Mediaphormedia, the people who make Django, fired 10 employees yesterday and pulled the plug on Ellington, a CMS geared toward news organizations.

In a Django vs. Rails thread on Quora, Naveen Kumar wrote: "If your motivation behind learning either framework is job or employability then go with Rails. On the other hand if your focus is more towards development friendliness, language design or readability go with Django or web2py. Django is Rails without magic."

In the Rails vs. Django debate, is Rails "winning?" What do news hackers use nowadays? What do you tell people getting started?

(I'm inviting a group of news apps developers to answer this question. I would like to use your responses in a Nieman Lab story.)

I don't really like the adversarial tone of this question, to be honest. I don't really see myself opposed to Django in any real capacity, and I think most of Rails' (and Django's) gains have come vs. PHP and other simplistic web-programming approaches.

Wasn't going for adversarial. I'm wondering which framework has the momentum and whether there are concerns about the fate of Mediaphormedia. Perhaps I shouldn't frame this as X vs. Y so much as "Thoughts on the latest with Django?" But I'm such a journalist.

I have to agree with Jacob with regarding to the way you're framing the question.

That said, I think the larger issue, really, is not "which framework is winning" but rather, are there enough knowledgeable developers interested in working in the news industry (regardless of framework) for this to even be a question worth worrying about?

Find the right tool for the job, and find the right team to make stuff happen.

(Not necessarily in that order.)

But I suppose it might be useful to explain why the NYT Interactive Newsroom Technologies team chose Rails instead of Django. For starters, even three years ago there was a huge Ruby/Rails user group presence here, a trend that has been accelerating with all the startups. We also had some Rails talent in-house already (myself included). And we liked the magic, dammit. Database reflection is pretty useful when you are handed a database like the Wikileaks war logs and you want to get it online.

I don't know about Django in general, but here in NYC there are a ton of Rails developers and if you were a NYC news organization, you'd want to pick Rails so you could more easily hire for positions or even find consultants in a pitch.

One random question: I seem to recall that Django was either formally or informally branded as the web framework for newsrooms. Was I hallucinating that?

If you're debating Rails vs. Django, why not CodeIgniter vs. Backbone.js?

The tools don't necessarily matter, in my humble opinion. I picked Django because it's what the St. Pete Times folks used. I have created a CRUD application on my laptop with Rails. It's fun and neat. I just know how to set up servers for Django and have worked with it more, thus I have a bias toward it.

But great work can be done in anything, even Java!

@Jacob Nope. Not at all Django was invited by a team at the Lawrence (Kansas) Journal-World that included Adrian Holovaty, Simon Willison, Jacob Kaplan-Moss and Wilson Miner. FWIW, you might find this interesting: quora.com

Hey this is a safe space to ask the wrong question. We're all journalists here :)

I'm also not morally opposed to Django. I like Ruby more than I like Python, but it's not like I'd quit news and become a potato farmer if somebody forced me to switch to Django.

But ignoring that. I think no framework is winning right now. We use Rails, but to be frank, Rails is going through a tough period, with some decisions being made by the Rails core team that a lot of us down here aren't thrilled with. Also, there've been some SQL injection bugs in the new ActiveRecord that have been staling time away from things we'd rather be doing, like making new stuff, so this isn't quite Rails shining moment.

If anything, I think a lot of us are kind of annoyed with our frameworks and experimenting more with lighter frameworks like Sinatra, Padrino, Flask, etc. And there's a lot of love for JavaScript and Backbone, too.

I've fought with both Django and Rails and they are both imperfect, I'll second that Rails is going through a rough batch, but from afar it looks like Django is also having some issues.

But that is really the thing, innit? A mantra of developers the world over is that your framework only gets you so far, but in the end that is pretty damn far.

Two other things I'd quibble with is all frameworks are magic. No one writes int main(int argc, char **argv) in web apps, there are layers upon layers of things that are just opaque.

Also, development friendliness is another canard for obvious reasons.

In the end it doesn't really matter, we're just slinging big strings down the wire. At ProPublica we chose rails because we wanted to share code with the NYTimes.

One framework is not winning, but many frameworks are - or the idea of using frameworks to better organize code and make it more readable by others. Consider the practices that both Django and Rails have helped to popularize and/or reinforce among web developers: Don't Repeat Yourself/modularization, better control and design of URLs, a wide variety of templating systems, among others.

The fact that we've now got frameworks ranging from the basic to the feature-rich in many languages is a testament to a "battle" that had many winners. If developers have learned and adapted these lessons, then it doesn't much matter which one is in favor at the moment.

In addition, the communities around Django and Rails are different, and in some respects have differing visions (although they also agree on quite a bit). Competition is how we tend to view these kinds of situations, but I'd argue that we should focus on the ecosystem as a whole, and the improvements we've seen among users.

Also, please don't conflate Ellington with Django - Ellington is built atop Django, and if anyone wanted to, he/she could invest the time and effort into building a new Ellington because, as others have noted, Django provides the framework for doing so. Even if MPM doesn't license another copy of Ellington, there are other Django-based publishing systems and there will be more.

It's odd to even think of Rails and Django as competitive. We have a common enemy: The deadline. And I compete against proprietary Java CMSes and ancient PHP/Perl scripts, rather than against Rails hackers.

Now, as for "thoughts on the latest in Django."

The Post is a crazy old-school Django shop. Yet our CMS is in Java. Why? Because our nimble apps team doesn't run our bulky CMS for the same reason why an Army reconnaissance team doesn't build bridges.

MPM and Ellington were already out of scope for what we do, and I suspect that this isn't a unique point of view. We'll keep building our micro-CMSes in Django and leave the core CMS to the guys who enjoy things like:

public static void main(String args[])

This probably says more about MPM and the state of the newspaper CMS market than anything about the underlying tech. Java has a big footprint, but I suspect that's inertia more than anything else.

I got started with Rails because I was embarking on a project at TPM (PollTracker) that was just too complex the blogging software we were using. When I realized I needed to learn a web framework, I chose Rails for two primary reasons: There were a ton of resources out there (tutorials, screencasts, StackOverflow answers, etc.), and I liked the stuff built in it: Basecamp, Github, Twitter, Slicehost. At that point, I didn't even know that NYT and ProPublica were using Rails for news. I wrote a blog post back then about my first impressions of Rails, and it still holds up pretty well:

shhhaw.com

Here's a possibly related question: All of us in here build news apps, mostly outside our main publishing systems. Would you rather spend your days hacking on a newer, better CMS (whether that's Armstrong or Baroque or Vox Media's new toy)? Or do you see an ongoing separation between daily publishing and app building? Or is the future more bespoke mini-CMS things like PolitiFact and Homicide Watch?

That's a good question. While I do want a better CMS from time to time, as a developer who is also a reporter, I want to spend more time building stuff that tells a story and not stuff that publishes stories.

Plus I'm skeptical we'll ever have a CMS to rule them all, and our current one (Expression Engine) does just fine putting words on a page.

All I want is a CMS that has a decent API. I don't want to hack on a newer, better CMS; I want a CMS that does exactly what it needs to and gets out of my way.

Building a CMS doesn't excite me. Having a CMS I can modify to my heart's content does. Since such a CMS that suited our needs didn't exist back in 2009, we built one ourselves, and hopefully the next team with similar desires will find Armstrong to be a useful starting point.

There is no idea that comes up in our tech office or in the newsroom that is infeasible because our CMS can't do something. This is an incredibly liberating property, and we've barely scratched the surface of the possibilities that opens up. A CMS with an excellent API is almost as good, but a team of news app developers and a team of CMS developers (with or without overlap) who are comfortable in each others' worlds is the ideal that news orgs should be aiming for.

With that said, the API approach to integrating with a CMS is clearly the most sustainable one. The technology du jour tends to change rapidly, and if you have a rich API, that's less of a problem. Our approach works for us because we were starting something new. For those with legacy concerns, a solid API is probably almost always the best approach to ending CMS/news app apartheid.

Speaking as a developer at a newspaper that received its share of Ellington sales cold calls, I think Ellington's No. 1 competition was the excellently documented, free, open-source Django and not Rails, which seem to be the premise of the question at the top of the page.

Had there been a paid turnkey Ruby on Rails newspaper app (to my knowledge, there isn't), I think it would've struggled against the free, open-source RoR as well.

Also your headline: "The Lawrence Journal-World gets out of the CMS business, losing out to freebies like WordPress"; it's an apples-oranges comparison (that's been made many, many times). WordPress idn't a CMS. Once you try to do much more than blog with it, it gets very cumbersome, very fast.

WordPress may not be a CMS in the same sense, but a lot of news organizations (including my previous) are hacking it up that way.