Programming skills self-assessment matrix

Writing

B1

B2

C1

C2

Intermediate User

Intermediate User

Proficient User

Proficient User

I can determine a suitable interface and

produce a correct implementation, given a
loose specification for a simple function,
without help from others. I can break down a
complex function specification in smaller
functions.

I can estimate the space and time costs of

my code during execution. I can empirically
compare different implementations of the
same function specification using welldefined metrics, including execution time and
memory footprint. I express invariants in my
code. I use stubs to gain flexibility on
implementation order.

I use typing and interfaces deliberately and

productively to structure and plan ahead my
coding activity. I can design and implement
entire programs myself given well-defined
specifications on external input and output. I
systematically attempt to generalize functions
to increase their reusability.

I can systematically recognize inconsistent or

conflicting requirements in specifications. I
can break down a complex program
architecture in smaller components that can
be implemented separately, including by
other people. I can us existing (E)DSLs or
metaprogramming patterns to increase my
productivity.

I can reliably recognize when underspecification is intentional or not. I can exploit

under-specification to increase my
productivity in non-trivial ways. I can devise
new (E)DSLs or create new
metaprogramming patterns to increase my
productivity and that of other programmers.

I can adapt my code when I receive small

changes in its specification without rewriting it
entirely, provided I know the change is
incremental. I can change my own code
given detailed instructions from a more
experienced programmer.

I can determine myself whether a small

change in specification is incremental or
requires a large refactoring. I can change my
own code given loose instructions from a
more experienced programmer.

I can derive a refactoring strategy on my own

code, given relatively small changes in
specifications. I can change other people's
code given precise instructions from a person
already familiar with the code.

I can estimate accurately the effort needed to

adapt my own code base to a new
specification. I can follow an existing
refactoring strategy on someone else's code.
I can take full responsibility for the integration
of someone else's patch onto my own code.

I can reverse-engineer someone else's code

base with help from the original specification,
and estimate accurately the effort needed to
adapt it to a new specification.

I can reverse-engineer someone else's code

base without original specification, and
estimate accurately the effort needed to
adapt it to a new specification.

I know the entry and termination points in the

code I write. I can use the main I/O channels
of my language to input and print simple text
and numbers.

I am familiar with recommended mechanisms

to accept program options/parameters from
the execition environment and signal errors,
and use them in the code I write.

I can delegate functions to an external

process at run-time. I know how to
productively use streaming and buffering to
work on large data sets and use them in my
code. I am familiar with the notion of locality
and use it to tailor my implementations.

Given a library of mostly pure functions and

and detailed API documentation, I can reuse
this library productively in my code.

I can recognize when existing code requires

a particular overall architecture for reuse (eg.
an event loop). I can adapt my own code in
advance to the requirements of multiple
separate libraries that I plan to reuse.

I can implement both client and server

software for arbitrary protocol specifications. I
can quantify accurately the time and space
overheads of different communication
mechanisms (eg, syscalls, pipes, sockets). I
am familiar with hardware architectures and
can predict how sequential programs will
behave when changing the underlying
hardware. I can estimate the scalability of
parallel code fragments on a given platform.
I can systematically remove constraints from
existing code that are not mandated by
specification, to maximize its generality. I can
read and understand code that uses APIs
most common in my domain without help
from their documentation. I can interface
code from different programming languages
with distinct operational semantics.

I am familiar with most software architectures

in use with systems I develop for. I can work
together with system architects to mutually
optimize my own software architecture with
the overall system architecture. I am familiar
with most design and operational cost/benefit
trade-offs in systems that I develop for.

I can assemble program fragments by

renaming variables until the whole becomes
coherent and compatible with my goal.

I am familiar with at least one API for bidirectional communication with other run-time
processes. I can write client code for simple
Internet protocols. I am familiar with the most
common packaging and redistribution
requirements of at least one platform and use
them in my own projects. I can use
predetermined programming patterns to
exploit platform parallelism productively in my
code.
I can recognize and extract reusable
components from a larger code base for my
own use, even when the original author did
not envision reusability. I can package,
document and distribute a software library for
others to reuse. I can interface stateless
code from different programming languages.

I can read the code I wrote and explain what I I can read code from someone of a similar or
intend it to mean to someone more
lower level than me and explain what it
experienced than me.
means. I can recognize and explain simple
Explaining /
mismatches between specification and
Discussing code
implementation in my code or code from
someone at the same level as me or lower.

I can show and explain code fragments I

write in either imperative or declarative style
to someone else who knows a different
programming language where the same
style is prevalent, so that this person can
reproduce the same functionality in their
language of choice.

I can explain my data structures, algorithms

and architecture patterns to someone else
using the standard terms in my domain,
without reference to my code.

I can gauge the expertise level of my

audience and change the way I talk to them
accordingly. I can recognize when an
explanation is overly or insufficently detailed
for a given audience, and give feedback
accordingly.

I can take part effortlessly in any

conversation or discussion about the
language(s) I use, and have a good familiarity
with idiomatic constructs. I can come up
spontaneously with correct and
demonstrative code examples for all
concepts I need to share with others.

Refactoring

Embedding in a
larger system

Understanding

A2
Basic User

I can produce a correct implementation for a

simple function, given a well-defined
specification of desired behavior and
interface, without help from others.

Writing code

Reusing code

Exploring, selflearning

Interacting

A1
Basic User

Mastery of the
environment

Troubleshooting

I can discover and reliably exploit

undocumented/unintended behavior of any
code written in a language I understand,
including code that I did not write myself.

I can distinguish between a command prompt

at a shell and an input prompt for a program
run from this shell. I can follow online tutorials
without help and reach the prescribed
outcome. I can search for the text of common
error messages and adapt the experience of
other people to my need.

I can distinguish between features general to

a language and features specific to a
particular language implementation. I can
read the text of error messages and
understand what they mean without external
help.

I can read the reference documentation for

the language(s) or API I use, and refer to it
to clarify my understanding of arbitrary code
fragments. I can understand the general
concepts in articles or presentations by
experts. I can track and determine who is
responsible for an arbitrary code fragment in
a system I use or develop for.

I can infer the abstract operating model of an

API or library from its interface, without
access to documentation, and write small test
programs to test if my model is accurate. I
can recognize when a reference
documentation for a language or API is
incomplete or contradictory with a reference
implementation.

I am able to read and understand most

expert literature applicable to the languages I
use. I am able to recognize when an
academic innovation is applicable to my
domain and adapt it for use in my projects.

I can recognize and expose tacit

assumptions in expert literature in my
domain. I can reliably recognize when the
narrative or description of a programming
achievement is false or misleading, without
testing explicitly.

I can use a common programming

environment and follow common workflows
step-by-step to test/run a program.

I can integrate my source files in a

programming environment that automates
large portions of my programming workflow. I
use version control to track my progress and
roll back from unsuccessful changes.

I express and use dependency tracking in my

programming environment to avoid
unnecessary (re)processing in my
development cycles. I can use different
development branches in version control for
different programming tasks.

I use different workflows for different

programming assignments, with different
trade-offs between initial set-up overhead
and long-term maintenance overhead. I can
enter the environment of someone else at my
level or below and make code contributions
there with minimal training.

I modify my programming environment to

tailor it to my personal style, and can quantify
how these changes impact my productivity. I
can productively use the preferred
programming environments of at least 80% of
all programmers at my level or below.

I can reliably recognize and quantify friction

between other programmers and their
programming environment. I can measurably
improve the productivity of my peers by
helping them tailor their environment to their
personal style.

I can distinguish between correct and

incorrect output in my own programs. I am
familiar with the etiquette for asking help from
experts in my domain.

I can reliably distinguish between incorrect

output due to incorrect input, from incorrect
output due to program error. I can narrow
down the location of a program error in a
complex program to a single module or
function. I can isolate and fix Bohr bugs in my
own code.

I can translate human knowledge or

specifications about invariants into assertions
or type constraints in my own code. I can
inspect the run-time state of a program to
check it matches known invariant. I write and
use unit tests where applicable.

Ican reduce a program error to the simplest

program that demonstrates the same error. I
have one or more working strategy to track
and fix heisenbugs in code that I can
understand. I write and use regression tests
for code that I work with directly.

I can devise systematic strategies to track

and fix mandelbugs in code that I can
understand. I can recognize a hardware bug
in a system driven mostly by software I
designed.

I can track and attribute responsibility

accurately for most unexpected/undesired
behaviors in systems that I develop for. I can
track and isolate hardware bugs in systems
where I have access to all software sources.

Copyright 2014 Raphael kena Poss. Permission is granted to distribute, reuse and modify this table according to the terms of the Creative Commons-ShareAlike 4.0 International License.
Explanatory notes are available online at:t http://science.raphael.poss.name/programming-levels.html