Ask HN: What startups here are focusing on developing on top of .NET?
I perceive an increase on the number articles about C# and .NET. Since HN activity reflects our own interests, I am curious as who is developing what for .NET.
Anyone wants to tell their story?
Disclaimer: I was a Windows user and did a lot of Windows development up until about 2002. Since about 1999, I have been preferring Unix-like environments and languages such as Perl, Python and Ruby (and, sometimes, Java) for developing and deploying web applications. Sometimes, I am also quite vocal about my personal preferences.
(
66 comments
| 45 points
| on HN
)
Deploying applications on Windows is pretty easy, but there are far fewer hosting options for Windows; and the options that are out there will either give you a very locked down environment or require you to have a dedicated box.
Beyond that, it's kind of a pain in the ass to do a quick edit of a file to fix something when you're on the road if you don't have all of your tools available. It's just not a big deal to make single file changes in Perl, Python or Ruby because most people don't use a full-featured IDE to edit that code and as a result files are often structured differently (which is to say generally broken up into smaller chunks).
If you're thinking about starting up on the .Net stack, there's really no reason not to if that's the stack you prefer to use. Bizspark is a nice program which allows you to delay the purchase of most licenses for 3 years. Microsoft has a long history of providing nice tools for its developers as well.
There are definitely some big sites out there that are using .Net as well (Loopt, Newegg, Woot), and there are tons of developers out there that are familiar with the stack.
I would say if that's what you are the strongest with, and you like it, there's no real reason not to use .Net - as long as you don't mind using a proprietary stack. I haven't used Mono seriously, but I've heard good things about it. However, it's still a second-class citizen when compared to the canonical .Net stack.
I'm using a Windows/Apache/Django/Postgres back end for http://newsley.com. It's all running on an EC2 Server 2003 instance. I've used their Windows slices for about a year, and I've been really really happy with them. It's a bit more expensive than a Linux VPS @ $80 a month, but I think it's worth it.
I would really like Amazon to support Server 2008 on EC2, especially to make PowerShell Remoting not such a hack. You can get it to work, but it's a little dicey.
Beyond that, it's kind of a pain in the ass to do a quick edit of a file to fix something when you're on the road if you don't have all of your tools available.
Granted, I'm developing using Python on Windows 7, but I have a full dev setup on a $300 Netbook that I installed Windows 7 on. I've been building Newsley while traveling 50k miles the past 4 months, and I haven't really had any problems.
I can run Win 7 Ultimate and my dev setup on a netbook pretty seamlessly. Would there be a reason that Visual Studio wouldn't run on a Netbook with 2GB ram?
You make a good point about tossing Visual Studio on a netbook, I can't think of a reason it wouldn't run, but it's still got the requirement of carrying my own gear with me.
Honestly, it's not a big deal in the end, I just remote in to my main machine on the rare occasion that I have to make a fix while I'm on the road; I would just prefer to be able to do it anywhere regardless of what machine I've got handy.
Beyond that, it's kind of a pain in the ass to do a quick edit of a file to fix something when you're on the road if you don't have all of your tools available.
If this is a priority for you, you can use ASP.NET Web site projects and edit live versions without recompiling/redeploying (for the most part). Web site projects are Webforms-only though, no MVC. However, I think not having the ability to make ad-hoc changes to production code is a feature ;)
True, but in my defense we are a continuous integration / almost continuous deployment shop ;). As soon as commits are made our tests get run, the product gets built and configured and automatically staged. Usually what happens is I get a phone call that there's a problem, fix it and then let someone test it and push the big red "RELEASE" button.
It's actually a very good technology stack if you lean on the available open source options instead of just using everything Microsoft suggests. C# is a good language, ASP.NET is very efficient and scalable, NHibernate is a passably good ORM, and ASP.NET MVC is a solid foundation to build webapps on.
It hurts to have to pay license fees for hosting, but we're enrolled in BizSpark, which gives us free licenses for 3 years, and you can always look into running your app on Mono. (We are.)
Not sure what you're looking to find out, but feel free to ask anything. :)
I am a very curious person. I never know what I am going to find out until I do. The other way to read this is that I never know what I am after until it's too late... ;-)
I ended up here starting back when I got VB3 for a birthday present in high school (actually if you go back further, I started programming with BASIC on an Apple in middle school). When the web came around that translated in to ASP/VBScript and then around the time .NET 2.0 came out I switched to C# and haven't really looked back. So far, the framework hasn't limited what I want to accomplish. That may not always be the case, but for now I'm satisfied with it.
I'll echo what nkohari said about licensing. BizSpark helps a lot, and you can easily use MySql and the like if you don't want to pay for MS SQL Server. SQL Azure might relieve some of the pricing pressure for SQL too.
Speaking of Azure, I really like it. It's completely intuitive if you've done any ASP.NET development. The only problem is that it is trailing about 6-12 months behind the other cloud app hosting options in features. And since SQL Azure was announced there's been far less emphasis on Azure Table Storage which is a shame. The querying capabilities for table storage are ABYSMAL. It needs to be on par with SimpleDB before I could see myself using it (and I really, really want to use it).
I was contemplating jumping ship before ASP.NET MVC came out since the previous versions of ASP.NET (known as Webforms) really hindered more complex web app development. MVC is awesome though. Almost every part of the framework is swappable so if you have a specific need, you can quickly swap in your own or an OSS implementation. ASP.NET MVC + Spark View Engine is pure hotness.
There's been a shift going on inside Microsoft's DevDiv too. ASP.NET MVC is open source, jQuery is shipped with Visual Studio, MS employees are active bloggers and twitterers, etc. And speaking of Twitter, there's a strong .NET community there.
This is particularly dangerous since MS has shown in the past they don't have a problem totally killing off platforms (look what happened in SDS before it even launched).
2. HTML encoded by default
3. <li each="var foo in bar">${foo.Name}</li>
4. Client-side rendering: http://odetocode.com/blogs/scott/archive/2009/03/12/client-r...
5. conditional attributes (kludgy syntax, but better than Webforms VE)
6. Content "spooling". You can create a partial that registers a script tag/script specific to the partial, but only if it hasn't been registered already. Partials can also spool content to any other content area.
6a. <script once="jquery" src="...">
The video quality of this is a little rough, but here's a good overview by the guy that started Spark, Louis Dejardin: http://whereslou.com/2009/06/24/spark-releas-and-presentatio...
http://www.zope.org/Documentation/Books/ZopeBook/2_6Edition/...
I know C# better than any other language so it was an obvious choice for me. The option of at least offering a Linux product on Mono has crossed my mind a few times so that's a plus too.
Mostly the fact that I can "query" any data source (an array, a file, etc) in the same way I query a table. Lovely.
DataMapper is great. Doctrine is good. Linq is my fave.
I really enjoy the tooling, the documentation, and the libraries. There is exactly one time we've run into an issue with the built in libraries: The documentation didn't match what really happened, but Microsoft confirmed this for free and fixed the docs.
Visual Studio is rather polarizing. I'd say the developers at Loopt fall into three camps: Like it, miss Eclipse, and still use Emacs. You don't have to use it. I can't type more than 20 WPM; Intellisense is so good I haven't felt the need to learn to type properly.
Cost wise it's really not a problem. There's BizSpark which is great for three years, but after that you can get into their SPLA licensing program. It lets you pay per month for only what you use. See http://www.microsoft.com/serviceproviders/licensing/default....
Loopt started before BizSpark, so we know what it costs. Here's a hint: we paid more per month for Office and Exchange than we did for the software running our service. All in it was less than hiring an intern.
If you're super strapped for cash, just get an MSDN subscription, and use it for everything. It's not legal, but Microsoft knows there's more money to be made by letting you become successful (at which point you pay real money for licensing) than there is in going after your initial few counts of infringement.
If you like getting into the guts of things you might not like .Net, even though they've released many of the symbols and source code. See http://msdn.microsoft.com/en-us/library/b8ttk8zy(lightweight... That said, you can't edit it. I'm ok with this though because in my experience it just works. I'd rather focus on developing my product than fixing the tools and runtime I'm using.
The open source .Net scene is a little touch and go. There are some great projects out there, many by Microsoft itself, but it's simply not as robust and vibrant as the Linux open source community. This seems to be changing (slowly), but as I said earlier we've found compelling reasons to have a mixed environment.
I'm not saying it's all good. ASP.Net is a valiant attempt to solve the wrong problem. Drag and Drop doesn't work for the web. .Net MVC is much better. Despite that, our website is written in Rails, and communicates with our back-end using Thrift. It works great.
If you want to know more let me know. Maybe I'll write a blog post.
Edit: Grammar
- Risking Microsoft's ire by breaking the licenses
- Locking yourself in to long-term Microsoft licensing
This was a pattern I saw in many, many startups in the 90s -- organizations that relied first on UNIX systems for their servers (often times Solaris, Linux, and in the early-to-mid 90s, IRIX), would eventually adopt Microsoft products for some component of their infrastructure -- usually Exchange, based on heavy political pushes from the growing number of business users.
This always to the same effect -- the Microsoft products spread due to lock-in network effects until it began to near 100% MS on the desktop, e-mail, and corporate servers. Production servers generally remained UNIX due to the costs involved. The IT organization shifted to MS, costs went up as the staffing necessarily increased to support administering Windows systems, expensive political maneuvering arose between engineering and business as they both fought to move to systems that they thought better aligned with business priorities.
At the end of the day, everyone wound up locked to MS with heavily entrenched political interests in keeping the status quo (and the job security that went with it).
Having seen the fall-out from migrations to MS products in corporate IT, I can't imagine voluntarily implementing MS as the primary production and developer system at a startup. You're locking yourself out of alternative solutions, you're signing up for significantly increased capital and operational expenditures, and perhaps most importantly, you're excluding an entire class of software engineers that primarily work on UNIX-based (including Mac OS X and Linux) systems.
There's tremendous value if it's what you already know. The opportunity cost associated with learning and becoming reasonably competent in a different development stack can be non-trivial, especially if you had 5 to 10 years of experience with a different stack or set of technologies. I've yet to encounter a new framework/API/stack/language which I've been able to move to in only a few weeks and be 80%+ as productive as what I already knew, as there's often lots of useful minutiae, subtleties and new concepts you need to acclimatize over time.
>You're excluding an entire class of software engineers that primarily work on UNIX-based (including Mac OS X and Linux) systems.
The converse also applies. Granted, a competent engineer can learn anything, however you're initially inclined to pass on them because of the associated ramp-up time.
If all you know is Microsoft, then that's all you can do, and the organization's technical direction will almost certainly model that initial decision.
However, if you have the luxury of choice -- and I think most people do -- I'd advise against it. The short and long-term costs of Microsoft lock-in are decidedly "non-trivial". The considerable investment in staff, licenses, and technologies make it very difficult to migrate to a non-Microsoft stack in the future.
The converse also applies. Granted, a competent engineer can learn anything, however you're initially inclined to pass on them because of the associated ramp-up time.
In my experience, you're entirely locking out engineers who use Linux and Mac OS X when you move to a Microsoft stack, but you're not necessarily locking out Windows engineers when you use a cross-platform stack.
I speak from personal experience -- I know the .NET stack very well, and am very familiar with the LAMP stack from my days with PHP, but I've only dabbled in Ruby and Python. Before starting work on our startup, I considered writing it in Rails, but the reality is that it would have taken me much longer to get to market. We hit the market at a very opportune time, and were able to ride a nice wave of buzz to sustainability. If we'd waited another couple of months (my guess at the added overhead of learning Rails), we may have missed our window.
The same would naturally apply to someone who already was well-versed in, say, Rails -- it would make no sense to learn .NET unless there was a very compelling reason (say, you were writing a Windows desktop app).
Not to mention that the first app you write using any technology you don't understand well is going to suck. If I'm going to base my livelihood on a product, I want it to be as well-written as possible, so I'm going to go with what I'm best at.
Also, it's absolutely not true that you're locking out Linux and Mac developers when you write .NET. Some of the best .NET developers I know use a MacBook Pro as their primary machine.
Sometimes price isn't the only factor when deciding which tools are best for the job. After using C#, switching back to Java or a non-compiled language would be very painful. I'd lose a lot and miss it.
What do you use for desktops, e-mail, corporate servers, etc?
My goal was elucidate the hefty cost of Windows lock-in and Window's demonstrably strong tendency to pervade your organization in terms of political, staffing, and technological investment.
If it's considered the best tool for the job and these recognized costs are acceptable, then there's no reason to not use it -- but history should also provide some insight into how Microsoft behaves with organizations that become reliant on their technology.
Microsoft isn't being magnanimous with Bizspark -- they're trying to ensure that more startups are locked into their technology stack early, when the perceived immediate advantages may seem most attractive.
VMware is not an adequate solution for day-to-day development. There's a performance impact, significant memory overhead, and running Windows inside Linux/Mac OS X is not any more accommodating to users familiar with a UNIX environment than simply running Windows natively.
I have been writing .NET code exclusively in VMware for 5 years now, 2 years on Windows and 3 years on a Mac. The performance impact is there but it isn't a deal-breaker (especially if you have a fast HDD and enough memory (6GB)).
Confirming neither is a drop-in replacement, which isn't to say that Mono isn't useful or valuable, but simply that it is necessarily 2nd-tier in a Windows-focused development environment.
Given that you've spend 3 years developing C# in VMware on a Mac, I'd ask why you're not using native Mono/MonoDevelop entirely?
I have been writing .NET code exclusively in VMware for 5 years now, 2 years on Windows and 3 years on a Mac. The performance impact is there but it isn't a deal-breaker (especially if you have a fast HDD and enough memory (6GB)).
What do you see as the significant advantage of .NET that has warranted using non-native tools with an admitted performance impact exclusively for 3 years?
This sounds like individual developer inertia, not a high-level business decision.
In the last 6 months I have been investigating moving my application to Mono and running on Linux servers, when that happens I will use Mono and MonoDevelop exclusively for that project. Of course other client projects I will still use VMWare and Visual Studio since they will be deployed on Windows.
The reason I didn't look at Mono earlier is that I had dedicated Windows boxes and didn't see a benefit in moving, now that I am on EC2 there is a price motivation to move.
Running in a VM is not a big deal, most .NET developers I know do all their client development in a VM even if they are running Windows.
That's a very bad argument. If you're super strapped for cash what are you doing with closed source software anyway ?
It's not like there aren't plenty of free alternatives.
I think that mostly boils down to what your experience is.
But the GGP effectively encourages people to break both their license agreements and the law.
If you're strapped for cash you also don't get a free ride to go and steal stuff.
If the price of being strapped for cash would be to have to learn how to set up open source software then so be it. But given the fact that you could easily make a business turning over a couple of 100 grand annually using just apache and (forgive me) PHP means that if you choose to abuse an MSDN subscription to be able to use closed source proprietary stuff that you are doing so deliberately.
I really have a problem with effectively encouraging people to break the law in a public forum, even if the manufacturer of the product has some hidden gain from this behaviour.
I tried for a few days to use Visual Studio to link some C++ code to a DLL that was supplied by a third-party vendor. I never got it to work. This was literally "hello world" calling a function in a DLL, and I couldn't get it to compile.
After much googling and asking around, I gave up and used gcc instead. It worked the first time.
As someone who has always used free software, I always assume that if someone gives me very expensive proprietary software, it will be problem-free. It hasn't happened yet.
This is why you haven't success with MS stuff. For people who have always use MS stuff, it's faster and easier to stick on the stack.
Learn more at http://www.microsoftstartupzone.com/BizSpark/Pages/At_a_Glan...
Microsoft has been a great partner and deserves better from me. If I could still edit my original post I would. For the record, Loopt was and is always compliant with all commercial and OSS licensing.
> # Enrollment is free, just pay a USD $100 program offering fee at program exit.
That's like saying 'Free car, pay in 36 months'.
Well suited to a startup, whom are usually looking for minimal outlay from the get go.
I must disagree with this one point. I cannot see a justification for taking software illicitly like that. In my opinion, VS and SQL Server are well worth their price, but if you disagree the answer is to use open source software not take it like that.
Aside from that, I think you have some very good points. I am primarily a SQL Server DBA and I think that for many applications it is the best RDBMS available. I am also a fan of Visual Studio for certain types of projects (though I also make use of Python).
I really like .NET the language (the CLR, the libraries) and really could care less for all the overblown frameworks, like ASP.NET or WPF
But I decided to go html/jsonp/F#/SQL-server this time. It lets me completely separate my layout from my business layer, and with a few bits of auto-code-gen my data access layer is always up-to-date. So it really just leaves me a few hundred lines of F# code to write a full-fledged application. I'm also looking to move to mono and perhaps MySQL, and this architecture is probably the most amenable to that move further down the road. Plus by writing functional code in F# with immutable data structures my code is automatically parallelizable and ready for moving up to super performance using things like MPI if I want :)
I could have done the same thing with CGI, and I anticipate it being easy to move to other platforms without much or any rework.
C# 4.0 HAML LESS MongoDB
we (http://friendsell.com) are running c#/bistro/ndjango/couchdb
If you are going for a small project, choose PHP, pyhton or Ruby. If you are building a giant app for a company, you'd better give Asp.net and SQL server a try.
For me, I use .NET to develop Windows application. Right now, I made a barcode software, no more and I should say it's the most advanced and coolest IDE + language (C#)I have ever used. For speed, it's Okay.
.NET speed is about the same as Java, it all depends on your architecture.
I should say this points to inexperience first. ASP.NET has taken huge steps in making development/deployment simple.
there are subtle implications to the platform decisions you make, but on the scale of most startups, the extra $40 bucks a month you'll pay for that windows ec2 ami isn't gonna make a difference. what will make a difference is you building on top of the stack you're most proficient with. for us that meant building on .net and couch. for someone else that'll be RoR and mysql.
I've been developing w/ Microsoft since the VB6 days and I will be the first to admit that there's no right answer to which is the "best stack". There's great advantages to LAMP, Java, and Microsoft. I think it all depends on where your comfortability lies. My honest opinion is that trying to answer this question is like to trying to figure out which is the "best religion". It's totally subjective.
Affordit.com, which received funding a few months ago from the Founders Fund, is also built on .NET.