mind¶

Technical Capabilities¶

Measure Lines Spent Edgar Djyskstra quote - tech debt The more code written the more there is to understand Writing lots of code, writing it quickly, is often the contractor curse - too much too soon. There is no such thing as hitting the ground running.
measure Onboarding times
measure code quality not feature velocity
measure effort to deploy
measure inter dependancy
measure metrics - is code well

Measured in production

web logging good enough practise

generate a unique ID per request, pass it through all stages

amatix 2 hours ago [-]

For us, a X-Request-ID header is generated by any app if it doesn’t receive it from upstream – but normally nginx or the CDN will generate it. There’s a few nginx modules to do it, we use https://github.com/newobj/nginx-x-rid-header Most languages/logging frameworks have some sort of per-thread context (eg. Filters in Python, MDC in log4j, etc) to be able to tag log messages with. If you’re using postgresql, you can call SET application_name=’{requestID}’; and that can be output as part of logs too. reply

Deployment can deploy code through promotion of fixed “builds” into production using totally automated means.
micro services architecture This is a major issue. Includes security, service discovery messaging logging and pr nose to only go through front door. See stevey blog post. https://news.ycombinator.com/item?id=12133670 Use fluentd logging

https://blog.treasuredata.com/blog/2016/08/03/distributed-logging-architecture-in-the-container-era/

OS installation and Hardening Choice of OS, location Ansible or similar automation scripts
Kerberos and Authentication Authentication and Authorisation are vital capabilities and federated, distributed security have so many advantages (think uUId as primary key )
Build promotion and use of containers PyHolodeck
message queues and worker pools Messaging Job scheduling based on events and then queues Is a resource problem always
performance monitoring metrics gathering with graphite
unit testing, integration testing, perfromance testing
user acceptanbce tests and UI tests Automated testing is vital and we must put huge amounts of effort into it simply to make the next release
continuous integrations and delivery There is no point batching up changes to go on a time based schedule. Code is ready as it is written and the decision to release should be based on automated tests. Only marketing or other business decisions should delay - so always push to pre-prod
pre-prod is live-clone
replaying actual traffic
Reporting Simple reports
SOPs and checklists
Measuring metrics
documentation
Dev Ops
statistics
distributed data and computation Use of Apache spark, Hadoop etc
Javascript
Angular vs Jquery
Programmable Corporation
functional programming and scalability Elixr and OTP
bug tracking, code analysis,
git and git workflow
Linux and workstations
Mobile development
Big Data
Maturity and keeping on a single course
seven league boots
earthquake-let theory of software development. Building blocks not planned features
Work harder on fewer projects and polish them to be CV-worthy shipping working software
Message Queue
Business Intelligence reporting
Software KPIs, buisness KPIs
Simple Marketing (Twilio approach), SEO

https://www.ducttapemarketing.com/blog/guide-to-small-business-keyword-research/

Debugging tools CHarles proxy Chrome debugger
Web technologies (html 5) Bootstrap, HTML 5, cimples CSS
Rekational Databases
NOSQL databases
Storage systems
Backup and recovery Business Disaster recovery and assurance Multiple data centres
Time, Unicode, basic data types
serialisation pickle, json etc
Cacheing
Configuration basics
command line is all
Publishing own work
contact management
email management
Practise on smaller complete projects
jenkins
chatbots and IRC. Team comms. Choose one one o said one!!
erlang
rethinkdb
training and just reading and exploring

if you do not explicitly make time to learn but sprint all the time all you will have is people who have learnt what they wrote