Re^2: Nobody Expects the Agile Imposition (Part VI): Architecture (Python 3)

Re Windows 95, thanks for the correction, I'll update the root node. It seems that Windows NT was a rewrite of Windows 3 and that Windows 95 was derived from the Windows 3 code base.

Python 3: I only claimed it was meeting "substantial resistance". Maybe that's unfair, depending on your interpretation of "substantial", but it's certainly meeting some resistance based on random web chatter on the subject. Well, I'm a Python user and I'm resisting it. ;-) My personal opinion is that breaking backward compatibility was unwarranted for a release with relatively modest improvements. Many businesses with large investments in Python 2.x code will resist Python 3 indefinitely because upgrading will prove too risky and/or too expensive.

Update: Even a company as wealthy as Google, as of 2017, are still heavily using Python 2. This is hardly surprising. Where is the ROI on spending millions of dollars rewriting millions of lines of already working code, without adding any customer value, while being almost guaranteed to suffer numerous breakages to critical business systems? You also pay an Opportunity cost. Curiously, I see some of Google's legacy Python 2 systems are being rewritten in Go perhaps because at least there is some perceived customer value (faster performance) in a Go rewrite. For smaller less wealthy companies, rewriting millions of lines of working Python 2 code in Python 3 could well put them out of business. Of course, if you don't have much Python 2 code, switching to Python 3 is a no brainer. Further update: I see Jython is still Python 2 in 2020 and IronPython3 unfinished. At least Perl doesn't have to worry about updating Java and CLR versions of the language. :)

See also:

Martian Headsets by Joel (on breaking compatibility)
How Microsoft lost the API war by Joel
Robustness Principle (sadly Postel's Law "Be conservative in what you send, be liberal in what you accept" is double-edged)
Improving p5p: Perl is going to stay Perl (and Stroustrup quote in this reply)

Comment on Re^2: Nobody Expects the Agile Imposition (Part VI): Architecture (Python 3)

Replies are listed 'Best First'.
Re^3: Nobody Expects the Agile Imposition (Part VI): Architecture by BrowserUk (Patriarch) on Jan 23, 2011 at 12:13 UTC
It seems that Windows NT was a rewrite of Windows 3 Absolutely not. Windows NT was an entirely separate, new development of 32-bit code. Ie the Win32 API. Win32s was a thunked win32 emulation retrofitted to the 16-bit Windows 3. Examine what is said, not who speaks -- Silence betokens consent -- Love the truth but pardon error. "Science is about questioning the status quo. Questioning authority". In the absence of evidence, opinion is indistinguishable from prejudice.	[reply]
Re^4: Nobody Expects the Agile Imposition (Part VI): Architecture by eyepopslikeamosquito (Bishop) on Jan 23, 2011 at 12:24 UTC
Yes I knew that, so maybe I'm using the term "rewrite" imprecisely. To me, it qualifies as a rewrite because Microsoft chose not to adapt the existing Windows 3/95/98 code base for future Windows versions (e.g. Windows XP, Windows Vista) but rather to write a new code base from scratch. How about: Windows NT was a rewrite of the Windows operating system?	[reply]
Re^5: Nobody Expects the Agile Imposition (Part VI): Architecture by BrowserUk (Patriarch) on Jan 23, 2011 at 12:50 UTC
Hm. I'm not sure that the term "re-write" comes into it anywhere. Windows NT was a completely new OS written from scratch. Bits (Kernel API; UI; etc.) of that new 32-bit OS were backported to the newer versions of the old 16-bit versions of Windows in order to provide a migration path. Saying one is a re-write of the other is like saying OSX is a re-write of MacOS. They are both essential very thin skins of look&feel layered over completely different underpinnings, design philosophy and target audience. Like saying this is a "rebuild" of this, because of their superficial similarities, completely ignoring the difference in their purpose and technology. Examine what is said, not who speaks -- Silence betokens consent -- Love the truth but pardon error. "Science is about questioning the status quo. Questioning authority". In the absence of evidence, opinion is indistinguishable from prejudice.	[reply]
Re^6: Nobody Expects the Agile Imposition (Part VI): Architecture by eyepopslikeamosquito (Bishop) on Jan 23, 2011 at 19:25 UTC
Re^7: Nobody Expects the Agile Imposition (Part VI): Architecture by BrowserUk (Patriarch) on Jan 23, 2011 at 23:09 UTC
Re^7: Nobody Expects the Agile Imposition (Part VI): Architecture by tilly (Archbishop) on Jan 24, 2011 at 06:02 UTC
Re^7: Nobody Expects the Agile Imposition (Part VI): Architecture by Jenda (Abbot) on Jan 25, 2011 at 17:58 UTC
Re^4: Nobody Expects the Agile Imposition (Part VI): Architecture by girarde (Hermit) on Jan 25, 2011 at 03:59 UTC
If NT was a re-write of anything, it was a re-write of VMS. ;-)	[reply]
Re^5: Nobody Expects the Agile Imposition (Part VI): Architecture by Limbic~Region (Chancellor) on Jan 25, 2011 at 14:34 UTC
girarde, `Windows NT (WNT) - 1 = VMS IBM - 1 = HAL` [download] Coincidence, I think not ;-) Cheers - L~R	[reply] [d/l]
Re^6: Nobody Expects the Agile Imposition (Part VI): Architecture by sundialsvc4 (Abbot) on Jan 25, 2011 at 15:01 UTC
Re^5: Nobody Expects the Agile Imposition (Part VI): Architecture by BrowserUk (Patriarch) on Jan 25, 2011 at 04:09 UTC
Indeed. With big chunks of RSX thrown in the courtesy of Dave Cutler.	[reply]


P is for Practical
	PerlMonks