Single point of failure

If you're going to have a single point of failure, make it replaceable.

We strive to avoid single points of failure. They hold risk -- if a single point of failure fails, then the entire system fails.

It is not always possible to avoid a single point of failure. Sometimes the constraint is cost. Other times the design requires a single component for a function.

If you have a single point of failure, make it easy to replace. Design the component so that you can replace it quickly and with little risk. When it fails, you can respond and install the replacement component. (Kind of like a spare tire on an automobile. Although the four tires on a car are not a single point of failure, because there are four of them. But you get the idea.)

A simple design for a single point of failure (or any component) requires care and attention. You have to design the component with minimal functionality. Move what you can to other, redundant components.

You also have to guard against changes to the simplicity. Over time, designs change. People add to designs. They want new features, or extensions to existing features. Watch for changes that complicate the single point of failure. Add them to other, redundant components in the system.

Do you want to be time and on budget, or do you want a better product?

Project management in IT is full of options, opinions, and arguments. Yet one thing that just about everyone agrees on is this: a successful development project must have a clear vision of the product (the software) and everyone the team has to understand that vision.

I'm not sure that I agree with that idea. But my explanation will be a bit lengthy.

I'll start with a summary of one of my projects: a BASIC interpreter.

* * * * *

It started with the development of an interpreter. My goal was not to build a BASIC interpreter, but to learn the Ruby programming language. I had built some small programs in Ruby, and I needed a larger, more ambitious project to learn the language in depth. (I learn by doing. Actually, I learn by making mistakes, and then fixing the mistakes. So an ambitious project was an opportunity to make mistakes.)

My initial clear vision of the product was just that: clear. I was to build a working interpreter for the BASIC language, implementing BASIC as described in a 1965 text by Kemeny and Kurtz (the authors of BASIC). That version had numeric variables but not text (string) variables. The lack of string variables simplified several aspects of the project, from parsing to execution. But the project was not trivial; there were some interesting aspects of a numeric-only BASIC language, including matrix operations and output formatting.

After some effort (and lots of mistakes), I had a working interpreter. It really ran BASIC! I could enter the programs from the "BASIC Programming" text, run them, and see the results!

The choice of Kemeny and Kurtz' "BASIC Programming" was fortuitous. It contains a series of programs, starting with simple ones and working up to complex programs, and it shows the output of each. I could build a very simple interpreter to run the initial programs, and then expand it gradually as I worked my way through the text. At each step I could check my work against the provided output.

Then things became interesting. After I had the interpreter working, I forked the source code and created a second interpreter that included string variables. A second interpreter was not part of my initial vision, and some might consider this change "scope creep". It is a valid criticism, because I was expanding the scope of the product.

Yet I felt that the expansion of features, the processing of string variables, was worth the effort. In my mind, there may be someone who wants a BASIC interpreter. (Goodness knows why, but perhaps they do.) If so, they most likely want a version that can handle string variables.

My reasoning wasn't "the product needs this feature to be successful"; it was "users of the product will find this feature helpful". I was making the lives of (possibly imaginary) users easier.

I had to find a different reference for my tests. "BASIC Programming" said nothing about string variables. So off I went, looking for old texts on BASIC. And I found them! I found three useful texts: Coan's "Basic BASIC", Tracton's "57 Practical Programs and Games", and David Ahl's "101 BASIC Computer Games".

And I was successful at adding string variables to the interpreter.

Things had become interesting (from a project management perspective) with the scope expansion for an interpreter that had string variables. And things stayed interesting: I kept expanding the scope. As I worked on a feature, I thought about new, different features. As I did, I noted them and kept working on the current feature. When I finished one feature, I started another.

I added statements to process arrays of data. BASIC can process individual variables (scalars) and matrices. I extended the definition of BASIC and created new statements to process arrays. (BASICs that handle matrices often process arrays as degenerate forms of matrices. The internal structure of the interpreter made it easy to add statements specific to arrays.)

I added statements to read and write files. This of course required statements to open and close files. These were a little challenging to create, but not that hard. Most of the work had already been done with the input-output processing for the console.

I added a trace option, to see each line as it executed. I found it useful for debugging. Using my logic from the expansion for string variables, if I found it useful then other users would (possibly) find it useful. And adding it was a simple operation: the interpreter was already processing each line, and all I had to do was add some logic to display the line as it was interpreted.

I added a profiler, to count and time the execution of each line of code. This helped me reduce the run-time of programs, by identifying inefficient areas of the code. This was also easy to add, as the interpreter was processing each line. I simply added a counter to each line's internal data, and incremented the counted when the line was executed.

Then I added a cross-reference command, which lists variables, functions, and constants, and the lines in which they appear. I use this to identify errors. For example, a variable that appears in one line (and only one line) is probably an error. It is either initialized without being used, or used without being initialized.

I decided to add a debugger. A debugger is exactly like trace mode, with the option to enter a command after each statement. This feature, too, helps the typical user.

* * * * *

Stepping back from the project, we can see that the end result (two interpreters each with profiler, cross-reference, trace mode, and debugger) is quite far from the initial vision of a simple interpreter for BASIC.

According to the predominant thinking in project management, my project is a failure. It delivered a product with many more features than initially planned, and it consumed more time than planned, two sins of project management.

Yet for me, the project is a success. First, I learned quite a bit about the Ruby programming language. Second -- and perhaps more important -- the product is much more capable and serves the user better.

* * * * *

This experience shows a difference in project management. As a side project, one without a firm budget or deadline, it was successful. The final product is much more capable that the initial vision. But more importantly, my motivation was to provide a better experience for the user.

That result is not desired in corporate software. Oh, I'm sure that corporate managers will quickly claim that they deliver a better experience to their customers. But they will do it only when their first priority has been met: profits for the corporation. And for those, the project must fit within expense limits and time limits. Thus, a successful corporate project delivers the initial vision on time and on budget -- not an expanded version that is late and over budget.

The cult of fastest

In IT, we (well, some of us) are obsessed with speed. The speed-cravers seek the fastest hardware, the fastest software, and the fastest network connections. They have been with us since the days of the IBM PC AT, which ran at 6MHz which was faster than the IBM PC (and XT) speed of 4.77MHz.

Now we see speed competition among browsers. First Firefox claims their browser is fastest. Then Google releases a new version of Chrome, and claims that it is the fastest. At some point, Microsoft will claim that their Edge browser is the fastest.

It is one thing to improve performance. When faced with a long-running job, we want the computer to be faster. That makes sense; we get results quicker and we can take actions faster. Sometimes it is reasonable to go to great lengths to improve performance.

I once had a job that compared source files for duplicate code. With 10,000 source files, and the need to compare each file against each other file, there were 1,000,000 comparisons. Each comparison took about a minute, so the total job was projected to run for 1,000,000 minutes -- or about 2 years! I revised the job significantly, using a simpler (and faster) comparison to identify if two files had any common lines of code and then using the more detailed (and longer) comparison on only those pairs with over 1,000 lines of common code.

Looking for faster processing in that case made sense.

But it is another thing to look for faster processing by itself.

Consider a word processor. Microsoft Word has been around for decades. (It actually started its life in MS-DOS.) Word was designed for systems with much smaller memory and much slower processors, and it still has some of that design. The code for Word is efficient. It spends most of its time not in processing words but in waiting for the user to type a key or click the mouse. Making the code twice as fast would not improve its performance (much), because the slowness comes from the user.

E-mail is another example. Most of the time for e-mail is, like Word, the computer waiting for the user to type something. When an e-mail is sent, the e-mail is passed from one e-mail server to another until it arrives at the assigned destination. Changing the servers would let the e-mail arrive quicker, but it doesn't help with the composition. The acts of writing and reading the e-mail are based on the human brain and physiology; faster processors won't help.

The pursuit of faster processing without definite benefits is, ironically, a waste of time.

Instead of blindly seeking faster hardware and software, we should think about what we want. We should identify the performance improvements that will benefit us. (For managers, this means lower cost or less time to obtain business results.)

Once we insist on benefits for improved performance, we find a new concept: the idea of "fast enough". When an improvement lets us meet a goal (a goal more specific than "go faster"), we can justify the effort or expense for faster performance. But once we meet that goal, we stop.

This is a useful tool. It allows us to eliminate effort and focus on changes that will help us. If we decide that our internet service is fast enough, then we can look at other things such as database and compilers. If we decide that our systems are fast enough, then we can look at security.

Which is not to say that we should simply declare our systems "fast enough" and ignore them. The decision should be well-considered, especially in the light of our competitors and their capabilities. The conditions that let us rate our systems as "fast enough" today may not hold in the future, so a periodic review is prudent.

We shouldn't ignore opportunities to improve performance. But we shouldn't spend all of our effort for them and avoid other things. We shouldn't pick a solution because it is the fastest. A solution that is "fast enough" is, at the end of the day, fast enough.