IT Werkz Sometimes

The Epson DX8450 is still painfull in that it doesn’t feed paper too well and the HP PSC2355 has completely given up the ghost, well it’s got a permanent carriage jam problem.

The cost of print catridges is so much I’m going to look at getting a laser printer that has very expensive cartridges but that do last for years with my kind of usage. If I get a laser without the scanning and copying functionality they are relatively cheap but that does mean that I will need to keep the old Epson connected so that it can do that stuff. Now which laser do I get?

On Linkedin I belong to a load of testing groups and I get email whenever someone posts a question. So many of these questions are banal beyond belief and some are just downright begging letters for others to brain dump testing knowledge on some hapless group members. I know it’s supposed to be a forum for people to share knowledge but jeez it’s so one way.

I’m starting to look into getting a new PC with Windows 7. I have installed 7 Ultimate on an old laptop but it runs like a dog and hasn’t got the hardware necessary to do the XP virtualisation thing. I might get some more memory for it. My old XP PC also runs out of steam soon enough after starting anything and the power supply fan starts up in a vain attempt at letting the processor do something. That PC also needs more memory and a bigger power supply. The Vista PC also needs more memory, currently with only 2 gig. When I start VM’s soon enough that PC’s power supplies fan starts up, so that PC also needs more memory. This could get expensive.

(More of this stuff? Wake me up when it’s over. Ed)

Experienced-based testing
- Strengths
— Works in short timescales
— Enables reactive and flexible testing (tester learns from results of previous tests)
— Often applied where there is minimal documentation
- Weaknesses
— Poorly documented
— Relies on experience of tester
— Unpredictable application by different testers
- Examples
— Exploratory testing
— Error guessing

Scripted vs unscripted testing
- Exploratory testing, while often considered an unscripted approach, is a systematic technique, while ad-hoc testing is the testing equivalent of ‘hacking’.

Non-functional testing
- Strengths
— Evaluates quality characteristics
- Weaknesses
— Often overlooked
— Poorly understood
— Often needs specialists
- Examples
— Performance testing
— Usability testing
— Reliability testing

Specifying Software Reliability
- Probability of Failure on Demand (POFOD)
— the likelihood that the system fails on demand
— POFOD = 0.01 (or 1% means that at least 99 out of every 100 requests must be dealt with successfully
— The Sizewell B Primary Protection System has a requirement of POFOD = 0.001
- Mean Time to Failure (MTTF)
— the time between system failures
— MTTF = 10 to the power 3 hours means that the average time between failures is no less than 1000 hours
— the Flight Control Software for the Airbus A320 and Boeing 777 has a requirement of MTTF of 10 to the power 9 hours

TPI
- Based on:
— Experience
— TMap
— Other process improvement models
— Boehm’s 1979 rework stats
- “..more improvement steps, practical details and instructions”
- Continuous model
— equivalent to a 13 stage model
- Includes optional improvement suggestions
- Survey results (113 responses)
— 81% felt TPI had a positive effect
— 66% felt TPI resulted in a product with less failures

TPI Structure
Key Areas -> Levels -> Checkpoints or Improvement Suggestions

TPI Assessment Activities
Example TPI checkpoint:
- Level A
— A test strategy is defined for a single high-level test
- Checkpoints – Level A
— A motivated product risk analysis
— There is a variation in test depth, depending on risks
— One or more test design techniques are used, in line with required depth of test
— A re-test strategy is defined

Example TPI improvement suggestions:
- Level A
— A test strategy is defined for a single high-level test
- Improvement suggestions Level A
— Involve proper parties when defining test strategy
— If only one test design technique available, make variations to vary depth of rigour
— Define a re-test strategy, consider full re-test, a thin re-test, or even no re-test
— Identify subsystems and quality characteristics, with defined relative importance

‘Ideal’ Test Process Improvement
- Easy to use
- Continuing publicity
- Publicly available
- Available consultants
- Not marketing
- Accepted by professional bodies
- Include provision for improvement
- Provide many small evolutionary improvements
- Sound basis:
— Practical
— Empirical
— Theoretical
— Published and justified
- Detail of how to:
— Assess
— Identify improvements
— Make improvements
- Quantifiable improvements
- Tailorable (project specific)

Comparison of TPI and TMMI:
TMMi will look at the whole organisation, strictly staged improvement jumps levels 1-5
TPI allows improvements of parts of the organisation, small continuous improvement moving A-B-C-D with a quality profile.