2007/10/20

MinWin: 25Mb WIndows. Hypervisor expected?

Could this be the start of a real change at MSFT? [i.e. doing software 'properly' (small, fast, secure)]

First question:
  • What if they pick up GPL or similar code & illegally include it.
  • How would that be detected??

Timeframe for 'commercial' MinWin is 2010.
The real news here is MSFT's focus on virtualisation...
With their purchase of "Virtual PC", they have the tools to build their next-gen O/S products around VM's.
Licensing???

Another question:
  • If MS kernels ship with a hypervisor, how do we dual-boot or run our own VM like XEN?
  • Would they be stupid enough to "embrace & extend" the VMI API/paravirt_ops?

The actual talk was on virtualisation and its impacts.. [143Mb]
<http://www.acm.uiuc.edu/conference/2007/video/UIUC-ACM-RP07-Traut.wmv>


<http://blogs.zdnet.com/microsoft/?p=842>

Traut spent most of his time describing Microsoft’s thinking around virtualization, and how virtualization can be used to ease backwards compatibility and other problems Windows users incur.


Microsoft has created a stripped-down version of the Windows core, called MinWin, that will be at the heart of future Windows products, starting with Windows 7, the Windows client release due in 2010.


MinWin is 25 MB on disk; Vista is 4 GB, Traut said. (The slimmed-down Windows Server 2008 core is still 1.5 GB in size.)

but no graphics subsystem


The MinWin core is 100 files total, while all of Windows is 5,000 files in size.

Runs in 40Mb memory. Ascii only

MinWin will be at the heart of future versions of Windows Media Center, Windows Server, embedded Windows products and more.

First good MSFT decision I've heard in ages

Traut said he is running a team of 200 Windows engineers working on the core kernel and Windows virtual technologies.

C.f. 10,000 total on Longhorn/Vista. Say 3,000 coders

(he) said that Microsoft is operating under the premise that “at some point, we’ll have to replace it (the kernel),” given that it “doesn’t have an unlimited life span.

That's important news

2007/10/05

Open Source - Barriers to Entry

Open Source - Barriers to Entry


I think I have a short, coherent description of the underlying cause of
the barriers to adoption to Open Source:

"Some Thinking/Expertise Required"
(as in "Some Assembly Required" or "Batteries not included")

It stems from:
Is IT well-managed?

Which leads to:
Is "Mangement" generally practiced well??

To both of these, my answer is a strong "NO" - it's all about failure of
management.

The Usual Management Method


I've seen very consistent behaviours, attitudes and approaches across
every organisation I've worked in [a very large cross-section]. I don't
know where they arise or how - but there are best described as
'unschooled' or 'hard knocks'.
Certainly not 'insightful', educated nor informed... That appears to be
anathema.

I've met precious few managers that I'd call competent, let alone good.
And very few who'd bothered to train in their work.
One (a scientist in charge of 100 ppl and a $30M budget) bragged "I've
never done *any* management training".
His PhD in biology qualified him for everything...
[The subtitle of "Other Peoples Money" is: 'Arrogance, Ignorance and
Self-Delusion'. Wide-spread and

Perhaps this one point, consistent management training, is the reason
IBM dominated the computing industry for 3 decades...
[And their avarice/denial brought them undone]

Professional Management & Management Profession


'Management' doesn't qualify as a 'Profession' under my definition:
  • an identified & testable set of practices, skills, competencies
  • (behaviour?) [think pilot or surgeon]
  • means to provide barriers to entry and disqualification/discipline
  • Improvement/Learning mechanisms:
  • by invention/discovery
  • by incremental improvement
  • analysis of failure & root-cause analysis + corrective actions (think 'bridge falling down' or plane crash)

IT Management and general Management


Without Professional & competent business managers, there can be good
management of IT.
Without good IT management, good practices and competent practitioners
are rare and can't be maintained...

Summary:
IT is populated mostly by a bunch of rowdy, undisciplined 'cowboys'
that are set in their ways and do what they please.
IT management is about politics, influencing and pleasing, not any
rational, objective measures.

That explains the Fads & Fashions of Management, and the almost
universal CIO mantra "nobody got fired for buying <fad/fashion>".
And of course:
  • Risk Avoidance & Blame Shifting [consultants & outsourcers]
  • CYA

Implications for Open Source business


How to use this premise?
  • Wait for the current fashion to collapse [or have cracks]. All fads & fashions change.
  • Find the few competent business & IT managers out there and sell to them...
  • Sell them camouflaged/disguised systems - like embedded devices or appliances (e.g. network, storage, security)