It would be really nice to know what technical reason there was with freezing IE development on Windows XP before the OS itself reached EOL.
It has been standard practice for Microsoft for many years to ship the DLLs required for a feature in the OS as redistributables, features such as updates to OLE, new runtimes for Visual C++ or Visual Basic, new APIs, new images processing libraries and audio and video codes, and many other things. Software like Office would even ship with new UI components that could be reused.
Building a version of IE that did not require OS sandboxing features was also not an impossible task and would have maintained the backwards compatibility that Microsoft holds almost sacred better than a version of IE that was is still frozen in time, and is the last version that will ever be shipped for Windows XP.
This has broken the trust and relationship with web developers, even those that used IE as the gold standard and built their sites to match every spec and feature promulgated by Microsoft in the past. (VML, behaviors, filters for instance).
I work in an industry where compatibility with IE8 is expected, and spend the same amount of time that many others do maintaining a veneer of compatibility with that browser, that rendering engine, that HTML parser. This requires heavier libraries than I would prefer to use (such as jQuery) because it maintains an abstraction over the incompatible parts of the DOM implementation.
Another example of this is the incompatibility with SNI which has driven a requirement for unique IP addresses per SSL site and driven those than do not want to pay for such as limited resource (assuming IPv4 addresses) to share SSL certificates, use wildcard certificates that open issues of their own, or forgo the security provided by SSL entirely if they still require compatibility with Internet Explorer versions and OS versions that don't implement SNI natively.
This also applies to current more secure versions of NTLM which are and were supported by third-party browsers on the same OSs that Microsoft did not offer support, by using a non OS provided SSL library which implemented the Microsoft driven standard.
It's true that the IE team has done a better job of communicating with IE Blog and other channels, and has done a better job of getting out in front of upcoming standards, delivering impressive performance on complicated implementations (WebGL) and otherwise kicking some ass.
But they and Microsoft still have to rebuild the trust and support of the developers they lost not just as users, but the developers who have come to dread working with technologies they once loved due to the heartbreak that IE6, 7, and 8 inflicted upon us.
I'll end where I began, with a plea for an explanation for not shipping a newer version of Internet Explorer on Windows XP before the end of life.
And best of luck.
Edit: I include 6 and 7 in reference to current support requirements, not the advances they represented when first released.