Arguments *against* using open source frameworks?

Question

I recently have a client comes asking me as a C#/.NET dev for reasons about why I have not picked any of the "wave of the future" frameworks out there like Rails/Django and instead choose "proprietary software" from M$ to build his website.

Frankly, I love all the open source stuff. But seeing non-technical people bugging me for reasons why I choose the tools I always choose just because it's "proprietary" and "locked in" is just putting unnecessary stress on me.

I have since stated my rationales to the client but end up w/ Python/Django anyway for the sake of the company. Lucky me that I can do some Python coding so it's a quick run of the mill for me. Nothing spectacular.

The tools and documentation is the biggest plus to me when using "proprietary software". For me, there's nothing like doing C# on Visual Studio especially with now C# becoming more functional. It's been so much fun since LINQ support was out.

But I have to wonder...

What are some arguments against using OSS frameworks?

From the your own perspective, from the company/business/startup perspective, from whatever angles...

Answer 1

For me, the topmost reason would be

I am not as familiar with those frameworks as I am with the .NET stack

I too have had such a request for a single project, where the person that originally came to me asking me to build his website for him, then later wondered why I had picked Microsoft ASP (which was what I used at the time) instead of php and apache.

Basically my response was that if you ask me to build the website the best way I can, then the tools I know are the ones that are going to be used.

Answer 2

Often times, licensing can be more of an issue with open-source software than with closed source, since proprietary software usually follows a simple licensing model.

For example: Some open source licenses require all derivative works to be entirely open source as well. This has become an issue for the LAME encoder, for example. You cannot link your copy of the LAME encoder into your executable, but must instead have your users download their own copy, separate from the product itself.

You'll often hear Jeff Atwood refer to programmers as 'Magpies' because we like to collect and use shiny things. (New frameworks and etc.)

For a real-world application, using a new or open source framework may actually be a detriment because you are less familiar with it, or it is not a mature.

Being pragmatic should be your guiding force.

Answer 3

First off, let me say that there are a lot of great OSS frameworks out there, and I don't want to come across as spreading FUD or anything. But from a business perspective, there are a lot of questions that need to be answered when using an OSS framework. Frankly, I believe that these are much the same questions that need to be asked of any software, but with some OSS software (maybe 80% of what's on SourceForge?) the answers to these questions aren't good:

• Who is going to support this software? Will we be left sifting through the source code to nail down a bug? If we do fix a bug, is the author willing to accept it or fix it in the next release?

• Will this software be kept current and secure 5 years from now? With a major vendor, they make contractual pledges for predictable support. Will this be available with this product?

• What licensing burdens are being placed upon us by using this software? If we go with a GPL version of a framework, will we have to make our product open source as well?

• Does this software have a large enough user base to be considered to be mature?

• Does the author have a rigorous testing and QA process?

• Will we have to build this product from source? If so, can we trust that the posted source has not been tampered with? If we're getting the binaries, can we trust who actually built the thing and posted it?

• Will I have someone to call on the phone if it all goes south? Will I have someone to blame who I can point a finger at who is not me?

This level of questioning typically does not get asked with major vendors -- the answers are pretty much taken for granted (rightly or wrongly). That means that OSS evaluation does require a bit of legwork in order to really understand what you're getting.

EDIT: A commenter pointed out that the second question I asked is very relevant to companies that go out of business. Very true. And in my personal opinion in a contest between OSS vs and a product from a small or unstable vendor, the OSS product wins every time, everything else being equal...

Answer 4

The main reason I've heard for going with a vendor (with most vendors being closed) would be support. If something breaks in your widget you can get support and maintenance from the vendor.

Answer 5

All software is broken and buggy. With Free software, you can fix the bugs yourself. With closed-source software, you can't do anything but wait for someone that already has your money to fix it for you.

I would never consider a closed-source application framework unless it had a proven track record of being amazingly good. (Qt from many years ago would have met this requirement.)