Regarding development of a conferrence session on choosing a web framework
Ryan Sullivan
My name is Ryan and I'm writing today because I'm planning a session at an upcoming conference regarding how to select a web framework. The session will focus on how a development team (or members thereof, or project leader, etc.) might go about choosing a web development framework.
My motivation for proposing and coordinating this session is that in ten years in IT I've yet to observe a group go about the process of choosing a web framework in a scientific way. Instead, emotions and politics generally drive the decision: the development team prefers rails, the business team heard that Django cures cancer, management has a cozy relationship with Oracle and chooses their proprietary stack, etc. I decided it’s be nice to find or develop a session that breaks away from these approaches and offers a method for choosing a framework/stack based on evidence and best practices.
Scouring the internet for discussion on this topic, I found there isn't a whole lot of good information. Your group's efforts stood out as you are pursuing a structured approach to framework comparison. I'd like to ask for your assistance/guidance on developing this session. I have a number of specific questions but first would just like to say hello and ask if you’re interested and have some time to discuss this with me. For some perspective, my background has been ecommerce web development with URBN, Inc. and then general web development for the last year at WRDS. The conference is the Wharton Web Conference.
I hope this email finds you all well look forward to hearing from you. Thanks in advance for your time and any assistance you can offer.
Ryan J. Sullivan | O:215.573.5089 | whartonwrds.com
Ryan Sullivan
Brian Hauer
Thanks for reaching out! The project you are describing sounds quite interesting. I agree, selection of web application frameworks is oftentimes capricious and not driven by science. Of course, it's hard to blame people since there is not a whole lot of evidence to leverage outside of popular consensus, conventional wisdom, and the ever-changing winds of opinion.
A previous draft of the Github landing page for our project in fact lamented the lack of data to use in decision-making as an impetus for our project. I've since revised the wording to remove that, but I retain the sentiment.
We'd be happy to explore how we can help with your project. As you might expect, beyond the raw data and lightweight editorializing we've added to our blog, we also have scores of opinions and anecdotes. Chief among those is one I've cited before: production deployment is an all-too-often overlooked subject in documentation and best-practice references. I'd like to see that improve across the board.
Incidentally, an area we repeatedly circle back to in our conversations here is measurement of developer efficiency. We consider efficiency the second key dimension. Pairing efficiency with performance would allow a two-dimensional graph that we think would be quite useful. But of course efficiency is an incredibly prickly and delicate subject.
I'm looking forward to hearing more about your project. Let us know what you had in mind for help.
I can't speak for the others from the developer community who read and post here, but I will say they are really bright people. Perhaps some of the other folks here, especially those who work on web frameworks, would be able to help as well.
Ryan Sullivan
Thank you for your response! It's been a heck of a day and I'm just catching up. I'll get back to you with a little more information. I look forward to discussing this with you!
Thanks.
Ryan Sullivan
Hi Brian,
Thanks again for getting back to me. To jump right in, I’m hoping for this session to provide project leaders with a method for how to objectively compare frameworks and chose the best one to fit their project. Ideally, I’ll find an expert who can give a presentation on a method for how this selection can be done and some examples of how the method has been successful in past projects.
Of course, it may not be the case that there is a concrete method or an expert that can speak authoritatively. However, it seems likely that an expert in this field exists who can speak to the considerations that are most important and how project leads may go about addressing them. What I’ve experienced and read on this subject so far indicates that most important in choosing a framework seem to be compatibility of the framework’s approach with the intended project, compatibility with the intended production environment, developer experience and anticipated learning curve (which is very much related to developer efficiency), the framework’s community support and activity, and a number of technical concerns such as type of persistence, component or request based, RIA or traditional, etc. What I haven’t yet found is any advice on a defined approach that can be followed to analyze the available frameworks and determine based on various considerations which is the most appropriate for a real world project.
Having done a bit of research on the subject and knowing that I’m not the authority that I would like to present on this subject, my current goal is to locate an individual or panel of experts who can present. I obviously have some ideas for what might be addressed in such a presentation, but I’m hoping that whoever leads the session will have an idea of what they’d like to present. Taking a look at your web framework benchmarks project and based on what you’ve mentioned about the importance of comparing production deployment and developer efficiency, it would seem that you/your group is addressing the same core problem that I’d like to see addressed in this session.
With that said if you have any thoughts on how this session might be lead, I’d be very happy to hear and discuss them. I’m also interested to know if you or anyone in your group has presented on this subject in the past. As you might imagine, you are among the few experts that I’ve spoken with about this; I’m curious to know if you’d be interested discussing an opportunity to lead this session.
Thanks again for your time and once more, my apologies for the late reply. I look forward to hearing your thoughts. Also, if you prefer, feel free to message me offline at rsull-at-wharton.upenn.edu.
Brian Hauer
You are definitely putting together something very interesting. I'd like to discuss this more next week after we've had a chance to post the Round 5 results. I will follow up via e-mail.