Why are all the 'cool' jobs no longer using .NET?
reddit.com
reddit.com
The only people that know about the new things MS comes out with are the few of us .NET developers also with a foothold in the Startup scene, and old enterprise companies who bought into the stack back in VB5/C++/MVC days.
The other uphill battle is young people's view on technology. Most of them would rather platform-hop in the same way people change churches (and for much the same reasons). New = better, and Microsoft = old, so no need to even see what they have to offer. Young people with ample free time have always driven the software industry, and young people want to use things built by startuppy or nonsensical names. "Microsoft? They actually have software in the name? No thanks!"
Finally, it's an issue of evangelism and online/community presence. Because enterprise is the bread and butter of MS, MS tends to target those kind of people more. It's not that they don't try to capture new minds, but when you attend an event where you're the lone startup guy among 50 enterprise people, how do you think presentations are going to be focused? When your community seems to be centralized on your corporate domain, are you really going to attract someone who isn't willing to type in "microsoft"? New platforms get popular not because they're actually better, but because they're cooler. Either cooler sounding, there's a new type of tutorial (to-dos and 5-minute-blogs are played out, people), or it's got a guy with weird hair promoting it, using Jobs-esque language like "beautiful" or "elegant". Of course it's elegant at version 0.9 before it really does anything. Look at Rails now that it actually does stuff.
Anyway, mini-rant over. I think MS deserves more attention than they get, but ultimately they dig their own grave by not investing more in capturing that next generation.
Oh and if you start building on that infrastructure using AD, Exchange, etc you're in CAL licensing hell. Next thing you know you've spend the wages of two really good engineers on just MS licensing. How is that cost efficient? I'd rather have a 5 man team running linux than a 3 man team running MS.
Not to mention, if I'm recruiting young talent, how many of them will have MS enterprise experience vs linux-based experience?
And don't get me started on the never ending stream of TLAs coming out of Redmond that get hyped for 3 years and then discontinued later for another TLA.
Also, I'd say that I'm more than willing to pay good money for VS, if it means not using Eclipse. I would easily lose $600 worth of time trying to get open source development tools to play well together. Microsoft's IDE technology is so far beyond anything in the FOSS world it's not even a debate anymore. Only mod it needs is Resharper. :)
It always pains me to hear such cries from someone who claims to be a software developer/engineer. I've used .NET (among several other) technologies for some years and had a few co-workers who would openly fret about Microsoft's direction and what horrible things they might have to endure if .NET went out of favor. Gosh, you mean you might actually have to learn a new technology? What a truly awful experience. I could never relate to that and I legitimately felt sorry for them. I hope to never become afraid or resistant to learning and am always ready to admit how little I really know.
The language/framework is but a means to an end. Each with strengths and weaknesses, but all of them suitable for a wide variety of uses. If you feel that your knowledge as a software developer is so language specific that you fear having to move away from it, you're doing it wrong.
As of .net 3.5, the .net framework contained 98 Assemblies, 309 namespaces, and 11,417 types. The framework is gigantic, and after you've spent long enough in .net, your domain specific knowledge of the framework itself is gigantic as well.
As someone who has been building applications on .net since beta, I have no fear at all in learning a new language (bring it on!). I have a fear of losing over a decade of experience with the thousands of classes provided in the .net framework that I use, and rely on, on a daily basis. I know exactly what I'm looking for, how it works, and what the performance ramifications are of each component I choose to use. I know what needs to be replaced with something custom to scale. I know what's doing extra work I don't need. I know the ASP.net page and control life-cycle (intimately), how the view engine works (and how to add features to it), how to replace critical components for custom functionality and how to build custom controls. I know the WPF layout system. I know the quirks in windows forms and GDI+. I know programmatic drawing in both frameworks. I have done machine learning in .net. I have done massive file IO in .net. I know the .net multi-threading model. I know how to do code generation in .net. I know how to use raw sockets in .net. I know how to bend this massive framework to my will, and get all of the performance and power I want out of it. I have built reliable, predictable, massive scale applications in .net. Over my entire career I’ve invited product managers to give me their crazy ideas, and I have never had to push back.
Having said that, I also don't want to learn Ruby. I don't think it's mature enough to get my attention. I don't see how it is in any way better (that's not to say that it isn't, I just don't see it). Is it easier to learn? Maybe, but I've already crossed that hurdle; just like OP on reddit has. Is it a faster platform to blast out a simple website? Likely not for me. Is it a better platform for building and scaling massive applications? I doubt it.
Your language/framework defines the outer capabilities of your application, the performance of your application (read: scale), as well as what parts of your application can be reused from parts made available by others. They are things to be mastered. .Net takes a long time to master. The "jack of all trades master of none" attitude is what pains me. Show me that ruby is better for me in an area that .net is lacking and I'll learn it. Show me that it's better in all of the areas I care about and I'll master it. That will be a hard sell. .Net is a fantastic platform; well organized, discoverable, consistent in quality, exhaustive in scale, performant, and still growing.
Will .net last forever? I don't know. Older Microsoft technologies are still around and .net is still in active development. I don't see Microsoft giving up the massive investment they have made in .net any time soon. Whether or not Microsoft stays relevant in web programming is up to them, and up to those of us using it.
As to startups using .net, I've worked for two. One provided a product that allowed you to order print ads made for your franchise. It accepted photoshop PSD files from designers, took them apart and allowed the ordering party to swap layers and text to fit their business, all with a WYSIWYG live-previewing interface. I wrote the original system myself in 2004 and the company is still operating today. The second was an online ad distribution company. The system I helped build was capable of targeting, serving, and tracking over 500 million ads per second at peak (a major accomplishment in 2008). The company is still operating today. I was not told what the companies did until after the interviews and pay negotiation. There was no way to know before I accepted the jobs whether or not they were "cool" jobs. Before starting my own business I was invited to work on a high frequency trading system built in .net. I turned it down, but they are doing very well. I've worked more for large businesses and mega-corps who also wouldn't tell me the nature of the job they were hiring me for until I was hired. How do you know ahead of time if a .net job is cool when all you hear about it is “Senior .net developer needed”?
Is there a lack of marketing from MS extolling the values and virtues of .net on the server (especially for start-ups)? Yes. But that's always the case with Microsoft. They are terrible at marketing. They will always be terrible at marketing. Should you be using .net at your start-up? That’s up to you. I’m using it for mine.
The "jack of all trades master of none"
To me, the trade in question is software development/engineering. If we are tradesmen, then .NET is one of the toolboxes available to us. Yes, it's a very advanced and developed toolkit, but I don't view it as an occupation. Now if you constrain yourself and simply call yourself a .NET developer and not a software developer, I suppose that would be an exception.
Will .net last forever? I don't know. Older Microsoft technologies are still around and .net is still in active development. I don't see Microsoft giving up the massive investment they have made in .net any time soon.
My experience is mostly on the desktop side (application dev for medical devices) and it has been the opposite. Look at technologies like WPF and Silverlight for example. I don't think you can honestly say that Microsoft developed these to maturity and didn't to some extent abandon them.
Your language/framework defines the outer capabilities of your application, the performance of your application (read: scale), as well as what parts of your application can be reused from parts made available by others. They are things to be mastered. .Net takes a long time to master. The "jack of all trades master of none" attitude is what pains me. Show me that ruby is better for me in an area that .net is lacking and I'll learn it. Show me that it's better in all of the areas I care about and I'll master it. That will be a hard sell. .Net is a fantastic platform; well organized, discoverable, consistent in quality, exhaustive in scale, performant, and still growing.
I think it's a bit contradictory to say that you need to master a language to understand it's capabilities, yet you won't explore other languages unless it's proven to you that they have capabilities that .NET lacks. That said, I think your viewpoint is mostly valid if you have full control over what technologies you work with (e.g. starting a startup, or only applying to .NET companies). However, that is true for very few developers and this is where the whole idea of fretting over the future of .NET comes from.
In my field (medical imaging and devices), I have used plenty of .NET, but I could never rely on it entirely. The foremost medical image processing and visualization frameworks and libraries still reside in C++ and although there are .NET wrappers, they are almost always limited, several releases behind, and with a much smaller community (there are a few exceptions).
To be clear, I love .NET and even moreso Visual Studio. However, if I'm asked to propose a technology for a product, I'm going to do my research and pick the one that I feel is best-suited for it and not let my .NET experience be a guiding force in that decision (unless this is an extremely time-sensitive project and I represent a significant amount of the developing team, then of course it needs to be considered).
I understand where you're coming from and I don't see your viewpoint as wrong. Mine is different. I see programming as part of an art, the way applying paint is part of creating a painting. It is by-far not the only part.
I totally agree that a good programmer should branch out and gain a better understanding of the world, but I believe the branching out should happen in a different direction.
To call someone that writes books a writer is an oversimplification of their job. Writers, like programmers, specialize. One writer might write non-fiction books in English in the murder mystery genre for an adult audience. This writer has had to develop a solid experience in and handle of the English language. They have had to research and understand murderers, the law and police. They need a solid handle on psychology to be any good. They need to understand the implements of murder. They need to understand the standard plot points of their genre, and how far they can be bent (to interest readers) before they break (and turn readers away). They need to understand pacing, and how to capture an audience's attention consistently. How to keep their material accessible to expand their audience as much as possible, or how to narrow their focus to be a must read for a certain niche. They must know how to get published and avoid getting swindled. Putting words on paper is only a very small part of their job.
If our hypothetical writer is going to branch out into new areas, he's more likely to gain success writing a sci-fi novel (where he can reuse most of his skill) or starting his own publishing company than he is in learning French to write books for the French audience.
Learning basic French is not hard, but it is very hard to know French like a native French speaker (an expectation of your French readers). If he is thinking in English and writing in French, his efficacy is already compromised. If you are writing for the French audience, you will have to take French customs and the French psyche into account. French readers are likely very different from the ones he’s used to, so he needs to understand the psychology and mind-set of his audience and his characters. Publishing a book in France is also likely completely different, with different ways to lose money in a book deal. Learning basic French will only help our writer when he needs to integrate a French character into one of his books.
It would be a little silly to tell a writer that he needs to identify as an English writer if he doesn't know French.
I make consumer products. I am a .net platform developer. I started my career as a designer 14 years ago; and soon realized that there is too much compromise in just graphic design. Every design I submitted at the company I worked for had to be changed (often dramatically) to fit within technical limitations. My designs were created to appeal to my audience, and due to technical limitations, that appeal was compromised. I had already been programming at home for a few years, and decided it was time to become a professional. The second month at my design job I asked to be put on the programming team as well. At the time, we were writing in legacy ASP. I never had a design compromised by technical limitation ever again. When .net was made available, I jumped on it immediately. I loved the ideas behind it and I couldn't wait to be a part of it.
A year or so later, I started realizing that the database design for a new product I was building was hampering my ability to write clean software (also due to “technical limitations”). I asked to be added to the Database team as well. I became a DBA. I was still a designer and developer. A year later, I became a PM as well, and so on.
At this point I have held the titles of senior .net architect, .net team lead, senior .net developer, senior MSSQL database architect, senior product manager, creative director, senior graphic designer, business analyst, and research scientist. I am always a .net developer, but in every job I've had at least two other titles as well. Each of these roles interface with the others, and knowing the others makes doing each of them much easier.
My job is to make quality consumer software. My job is to appeal to an audience. I learn whatever I need to in order to better appeal to that audience. As a whole my audience doesn't care what the software is written in. They care about whether or not it’s easy and pleasurable to use, whether or not it’s fast, intuitive, and solves their needs efficiently.
Now I have my own software company.
-----------------------
WPF works, has a number of great features and tricks, and it’s fast if you’re using it right. I assume the developers on WPF moved to the Metro dev team when that ramped up; I don’t know. Maybe they’ll be back. My only issue with WPF is its half-hearted 3D support. Silverlight was an experiment, and it showed. It’s still possible to write great stuff on Silverlight, but the limited audience keeps it from being a solid solution unless you’re doing something that can’t be done as well using something different (DRM video for example) or you are creating an intranet application and have control over all of the machines that will be using your application. It’s a shame, but that future was easily foretold when it was announced.
-----------------------
I don’t think you don’t have to master a language to understand its capabilities, but woe to he that tries to exceed them without having a solid understanding of what’s going on under the covers.
In my experience, .net shops are .net shops. Maybe I have been lucky.
In my experience it’s pretty easy to write your own wrappers for C++.
What sort of projects (other than projects targeting non-windows machines) have you encountered where you suggested not using .net?
I did .net for 8 years, and for a long time I used to take on freelance projects. There came a point where I could get the same project done part time in django or rails in 2 weeks that would take me 2 months to write using ASP.NET, nhibernate and an elaborate domain layer with everything split up in two projects.
If you wanted it to take anything less than 2 months then you had to start compromising and using some idiotic ms rolled ado datasets and any self respecting .NET engineer wont stoop that low.
With django or rails, I write clean, pragmatic, domain driven, and test driven code that is best of breed. It does not take 10 years of fiddling with an API to become productive with it - it would take maybe a month once you take the plunge. And for what its worth, I can care less about getting maybe 1/10th of the raw performance that I would get with the CLR because scaling the app layer is the easiest part anyways.
> I spent too much time with .NET I don't want to learn ruby!
Um. That answers that.
Also, after doing lots of things in JS and Python over the last couple years, going back to .NET just feels too "heavy" somehow. The changes in C# and .NET over the years have been very nice so it's hard to justify this feeling. One definite downside of the .NET ecosystem, I would say, is that they take forever to do completely sensible things that everyone else is doing. Things like MVC, package management, deployment tools.
BTW, I have found that in interviews my extensive past .NET experience is viewed negatively. There are assumptions out there about .NET devs and in fact I would say there is some basis for those assumptions. At the same time, I was hoping that smart people would be able to see through that and acknowledge that not everyone who works with a technology is "like that", but alas. So it is a bit of a liability, unfortunately.
At that time, attempting to run Linux, Eclipse, Forte, Java, etc. was difficult, buggy, and slow - sometimes too slow to be usable. Until Sun sued it off the market, Visual J++ was the only usable interactive Java toolchain/runtime/base-classes with UI base classes that resulted in applications that looked and acted polished and clean.
The big difference today is that you can take a laptop out of a box and within two hours have a Linux-based Android development toolchain up and running without any hitches, and have an even more coherent managed language runtime targeted by that toolchain running clean beautiful mobile touch apps.
Open tools and OSs have been completely transformed. I don't really see the 13 years of comparable progress in VS and Windows.
This experience has not much to do with .NET per se; I really like C# and F# and sometimes wish I could use it all the time (not a big fan of Mono yet...). Maybe I'm just weird or set in my ways, but I don't really find the MS toolchain to be very productive. I'll retry it again when I get my hands on a Windows laptop (we're running only Linux and Mac OS X besides for WP8 dev and i'm not doing that at the moment).