arabic numbers to understand what we need to know about A HUGE TANK OF MONEY

    This week, luck smiled and brought a real world scenario to my attention where I have been able to estimate the situation using this quantitative model. Sam told me about a programmer working for a start up that thinks they can score $billions. Sam is their lawyer. I said I want in on the programming work. Sam said they have their guy already. My model says a billion dollar enterprise will need more than one programmer and will find itself more than willing to spend a modest sum to add to the software one programmer will provide. My model says there is not any one programmer who is to going to provide everything needed or desperately wanted for a billion dollar operation, especially one depending on new software for the basics work of the business. "At some point he is going to need help..."

    The bad news they might not need help, should the unlikely possibility of business failure occur before they can fund the help they will need. Sam is sure this startup has what it needs. Then, he told me the business. I think there is a non-zero probability the bitcoin mining operation will not produce the results these guys expect. Maybe they will do well, some people found gold in California, but mining... that industry notorious for insane optimism? The programmer is the same one who told me he has a genius working for him who can write 5,000 lines of new code per day every day. I asked if I could see some of it and since then, that guy and I have been wary of one another. Can you think of any reason why he would become wary when I am doing nothing to threaten him? I just cannot think of any possible reason someone would react that way, I mean, assuming he is not full of shit, but why would he be full of shit, that is just ignorant. Nobody is ignorant any more, as a result of the internet.

    Anyway, back to business. If that bitcoining deal makes a billion dollars they will looking for programming help, so says my model. I'll let you know results when I have them.

    Why so MUCH Snakeoil?
    When we met, I was puzzled by the snake oil we all were swimming in. It's taken a long time to sink into my thick skull why there is so much snake oil. I could see why there is a lot, but I didn't know is it is the only thing anyone is making. No one knows what I am telling you. What I am telling you is no one knows how long software work takes, because no one does the work.


    Well, someone is doing some work, we can see some software. It comes from somewhere. Almost all of it is last year's software, which was from the year before. Here and there you have Schrivner. How long was it in development? How complicated it its function compared the function of your business? Answers to those questions is real information that almost no one has. Mr, Scrivner does. And look what he is doing. He is getting out, letting someone handle that shit for him, Writing fiction is easier for him.

    What other serious software can you name? There is a website template industry in India that has some sort of factory way of cranking that shit out. I guarantee you, there are armies over there with software they use for creating that stuff. I think their work is a an unappreciated pinnacle of software.

    Xenforo is another pinnacle. Look what it took. It took the reining world champion, VB net, getting bought by assholes prompting the top programmers leave and build their vision. They only a few years battle in court before settling with and thus they were able bring Xenforo to the world. A pinnacle and very weird software. Now, WE can see the scale of work that has gone into making it by looking at the product and we can see the connection between the work and the product. Xenforo is highly appreciated software, it seems mostly by people in awe of the result, where the awe is significantly boosted by wondering how it was done. It was done with as close to a miracle as you can find, almost. How many top developers are there from world-leaders in a certain type of software breaking off, winning in court, and putting together their vision of how it could better given a second shot at doing it? I don't think there are thousands of those around.

    Big Corpo
    You have big corpo software, such as PeopleSoft and SAP, which is braindead simple compared to what we have made for you . Their idea of a great app is they can have all their employees on a list and they work for years to achieve it. Their scale works against their software efforts, so they have to make everything simple and ironclad. Easy money for a programmer who doesn't like her job.

    Free Open Source
    This video explains free open source. Software is made. There is no money in making it.

    Bill Gates
    Very few people have a clear eye on these metrics. We usually cannot tell people about it, because no one believes us. Somebody like Gates, he knows to count on x hundreds of thousands or millions of dollars for certain work. He can afford it so he can deal with it like that. At that level he can't see see how monumental an effort every programmer he pays has to put in. Bill sees it as salary to an employee who of course is good or he would have hired the person. He doesn't get it that the smallest project he can think is a major multi year life event for every single person involved. For a time, he made up for half assed software, because the scope of DOS was small enough that he and a few dozen nerds could do it. Bill very possibly might not know that his industry still has not developed a capacity for routine development that is above the threshold of HORRIBLE.

    Here is some math for software


    A System Like Yours

    For more than your system:
    ((Work for system like yours
    Number of times bigger)
    *PHENOMENAL BUSINESS MANAGER)^an exponent bigger than one

    *does not program or operate the system. Views heroics as something for weak people. Imagine Mr. Spock joining with Scrooge to make a test tube baby. That baby has the right stuff for becoming a business manager.

    Yields: rich business manager, everyone else lives on fumes. Software might be produced might not.

    A HUGE TANK OF MONEY because that is the only way they can scale, paying people to give over their lives to a cause, mercenaries. They look for people who share the vision but they are so hard to find, big compromises get made here. This is an expensive proposition, but so few can do it, the competition among those who do is LOW, and there are huge markets, big enough to REPLENISH and PAY A PROFIT TO THE ENTITY SUPPLYING THE DEVELOPMENT MONEY. That entity gets rich, everybody else involved including the heros who do the work get salary no where near commensurate with their motivation and goals and then the whole thing down sizes, the heros get fired or put on other projects that take their life for inadequate compensation

    That is pretty close to the formula and with that in writing, it is pretty easy to see what the experience of the heros is going to be. They are going be hearing a lot of trains, but that will be mitigated by the business manager who knows he has to pay enough and distract enough with company bullshit (parties can go here, but I don't mean the party you are planning, I am talking about the annual golf day and generic holiday party at year end to keep the herd wrangled and celebrate the fortunes of the generous people who got rich and know they better at least butter a slice or two of bread for the people who did almost all the work but not the key work the rich person did, the one who wrote NO software that year, due to being too busy with phone calls, meetings and figuring out who can be squeezed and cutting everyone and everything else loose. Your party is void of that kind of thing. You are an honest GREAT worker who performs hugely more great than normal workers.

    Any war veteran will tell you, they didn't fight for Uncle Sam, or to defeat the bad guys. Truth be told, I think many of them (not Israelis dealing with Palestinians) discover the bad guys are our brothers and sisters, trapped in a situation that could just as easily have trapped us. It might have started out as cause, but those perspectives wear thin after months or years of battles, were you see friends or yourself blown to pieces on a regular basis. I don't think it is possible to imagine what it is like, but we can pick up some information from what the people who have done it say. They say they fight for the soldier next to them. I think that is how we heros are. Not many soldiers who have really been in the shit are wanting to start over once they have survived a campaign for its duration. Plenty do it anyway, because they don't know what else to do.

    My Objectives for this Week
    This math is not precise enough to say whether I can get that list done this week. I do know I have been unable to meet objectives estimated similarly for a long time.

    There is something else I know. It's a secret weapon. I have not been able to bring myself to use it for awhile. I used to use it all the time, to meet deadlines. I don't know if I can still use this weapon. I am going to try. I'll let you know how it goes.

    The target I am going after first target is the hardest: isolating the data on transactions after purchase, matching them to the purchase, tracking status and executing actions where needed. Were it not for the extensive effort from March - May this year to improve reliability of fulfillment, it would be hopeless to pull this together in a couple days. I had to put this work aside before it was 100% nailed down, which is at this point is unfortunate. I don't remember what the reason was, I think it was to focus on affiliates and not long after that is when we got JEM. JEM has saved us a lot of time on the shopping cart and affiliate processing, even though we are using a bare shell of it. If I had to have formulated that from scratch, that would have been tough. Turns out there is not much to it, but it would have been too exposed to raw elements for me to move through it quickly the way I was able to with JEM. As slow as it was, that was quick for what it was.
  2. Dan Allen

    Dan Allen Administrator Founder Not Banned Radio Button Problem - Leader

    Another application of the model came up when Gigster told me their plans for automatic custom development. Their plan is to increase the efficiency of custom development enough to facilitate huge cost savings no one else could match without their automation. I was able to tell them that plan will not work in that time frame, because they have not reached the point in their work where they are running up against the granularity problem. The granularity is that the cost of complete control of what you make is you have to address so many micro details to make a finished product, the time to do those little things becomes the bottleneck in development. If they are onto the granularity problem now, they are not going have it solved in 3-4 years, without limiting their definition of custom development to a tiny subset of what custom development can entail.

    I told them the best solution to the granularity problem for now is to have an effective work force of relatively lower paid workers. They seemed interested in that line of reasoning.

    None of this even starts touching the almost universal requirement to integrate custom development of new systems with previously-existing systems that are not being replaced, or the requirement to change those systems either as part of the integration or not.

    Our platform of universal hookups, a training system with dripped lessons, and some form of graphics editing software is looking pretty slick to me. I think that is the first time I have mentioned the graphics editing.
  3. Dan Allen

    Dan Allen Administrator Founder Not Banned Radio Button Problem - Leader

    I bet I can start scoring WordPress gigs
  4. Dan Allen

    Dan Allen Administrator Founder Not Banned Radio Button Problem - Leader

    I bet I can start scoring WordPress gigs

    I know a place with lots of WordPress expertise.

    This is part of my answer to the high cost of building software. It is too expensive except for enterprises with millions in revenue and high profit on that revenue.
  5. Kirsten Bolda

    Kirsten Bolda Administrator Radio Button Problem - 2nd Demo Parts A and B Founder Not Banned

    I started working with this and am in the midst of this course: Create A Custom Responsive Website for a Client
    The work in progress is here:

    A lot of this course is review, but one thing I've noticed is that the WP theme required for the course eliminates the need for enqueuing the child theme myself, and so removes potential for error. The premade theme also makes the site responsive without any effort on my part. I will report how this is going once I'm further in. My goal is to utilize the shortcuts the WordPress theme provides without sacrificing the flexibility of coding stuff myself.

    What is the place you have in mind with WordPress expertise?

Share This Page