Mar 02

Hey, software guy…. “How Long Will That Take?”

I was inspired to write this blog post in response to a post I came across today on LinkedIn about sizing software projects. (link below)

Sizing software projects is the thing that most everyone gets wrong, its hard and almost impossible to get an accurate estimate, why is that?

Well apart from scope creep, budget changes and all the usual business variables mentioned by Patrick in his blog post, developers and product teams will never be consistent in their output, not even if you average it out over a long time which is what SCRUM/other agile methodologies suggest when trying to establish velocity – that simply does not work, it is a fallacy. Writing software is an iterative and creative process so “how someone feels” will change output, and I am not talking about how someone feels about the project or work, I am talking about how someone feels about their cat dying, or their wife being pregnant or political changes or the weather, or the office banter or how unwell/well they feel today, in fact “life” guarantees this.

So I am going to be a bit outrageous here and suggest an alternative way of thinking about this. Let us start with asking the most important question – “what is the point of estimating”? there are only two possible answers to that question…

1. You are going to undertake some work which you will charge your client for so you need to know what to charge them.

The only possible way you can give your client a fixed price for work that is essentially a creative process is by substantially over-pricing the work estimate and giving yourself lots of fat in the deal to give you the best opportunity of making a profit at the end of it. If you think that you can ask a team of developers to tell you how long it is going to take so you can make a “fair” markup you’re deluded. The best option you have in this scenario is to work backwards, you need to understand the need the client has at a high level, then you need to establish the value that your customer is getting from the thing you would deliver, then you put a price on it, you are looking for the best possible price the customer is willing to pay, you should not at this point trying to establish “how much will it cost”, you must be asking the customer “how much are you prepared to pay”. Once you have a number, now you can work with your developers, but instead of saying “how long will it take” you are asking “can it be done in this timeframe…”, that may seem a subtle difference but it is actually huge because in answering that your developers will take “ownership” of the delivery commitment and that is what you need to stand any chance of being successful. The risk you are taking is on your team, not the project – if your team succeeds then you and your business do, if they fail so do you.

2. Your organization wants to know how much and how long this new software thing is going to take/cost so they can “budget” and control cost escalation.

The reason to budget is because managers and finance people (and he people that own the actual money that gets spent) generally need to *know* how much it costs to feel like they are doing their job, that is just business and really comes from an age where output was quantifiable (manufacturing for example), but creative IP output is in no way quantifiable there are too many variables that are outside of your control so trying to control them is futile.

Think about this, you wake up one day and have a great idea to write a piece of software that will change the world; you believe it is going to make your fortune. You are so confident that you leave your day job, set up your business and your first day you sit down and start to change the world. What is the first thing you are going to do?

I am going to bet that you will NOT crack open Excel and start to do a time estimate and a budget  Instead you will start making your software. So you get on with it, now project forwards, you make your software and you start to sell it and you get successful, in fact so successful that you have to hire a manager or two, then you go and raise some funding to go global with your idea. Now something important happens, instead of spending your time making that thing you believe in, now the people who invested money (which may well be yourself) want to protect it so they put management controls in place, and when the question “To get this thing to the next level, what do we need to do” is asked of you, and you answer “We need to build X, Y and Z” the dreaded question comes – “How long will that take?” which roughly translates to “how much will that cost”, this is because the people asking that question are in fact trying to protect cash and de-risk the investment – they don’t believe in the thing you are building in the same way that you do, the thing is just a means to an end – profit (which by the way is a good thing).

Going back to that first day, if you had tried to budget and determined it was going to take you six months before you could sell your first one, and after six months you realize you were not even half way there and you had another 12 months to go – would you stop? Well you would make sure that you could pay the bills and survive of course, but if you decide to stop, it would not be because of your budget, you would stop because with hindsight you no longer belief the idea was as good as you thought – or you would plough on regardless right?

So back to the boardroom then and the “How long will it take”? question. Well the answer to that question should be, “I have no idea– how important is it”? Because either its important and it had to get done or its not.

You would be a little more committal than that but you get the idea. If you assume that an acceptable level of estimating effort was going to be 25% of the overall development effort (which has been my experience in software development) and if you have an thing that needs to get done because its strategically important for the business to flourish – then how long it takes is irrelevant, its either strategically important and it gets done, or it is not so the distraction of having to do it is removed. Moreover, if it “just has to get done” what on earth are you trying to estimate how long it will take for – just get on with it and use that 25% for making the think you are trying to make – just like you did in the first six months.

You need to ask the same question about how important is it and what is it worth top the business, this is the question that the people trying to de-risk are not wanting you to ask, because they will find that question as difficult to answer as you will trying to answer the “how long will it take” question. Of course for trivial stuff like defects and general maintenance/tactical incremental development work this does not really apply, but for big projects that have strategic importance the “how long will it take?” question is a nonsense question to ask because any answer you get will be either fictitious or grossly over estimated.

If you want to get something strategically important created, hire a great team and empower them to get on with it – if you are making them spend their time telling you how much it will cost to develop instead of developing it then your failing – not them. As a manager, entrepreneur, director or investor, hire software developers to do what they do best – make software, it is your job to take and manage the investment risk, if the team fail then you either hired the wrong team or you did not manage the business well enough to sustain the effort required to make it happen, asking them of an estimate is just a way of getting a stick to blame them when things are not going well.

I have been managing (arguably very loosely) software development projects and a software business for the best part of 20 years, and I have learned a few things along the way. Perhaps more importantly I have been doing this largely investing my own money, so I think I know both sides of the “How long does it take” question very well.

The article I responded to

If you found this post useful or interesting, please consider giving me a tib.
Tibs are pocket-change for the internet™.
Jul 30

Fully Programmable Modular Bench Power Supply – Part 7

Now I have the requirements for the control ranges I need its time to get down to the nitty-gritty and get a DAC up and running so we can make some measurements. The maximum dynamic range I need for this project is 6000 individual steps – this was identified in the calculations for the voltage control for the 0-6V rage in Part 5, so let us start there.

There are many options for DAC’s to choose from, I want to keep the cost and component count down so my starting point is a low-cost single component solution from Microchip, part number MCP4822. This is a dual 12-bit DAC with two channels and a built in voltage reference. I get independent voltage and current control from one 8-pin chip – wow! However, there is a problem with this part, it only has 12-bit resolution which will only give me 4096 individual steps and my design calls for 6000. The problem with choosing components with a higher number of bits is that they start to get expensive. I want to see if it’s possible to extend the range of the DAC using software and a technique called “dithering” or “modulation”.

The idea here is pretty simple, to increase the resolution of the DAC you can continuously switch the output between two or more codes, feed the result into a low-pass filter and get the average voltage. If you switch between two adjacent codes with a variable mark space ratio like you do in PWM, it should be possible to extend the range of the DAC without creating large ripples on the voltage head making the low pass filter easy to construct. That is the theory at least, I need to try it and see what the results are in practice.

Before getting too complicated though, I thought it would be good to run the DAC in static mode, set some codes and measure what we get. The DAC has 4096 steps, the internal reference voltage is 2.048v and the chip has a x2 gain option so I should be able to program any voltage between 0v and 4.096v in 1mV steps by simply programming a digital code between 0 and 4096 into the DAC channel. To get this up and running I hooked up a PIC micro controller, the DAC chip and an RS232 serial interface. The firmware in the PIC will allow me to interact with the DAC through a simple serial terminal on my computer.

Before any MPS430, Atmel or Arduino die-hard fans start giving me advice on micro controller choice — forget it. They are all good parts, I just happen to personally like PIC’s because I know them and I have the tools and a whole bunch of them sitting here to play with – if you are not happy with my choice of micro controller thats tough…. I am not going to change it or enter into any debate over the pro’s and cons of other devices – I am sticking with PIC’s for this one and if you try to change my mind I will ignore you – sorry.

Here is the schematic diagram for the prototype I am using.

For the dithering I have decided to extend the DAC by 2 bits. Extending by two bits means I have to write a sequence of four codes continuously in succession to the DAC. I am writing approximately 1000 codes per second within a timer driven high-priority interrupt routine which ensures that the timing remains constant. Timing errors will introduce more DNL errors so the code stream needs to be constant and accurate. The codes written are the base code value followed by the base code value + 1. The two least significant bits from the now 14-bit word controls how many times each of the two values is written. For example, to get four steps between code 100 and code 101 we would write the following codes: –

100.00       100   100   100   100
100.25       100   100   100   101
100.5        100   100   101   101
100.75       100   101   101   101

Extending by 3 bits is an option and would mean I have to write a sequence of eight codes, again following the same explanation as above, here are the codes that would be written

100.00        100   100   100   100   100   100   100   100
100.125       100   100   100   100   100   100   100   101
100.250       100   100   100   100   100   100   101   101
100.375       100   100   100   100   100   101   101   101
100.500       100   100   100   100   101   101   101   101
100.625       100   100   100   101   101   101   101   101
100.750       100   100   101   101   101   101   101   101
100.875       100   101   101   101   101   101   101   101

I tried the three bits as an academic exercise but I have decided not to go to three bits because of the noise, ripple and integral errors generated. The cost of the filter circuitry and the expansion of the line items in the bill of materials would probably outweigh the cost of upgrading the DAC component to a higher resolution part.

I have selected a number of spot voltages in the range to benchmark what I get from the DAC. The following table sets out the results I measured. (I am using a calibrated HP 34401A meter for all measurements).

Spot VoltageDAC CodeMeasured
(12bit static)
ErrorDAC CodeMeasured
(14bit dith)

Well, that is disappointing given I am aiming for a precision of 1mV and to get a control voltage of 0-6V I need accurate 500µV steps. So whats wrong here? Having read the data sheet there are some gotcha’s that naively you might ignore as I did. Every DAC has two really important parameters called Integral Non-linearity (INL) and Differential Non-Linearity (DNL). The DNL defines the maximum deviation to expect from the “ideal” voltage for any given code, expressed in LSB’s (or counts from ideal) and INL is the accumulated DNL errors that occur over the whole range. Fundamentally, the DAC is based on a resistor string network and its not easy to make highly accurate resistors, as soon as you start needing more accuracy the cost of the part rises very steeply, and even with the best part money can buy there will still be errors. The more bits you extend the DAC by using dithering, the more error you introduce and the more noise you introduce too. While extending by two bits is probably acceptable with a decent low-pass filter, extending by three bits and beyond is not really practical. As an aside, the noise figures for the MCP4x22 parts are not that great – something one must consider when the reference voltage generated is going to be amplified, the noise will also be amplified.

In summary then, I want accuracy and precision but I want reasonable cost and even if I spend a lot of money I will still have errors. The lesson learned for me is I now no longer think of a DAC as an accurate programmable voltage source – its not, its a close approximation only. The MCP4922 (MCP4822) is a nice part for the $$$ and useful for some things I have no doubt, but its not good enough for what I want to achieve in this project. Even with the resolution extension to 14-bits it falls short. Actually to be fair, even with the errors in the DAC this would be make a pretty good degree of control, its probably more accurate than most of the lower end bench PSU’s out there, but my benchmark is the Agilent E3631A so I need to achieve much better than this. The MCP4x22 device is the best resolution DAC Microchip do so I must now search for other parts instead – Linear Technologies and Analog Devices are the logical starting point for my search.

There is one further possibility which I have yet to try, which is to combine both 12-bit DAC outputs to create a much higher resolution DAC, the block diagram for such a solution is shown in the data sheet for the part. This is well worth a look because if it works well enough, the cost of the two chips may well still be cheaper than an upgraded DAC part. I will build this out at some point and give it a try.

This project also needs to implement metering in order to monitor the output volts and current drawn by the load connected to the PSU, and this needs to be reasonably accurate to 1mV too, thats 5-digits I need which in its self is a tall order. However, it occurred to me that if I could get an ADC that was accurate enough and a DAC with enough resolution to provide headroom for trimming it might be possible to build a self-calibrating system that trims the DAC output to match the desired programmed voltage each time you set a new voltage. That is what I will look at in Part 8.

If you found this post useful or interesting, please consider giving me a tib.
Tibs are pocket-change for the internet™.