3 ms·
I would assume it wouldn't. The key to this decision is the fact that Azure is clearly in trouble. Almost no one uses it including .Net developers (I have to
by TomOfTTB 15y ago
I would assume it wouldn't. The key to this decision is the fact that Azure is clearly in trouble. Almost no one uses it including .Net developers (I have to assume Microsoft was counting on their support right off the bat).
That's key because it tells you how ScottGu is viewed within Microsoft (you don't put a failure in charge of an important project that's already failing). Which in turn means the direction he's leaving behind on the .Net platform must be one the upper execs are happy with.
So IMHO you won't see any dramatic shifts away from the current direction
- vyrotek 15y ago"Almost no one uses it including .Net developers" Sources? I know of 3 startups (mine included) which visit HN that are based on Azure. I am also familiar with many other larger companies which are using Azure. This seems to be the only sort of public exposure on companies using Azure. http://www.microsoft.com/windowsazure/evidence/ http://www.microsoft.com/windowsazure/evidence/ The perception that the platform isn't being used is usually because those companies don't care to advertise their infrastructures. They sort of have more important things to do than write a blog entry on how much they love Azure.
- TomOfTTB 15y agoWell the article itself says... "As I blogged last week, Microsoft is seeking ways to boost sales of its Azure platform. One way the company is hoping to do this is by tightening up its developer story around Azure." While Azure was largely blamed for Ray Ozzie's departure (http://itknowledgeexchange.techtarget.com/IT-watch-blog/ray-ozzie-leaves-microsoft-another-victim-of-clippy http://itknowledgeexchange.techtarget.com/IT-watch-blog/ray-...) You can look at the anemic selection of Azure Books: http://www.amazon.com/s/ref=nb_sb_noss?url=search-alias%3Daps&field-keywords=Azure&x=0&y=0 http://www.amazon.com/s/ref=nb_sb_noss?url=search-alias%3Dap... (By comparison look at WPF books a technology most people believe will disappear into Silverlight in coming years: http://www.amazon.com/s/ref=nb_sb_noss?url=search-alias%3Daps&field-keywords=Azure&x=0&y=0#/ref=nb_sb_noss?url=search-alias%3Daps&field-keywords=WPF&rh=i%3Aaps%2Ck%3AWPF http://www.amazon.com/s/ref=nb_sb_noss?url=search-alias%3Dap...) I'd also point to the number of views on the average Azure article on DotNetKicks: http://dotnetkicks.com/azure http://dotnetkicks.com/azure But to be honest I mostly based it on my own experience. I know many .Net developers and all of them have chosen to ignore the benefits of Azure and go with the more difficult task of implementing on EC2. Note I'm not saying Azure isn't a good choice. I'd even argue it's a better choice in many ways than services like EC2 (Scaling alone makes it worthwhile for .Net developers). I'm simply saying it hasn't been a success for Microsoft thus far.
- vyrotek 15y agoAzure definitely has some hurdles to overcome still. I was actually down in Redmond visiting the Azure team a couple of months ago and both complained and learned a lot. It seems their original focus for almost a year was building corporate migration tools and features which I think really hurt them in the community. But thats because those big companies were large paying customers. There are a lot of features that are just now getting the attention they need. (Table Storage, Caching, etc) Hopefully with those issues resolved it will be more attractive to other companies and developers. I think its funny that you bring up WPF/Sliverlight because I've been fairly against the technology in general and have avoided it whenever possible in my .Net career. I quickly fell in love with ASP.Net MVC and will stick with internet standards when possible.
- kenjackson 15y agoWPF/Silverlight and ASP.NET MVC are very different technologies. For one WPF/Silverlight are client technologies. More like HTML/CSS/JS than MVC. But what I really wanted to say is that while I like Azure in theory, it's a pain in the butt to use. I used AppHarbor and literally had my first app hosted in 10 minutes, maybe less. With Azure it seems like there's all these things they want me to know. If I want to get up a super simple web site with ASP.NET MVC why are you making me know about all these different types of roles and load balancing and such? My hope is that this is what Scott Guthrie will fix. ASP.NET MVC is surprisingly clean and pleasant. Silverlight is a triumph of moving WPF off the desktop. I'm really excited about the future of Azure, for once.
- luffy 15y agoIt's best to look at it like most of Microsoft version 1 products - it's going to be painful to use. But it has tons of potential - more so than any product in recent memory. And it's actually pretty reasonably priced, even if you aren't in BizSpark. You really don't need to worry about roles that much. A web role is a website, and a worker role is a non-website service (though you can connect to a worker role via http). In theory, load balancing his handled for you so you can achieve redundancy with at least two instances. Now, there are some real painful points with azure as it stands. My two two specific points are these: deploying takes a long time - some times over 10 minutes. So forget about agile build processes or continuous deployment or even making a quick, simple text change. The other major problem I have is you have to run the debugger to get access to the development storage ( blob & tables). Both of these are really useful. But running the debugger all the time is torture. Developer tools are hugely important to all of the MS ecosystem, and I bet these concerns get addressed - this is the big advantage for moving ScottGu over to the azure team. Azure is a big part of the future of that company.