Years ago I had the fortunate and frustrating experience of working with Charlie, a hardware and software developer of much experience.

Charlie told me of a guideline he used which I have found to be quite valuable over the years, and accurate too. I have taken to calling it Charlie's Law.

Let's say you're writing code that makes some work function much easier, or you develop some gizmo that's really cool and you want to spread it to your group or create a marketable product. It only took a day, a week, maybe a month to get it working pretty well for yourself. Call this one level of effort.

To put a little polish on it, get away from quirky steps that are difficult to explain (but you understand completely) to show it to a few coworkers will take about three times the level of effort. If it took one day to make something you were happy with, it will take three additional days to move it up to level two.

Showing a few close coworkers your problem went well or your gizmo was well received by your close friends, so now you want to present it too a bigger audience – your whole work group or people to review a potential product. Now the program has to shine. No unexpected problems. A good looking product in final from. Instructions and documentation. Charlie says three more times the effort of level 2 to get to level 3; nine units of additional effort to reach the next level, and 12 units more than it took to start with.

I'm going to extrapolate from what Charlie explained to me. The product level, ready to ship. Polished to a high degree. No errors. If it's hardware, plans for manufacturing in small to large quantities, packaging, user's manuals, the list of details is long. A million details to settle. Charlie's factor of 3 is a starting point, but a bare minimum in my experience. So to go up to this level, it's at least 27 units of additional effort from level 3.

This just covers "the thing" and not the business requirements supporting "the thing."

I've found Charlie's Law mirrors reality pretty will. I'm curious if others have similar experiences.

 

Add “compliance with national and international standards”.


Also, It perfectly explains the “mechanic’s car” syndrome!

 

In some companies, the path from working concept to getting it into a customer's hands can be gruelling years-long project, particularly if it involves software and firmware.

 

Similar to the 90 / 10 rule.

 

Years ago...

So, what's the advice?

 

So, what's the advice?

Don’t get too excited when you demonstrate a great prototype, unless you want to spend three years on the project. (been there, done that)

 

Don’t get too excited when you demonstrate a great prototype, unless you want to spend three years on the project. (been there, done that)

What's the point of a prototype, then?

 

So, what's the advice?

While I don't know that I agree with the specific scaling factors mentioned, my general takeaway is that we need to be realistic about the expectations of the time and effort to bring a product to market (whatever 'product' and 'market' happen to mean in a particular context). I think most people (myself included) tend to naturally think in near-linear terms about these things while they can be near-exponential in actuality.

 

While I don't know that I agree with the specific scaling factors mentioned, my general takeaway is that we need to be realistic about the expectations of the time and effort to bring a product to market (whatever 'product' and 'market' happen to mean in a particular context). I think most people (myself included) tend to naturally think in near-linear terms about these things while they can be near-exponential in actuality.

A properly laid out Gant chart quickly dispells any optimism about quick TTM.

Many decades ago, I presented such a chart to my manager. He was displeased with my proposed product launch date.

I explained my methodology in generating the schedule, and that the launch date was reasonable based on the work that needed to be done, the resources available, and the finite number of hours in any given day.

His response: "push the launch date up six months. It'll make everyone work faster."

 

His response: "push the launch date up six months. It'll make everyone work faster."

... and then they let someone else to do it because that person did not object to shorter deadline.
unsurprisingly things don't go as boss wished. then he points finger at you...
first rule of management - do not admit you are wrong, find someone else to blame.

 

... and then they let someone else to do it because that person did not object to shorter deadline.
unsurprisingly things don't go as boss wished. then he points finger at you...
first rule of management - do not admit you are wrong, find someone else to blame.

Oh, I just laughed at him and told him he can have whatever launch date he wants, but I'm keeping mine.

 

What's the point of a prototype, then?

Proof of concept to get someone else to do it.

Done that too.

 

Level 5: Design it so it lasts 3 decades 24/7 in a harsh (chemicals, radiation, high EM fields, late night operators) industrial environment.

 

I was hassled on a project for the FAA because I insisted on using 7-segment LEDs instead of an LCD to display two 3-digit numbers. Some FAA equipment from the 70s is still in use.

I want my project to outlive me, by many years. The LEDs may get dim in 15–25 years depending on usage, but I know even that far into the future and well beyond, 7-segment displays will almost certainly be available, and replaced without too much trouble by somebody who can solder.

How long will LCD displays last? There's no good documentation to make a prediction. But from current experience, I know finding an equivalent display (form, fit and interface) in even a couple years is all but impossible.