Coming home: a return to PHP’s roots by Adam Harvey
- I arrived late for this but he looked like he was having fun
- He wrote a PHP micro-framework live in the talk
- Cute
Get your PaaS into gear by Katie Miller
- An introduction to OpenShift
- Platform As A Service, include web servers, databases, storage.
- Iaas -> Paas -> Saas
- Why?
- Focus on code, not config
- Speed of deployment
- Convenience
- Scalability
- effeciency
- Flavours
- Origin – Source can be downloaded
- online – Hosted in Amazon
- Enterprise – Run you own supported
- Runs on RHEL
- Runs in containers
- Cartridges (programming environment) within the container
- Tissues!
The Cobblers Children Have No Shoes: Development Tools and the Unix Philosophy by Russell Keith-Magee
- Django Guy
- Over last 30 years – My tools have got appreciable worse
- Libraries, languages and computers better
- Programming tools have not
- Some have got worse
- Not language specific
- Debugging
- Borland C – GUI – see where in code you are
- 1994 gdb
- 2013 gdb
- Why?
- The Unix philosphy – small, seperate tools
- Cargo Cult
- Powerful tools only need a CLI
- Fully featured IDE?
- Only one tool the GUI
- Maybe a plugin interface, if you are lucky
- Keep good bit of Unix philosphy
- Different Window dressing
- BeeWare
- Common problem with bad interface – testing
- Current test suite outputs little to screen until it finishes.
- Cricket – runs test suites and lets you look at output while it is running, see progress, what is passing/failing, can browse errors, does just one thing
- cross platform – uses what Python provides out of the box, uses tk
- Zero configuration – uses pip and run from command line. Nothing set in tool
- Integration
- Duvet – Test coverage tool
- Why not a web interface
- Needs to have a server
- Live updates are still ard
- Don’t believe web browsers will be superior to native tools and time soon
- pybee.org
A Pseudo-Random Talk on Entropy by Jim Cheetham
- Random vs Pseudo-Random vs Cryptographically Secure Pseudo-Random
- Some Pseudo-Random generators not as random as was expected
- Good PRNG – Mersenne Twister,
- Blum Blum Shub – Crypto good if you pick right numbers
- Diehard test spot bad random numbers
- Random numbers for science – use publish “table” of pre-generated ones
- /dev/random very random enviromental noise , /dev/urandom not crypto-good
- /dev/random should block if not enough random bits available
- RdRand – Source of random data built into chip, measures physical noise, runs through AES, can’t acess original data though
- Sources from various hardware generators
- Turbit – reads noise from soundcards
- OneRNG – Speakers project, Open etc