Hacker News Reader

Err, that's kinda greatly simplifying things. I'd say the last 10 years can be described as:

- ~2005 to ~2009 was the Ray Ozzie era. Ozzie took over the role of Chief Software Architect from Bill Gates in 2006 and the company, internally and externally, was focused on very ambitious projects. Software 'architects' were common on every team and usually doing something the 'right' way trumped shipping quickly and iterating. There were lots of interesting ideas and cool tech, but turning that stuff into viable products and businesses proved to be difficult.

- ~2009 to 2013 was the Steven Sinofsky era. Sinofsky took over as the head of the Windows division after Vista and helped recover with the Windows 7 release. His influence quickly spread and most of the company adopted his management structure (very strict separation of concerns between development, test, and product management). The role of software architect quickly disappeared (and Ray Ozzie left the company, leaving the CSA position to disappear entirely) and although research was still important it was definitely not the top priority. There was much less focus on building platforms, ideas, or other intangible things vs. shipping to customers in controlled and predictable ways.

Post 2013 and Steve Ballmer stepping down I don't really know much since I left around that time.

> 2009 to 2013 was the Steven Sinofsky era... There was much less focus on building platforms, ideas, or other intangible things vs. shipping to customers in controlled and predictable ways.

For certain definitions of "controlled and predictable," yes. For others not so much. Sinofsky's tenure was also the era of the great push to turn Windows 8 into the one Windows to rule all devices, which resulted in it being the controversial, poorly-received product it has been. "Controlled and predictable" would have involved rolling out changes like moving to Modern/Metro/Windows-Style/etc. UI, removing the Start button, booting to the Start screen instead of the desktop, etc. slowly and incrementally, but Sinofsky bet instead on rolling them all out at once in a single "big bang" release. That was a big, risky bet that completely failed to pay off, and that MS has been retreating from step by step ever since his departure.

From my perspective, "controlled and predictable" with Sinofsky was about managing the release cycle, hitting the dates. e.g. We said we're shipping in May, so we're shipping in May. That means careful planning and aggressive feature cuts if necessary. The goal was on time with high quality, where quality is measured in metrics like crash rate and bug count, not public opinion surveys.

Incremental releases and minimizing controversial changes might make releases more "controlled" in different ways, but those weren't the target.

[Edited for clarity]

That sounds wonderful to me right now.

Sometimes I find "architects" curious folks. XAML was developed during the ray ozzie era right? That certainly has a "right way" with "Model-view-view-model" levels of abstraction insanity. Perhaps simpler is better.

MVVM is a very good model though. I certainly enjoyed working with it.

I just recently left my job at Microsoft, about 7 months ago. That said, I have nothing but mostly great things to say about my time there (2010-2014). By the time I got there, there was already the sense that even though we had huge market share on traditional desktop operating systems, that we were still far and away the underdog in many other respects that probably mattered more. Working on the dethroned Windows OS was an exciting and compelling thing. How do you stay relevant in a market that is rapidly changing without just outright copying the competition and while still giving it its own unique perspective? That's a really hard problem, and a really fun one to work on. Even though I know we didn't get it all right in 8/8.1, the excitement was all there and it was a really great ride.

Ask yourself if you would rather work for a big software leader that is resting on its laurels or for a company that knows it has some ground to cover if it wants to stay relevant, and then you'll know if you prefer the old Microsoft or the new one.

At the moment I'm looking at trying to compile some stuff on Windows, that has Visual Studio project files, with clang instead of Microsoft C++. This would involve making msbuild do things it wasn't exactly designed to do (assuming I don't want to spend a lot of time duplicating what it does from scratch). It occurred to me maybe what I really need to do is modify msbuild, but of course that's out of the question, right? But just for the heck of it I ran a Google search...

... and msbuild was recently open sourced. The code is right there on Github.

Yeah, Microsoft of 2015 is very different from Microsoft of 1995, in good ways.

You actually can extend out MSBuild to use alternative compilers.

It already comes with rules for C#, VB, C++, F#, etc. Replacing the rules for C++ with another command line and set of parameters of quite doable! :)

(I happen to be using such a build system, unfortunately I wasn't the one who wrote it so I cannot provide any direct help, just encouragement!)

It can actually all work from within VS as well.

Microsoft is definitely, under the lead of Satya, becoming more open and friendly. But does it necessarily make it cooler? That is very much left to debate.

The following claim might be a little generalized, but I believe Microsoft is still largely considered as an software company, who sells licenses and units of products for a living, and very much enterprise focused. This is quite different from the image that younger companies, especially those service-oriented ones, show or at least try to show to the public, something that is more ubiquitous and intimate.

Yes, Microsoft is cool again.

>Four or Five years ago, if you'd asked me what tech company I worked for, my answer would have probably been, "Microsoft, I know, it's a bummer. I'm trying to move to google". Now though people actually think MS is cool again. The stuff we're working on is exciting and the corporate culture is starting to get healthy.

It's embarrassing how fickle the technology industry is... so much reliance on hype and trend-hopping.

yeah, embarassing to want to work at companies doing innovative and interesting work. The whole side effect of probably getting a raise to go with a new job is terrible too.

I'm happy to hear the culture at your company has improved but I don't know if Microsoft is considered "cool" - or whatever that means.

Anyone who has been around long enough (and not really that long all things considered) knows better than to ever trust Microsoft. Their whole model has been, and always will be: "Embrace, extend, extinguish".

So Microsoft is opening some things up? They've had no choice as they are being left behind more and more. They know the modern developer is not as interested in an expensive IDE and MSDN account when there are many alternatives. We are creating so much more software than before and the top down model Microsoft sold has become niche after many years of dominance.

Any open project Microsoft gets involved in and any open standard they get involved in should come with a warning label. Why? Because this is in their DNA.

Anyone who has been around long enough knows not to trust Google, Apple, Facebook, Oracle, ... Microsoft is not uniquely untrustworthy when it comes to big multi-faceted companies. As for DNA, I haven't seen any dodgy behavior in the OSS area for 5+ years, and nothing like EEE since I started working here in 2007.

No, Microsoft is a special case because they have always been at best amoral; they have been a cutthroat, even malicious company. Recently they've softened their behavior and image, but it's really a very recent shift and given their past treachery it's hard to trust them.

These other companies may not always work together in harmony, but at least they aren't preoccupied with extinguishing their competitors and dominating ("owning") their users.

> at least they aren't preoccupied with extinguishing their competitors and dominating ("owning") their users.

That's weird, because my perception of the companies listed above is that they are exactly preoccupied with that.

Microsoft is a special case, but not for that reason.

Microsoft is a special case because they had a zero-sum game/scorched Earth mindset, and the ability to lose billions of dollars per year, indefinitely, achieving it.

Microsoft was not content at allowing potential new ideas/technologies develop which might grow the market and improve things because they saw everything as a potential threat to their position. So they would do everything from trying to poison the market, to pouring (losing) tons of money into competing products, and sometimes just buying up companies and let it rot.

The very special part is as long as Windows and Office were cash cows, stock holders weren't willing to lead a revolt for Microsoft losing billions on everything else.

Not all companies are not scorched earth (SGI gave away serious resources to Nvidia, Valve recently gave away their VR tech in the highly publicized layoffs a few years ago).

Other companies might like to play scorched earth, but they don't have the ability to piss away money like Microsoft. This is why Microsoft is a special case.

Microsoft still has tons of cash and Windows/Office is still a cash cow. And it is generally accepted that company culture is hard to change once established (especially here on HN). This is why people with historical context do not trust Microsoft, with good reason.

So how was MSR part of a scorched earth strategy?

I am really tempted to set up a new account just so that I can post "Embrace, Extend, Extinguish" to each MS article. Just that. Three words. Each article.

The only reason I don't do it is that I'm afraid it won't be distinguishable from a small percentage of the HN readership who do so seriously.

Really, it gets boring.

By going cross-platform on C#, Microsoft is coming to eat Java's lunch. Throwing their weight behind .NET Core ASP.NET on Linux is a clear shot across the bow at server-side Java, which, let's face it, is a shit-show compared to developing in C# these days.

As someone who hasn't used an MS OS for at least ten years (though Star Citizen has forced me to), the open source thing going on is really starting to turn my head. Now, that being said, I feel like MS has a long way to go before I'll start to prefer it over *nix (I would love to see some of the kernel source, for example), but it's a direction I am very pleased to see Microsoft headed in. I have mad respect for the people at MS pushing this forward.

> The stuff we're working on is exciting and the corporate culture is starting to get healthy.

... because the "positive" energy retired.

I love the way MS has smartened up in regards to open source and embracing developers in the modern era. However, I think they still have some very fundamental problems in terms of corporate culture. For a very, very long time the reviews system incentivized short-term, easily measurable work over more substantive work, with a fairly obvious impact on the products delivered by MS.

Was stack ranking already common practice when this article was written (2004) and is it actually over now (2015) ?

stack ranking had been going on for many review cycles at this point and I heard last night that it's gone for good (wink wink, nudge nudge) because they still have to allocate bonuses and stock.