Software Abstraction - business success stories

Couple of abstractions in the recent memory and their successes

1. Operating Systems abstract the hardware layer and present usability to end users and developers alike. (Microsoft)
2. Compilers - abstracts the machine code in higher level language (IBM, Microsoft, SUN)
3. Virtual Machines abstracts hardware & OS for generic machine codes running over it - (SUN - Java; Microsoft - C#)
4. Virtualization technology (vmware, Citrix - xen; Microsoft - hyperV)

This is likely to grow - as people would like heterogeneous systems interworking.

Apple iTouch

Apple ID seems to have multiple non integrated segments.

1. iTunes account
2. Apple ID

Password requirement on one account need not be the same as other.

If you are in the path to purchase free application, the app would let you sign up with no need for credit card. Otherwise, it requires a valid credit card.

No wounder Microsoft went in for SSO (Single Sign ON) with Passport. Folks are working on Open ID widely accepted by

AOL, BBC, Facebook, Google, IBM, MySpace, Orange, PayPal, VeriSign, LiveJournal, Yandex, Ustream and Yahoo!.

ODBC excel on Windows 7 - 64bit

http://social.answers.microsoft.com/Forums/en-US/w7hardware/thread/001c234b-dfd5-4378-a325-c4f1482fb6fd

Executable:

1. 32 bit version: C:\Windows\System32\odbcad32.exe is what is default (Bad Microsoft)
2. 64 bit version: c:\windows\sysWOW64\odbcad32.exe We need to use

The corresponding .dll file would be

1. 32 bit: c:\windows\sysWOW64\odbc32.dll
2. 64 bit: c:\windows\sysWOW64\odbccp32.dll

Action Driven Testing

open2test.org - requires registration

1. fit
2. Rasta
3. Robot Framework
4. Software Automation Framework Support

	Record-Playback	Data-Method Driven	Action Driven
Time to Market	Low	Medium	High
Learning Curve	Low	High	Medium
Maintenance	High	High	Low
Code Reuse	Low	Medium	High
Tool independence	No	No	Yes
Division of Labor	No	No	Yes
Abstraction Layer	No	No	Yes

Storage Startup Companies

1. autovirt
2. avere
3. caringo
4. digitiliti
5. egnyte
6. exar
7. the linux box
8. nasuni
9. nexenta
10. nimble storage
11. nimbus data
12. nine technology
13. ocarina networks
14. pivot 3
15. promise technology
16. storsimple
17. zetta

Misc

1. Drobo

Pure Storage

Testing opportunities in Agile environment

Agile methodology brings in some opportunities to software testing professional.

Automation is the general demand from project teams so that projects can be shipped as soon as the last code is checked in to source code. Automation of areas that don't change can be a great benefit to testers. Automating not well documented new feature can take much longer than any plan, unless it is done by the developer of the product.

Other changes in testing style that help agile projects

1. Do not call for rigid Testing Procedures as the requirements/features can change rapidly.
2. Test early and often.
3. Test from end user perspective and early
4. Build pass is a test case - don't you think so?
5. Adopt iterative testing - not waiting for the last build to start test cycle
6. Each recorded bug in the system is asset and liability. Liability in that there is fixed cost associated with the administration of the defect. Don't have the team motivation based on bug count - instead focus on quality of the product and the number of not already known issues returning from field
7. Module/unit testing (jUnit) can help - especially when the system level is not quiet ready
8. Practice light weight documentation style. Not a pretty verbose documentation.
9. Make reusable checklists - preferably in wiki
10. Testing Sequence for an incremental internal/external release

• New features
• Bug verification
• regression testing (run any automated test continuously)

Reference: Wikipedia article http://en.wikipedia.org/wiki/Agile_testing

Sparse Files

A sparse file attempts to use file system space more efficiently when blocks allocated to the file are mostly empty. This is achieved by writing brief information (metadata) representing the empty blocks to disk instead of the actual "empty" space which makes up the block, using less disk space. The full block size is written to disk as the actual size only when the block contains "real" (non-empty) data.