Philosophy for SW Engineers

• Bear in mind that Absence of Evidence of SW product malfunction does not always mean Evidence of Absence:

 

•   The world as it really is   >   the world as it appears to us (to our senses).  Human Eye senses much more the visual cortex perceives.

 

 

• As an engineer or an manager in a hi-tech company you should be aware that between God and CEO there is an intermediate level – Board of Directors.

 

• The Latin phrase memento mori means “remember that you must die”. For SW engineers this phrase means – all data you allocate must die (i.e. to be de-allocated).
  

 

• There is no flawless SW engineer, therefore there is no flawless SW product.

 

• Improbable is not the same as impossible (‘improbably’ means highly unlikely). In  other words, if something can horribly go wrong, it will; it is just a matter of time.
  

 

• Frustration can be a great teacher

 

• As computer program grows it becomes increasingly susceptible  to crashing and misbehaving, because conflicting instructions buried in miles of code. More staff is required to maintenance such growing monster.

 

• You need use both methods: Deduction and Induction, where Deduction works from a hypothesis to a conclusion while Induction works in the opposite direction: from evidence to hypothesis.

 

• Don’t pursue high job titles. Job titles are irrelevant. What’s relevant – your knowledge, experience, innovativity and skills.

 

 

• 

 

• Bear in mind success comes after long hard work.

 

• Bear in mind the correlation is not always transitive relation.Firstly, what’s the transitivity:  ‘a’ is greater than ‘b’ and ‘b’ is greater than ‘c’, hence  ‘a’ is greater than ‘c’.For counter-example, let’s consider the following situation:  I prefer ice-cream ‘a’ to ice-cream ‘b’. I also prefer ice-cream ‘b’ to ice-cream ‘c’. Does this mean that I prefer ice-cream ‘a’ to ice-cream ‘c’?

 

• Not all uncorrelated variables are unrelated.

 

• To develop technology, you need to understand how things work.

 

• Bear in mind the Pareto Principle (or the 80/20 rule) which states that for many phenomena 80% of the result comes from 20% of the effort.  Generally speaking, in each team 20% makes 80% of total work.
  

 

• Bear in mind: Low probability events are not zero probability events

 

• Learn how to learn and learn how to learn to learn …

 

• Three pillars of system observability: logs, metrics, and traces. Logs – timestamped record of relevant events while the system is running. Metrics is utilization data measured over intervals of time, e.g. CPU/GPU usage, GPU clock value. Traces are series of causally related events, e.g. in Server-Client system the trace can be: send a request, process request, send reply, receive the reply, process the reply.

 

• Law of Sufficiency: if a solution meets specifications and it is fast enough then this solution is sufficient.

 

• Gödel’s second incompleteness theorem states no consistent axiomatic system can prove its own consistency. In other words no system can be safe from hacking.

 

• Statistical inference is the act of generalization from a sample to a population with some degree of certainty.

 

• How to keep up with a rapidly advancing field – read technical literature and participate in technical blogs, ask directly experts.
  

 

• Taken from LinkedIn:

, an

 

• In natural sciences, an experiment and/or observation is the scientist’s way of asking Nature a question.

In Client-Server paradigm, Nature is Server, the scientist is Client and the experiment/observation is a query. You need arrange the query (the experiment) correctly to get a correct response from Server and not a garbage. Server can be stuck and you need wait for a response.

 

• There are many theories on how consciousness came to be, but no concrete proofs.
  There exist so called theory of emergent consciousness: create enough neural connections out of ordinary matter (e.g. organic molecules) and at some point consciousness emerges naturally.
  Alternatively, create sufficiently rich in nodes and connections a neural network and at some point consciousness emerges, then you press “the red button” and kill the conscious being.

 

• There is a wide gap between intention and reality

 

• Sometimes laziness is beneficial. Let’s suppose you are asked to supply 89 pencils from the box of 100 pencils. Instead of counting 89 pencils, it’s better to pick 11 pencils and put them aside. Then the box will contain 89 pencils and you merely supply these 89 pencils. So you counted 11 pencils instead of 89 ones.

 

• 

 

• Random sequence definition from Computer Science point of view: random sequence can’t be recorded by means of a program (an algorithm) which is shorter than the sequence itself.

 

• If you have to take a decision under uncertainty (most of cases), bear in mind that mistakes cannot be avoided.

 

• One of main questions of Data Analysis:
  Can we be confident in the patterns that we are seeing in the data or what we are observing is just noise?

 

 

• 

The best ideas often come from those who are willing to listen, rather than those who are eager to speak.

taken from LinkedIn, published by Tyler Smith

 

• Most Meetings Are Useless. Some meetings are merely a speech of one or two persons, escape them.
  

 

• from the book “Machine Learning God”, by Stefan Stavrev
  
  

 

• Boundaries exist only in the human brain (or mind).

 

• Never apply the following rule: if you don’t know to explain something (e.g. a bug in a program), pretend it does not exist.

 

• Generally speaking, learning is the process of converting experience into knowledge.

 

• Equations that changed our understanding of the Nature
• 

 

• Difficult does not necessarily mean impossible