Welcome!

Web 2.0 Authors: Elizabeth White, Dana Gardner, Roger Strukhoff, Keith Cawley, Harald Zeitlhofer

Related Topics: ColdFusion

ColdFusion: Article

i-Technology Viewpoint: We Need Not More Frameworks, But Better Programmers

"I view poor developers as a symptom of frameworks"

It's no secret that I've been outspoken about not liking frameworks for quite some time now. The truth is, I believe that frameworks have a lot to offer. The most significant benefit that organizations stand to gain from using frameworks is a standardized way to code and an environment that is generally more conducive to allowing multiple developers to work on a project at the same time. If frameworks help to standardize how things are done and make it easier for many developers to work on a project, why have I been vocal about not liking them? Am I just trying to create controversy?

The truth is that by and large, yes, I am trying to cause controversy...but only because I want to make people think - a statement that I make for a very good reason. I don't like frameworks because I've met a lot of developers over the years who have maybe one or two years of experience building applications with a framework but who don't have a good grasp of the CFML programming fundamentals you'd expect from someone with a few years' experience. I've met many developers who aren't capable of developing anything outside the context of the framework they're used to. In these scenarios, the framework is a crutch.

I view poor developers as a symptom of frameworks. Is this the fault of frameworks? No, not at all. There's nothing about a framework that says you can't use it and still be a terrific programmer. It's the fault of the developers, for sure, and it is for that reason that I don't truly dislike frameworks. I don't think any one framework should be used to build every application, and I personally prefer to develop without them, but when it's appropriate I see nothing wrong with using them. If developers are really good at what they do, they should feel comfortable and be competent working without a framework and/or bouncing in and out of several frameworks. This is unfortunately not the case for the majority, though not all, of the people who use frameworks. It is especially true for developers who began their programming career working on projects that are built with frameworks and have never had the need to build an application without one.

You might be surprised to hear that I have a similar dislike of object-oriented programming and an even stronger distaste for design patterns. That's an extremely controversial statement to make and I'm sure that if my commentary on frameworks didn't upset or confuse the majority of our readers, talking poorly about OOP and design patterns will. Now that I've managed to get everyone worked up, let's talk about it.

What could I not like about object-oriented programming? Similar to frameworks, I actually do like OOP, but I don't like what it's done to ColdFusion developers. This is truer for design patterns, which I view as the buzzword of the day right now. The reasons behind these statements are simple to explain, but very difficult to do anything about. I spend a lot of my "free time" reading about object theory and various thoughts and approaches to software architecture. The vast majority of the reading is not specific to ColdFusion but is certainly applicable. Likewise, my statements about OOP and design patterns are not specific to CF but apply to developers in all environments. An overwhelming number of programmers, not just CF but Java, .NET, C++, etc., talk about object-oriented programming without really understanding it at all. The roots of OOP are grounded in ideas, radical at the time they came out, about how to think about software development. Most of the books, courses, and, therefore, developers today have completely missed the point when it comes to OOP. These misconceptions and misapplied ideas are magnified greatly when developers begin throwing around talk about design patterns, which is why I said that I don't like them. The truth is, I think design patterns are an excellent learning tool and I am an avid supporter of OOP...but only when they're approached the right way.

My definition of "the right way" is for developers to learn "Object Think." Object Think is an approach to thinking about the objects that make up an application and is used for object modeling solutions. The original OOP concepts were based on this line of thought but it was quickly lost among developers, as they tend to use sequential thought when formulating solutions. Sequential thought comes more naturally to most people with a background in traditional programming and, to be honest, tends to be a more natural line of thought when formulating machine solutions, as programmers do. I have spent the better part of my free time for the past five or six years studying the writings and interpretations of the architect Christopher Alexander, software architecture pioneer Dr. David Parnas, Kent Beck (best known as the founding father of XP), and a handful of other early thinkers in the realm of object thought and behavioral programming. To apply these philosophies and thoughts to ColdFusion development requires changing the way we plan and execute solutions. At the heart of this change is the need to be more object-centric, achievable via object-personification, in our approach to planning applications.

Describing a philosophy of how to model applications is obviously beyond the scope of this editorial. It's beyond the scope of an article and, in all fairness, is even difficult to achieve within the limitations of a book. My Sensible Assembly Methodology is based on these teachings, taken from Streamlined Object Modeling, Object Think, other thoughts and writings, and my architecture and development experience in the real world. Due to an extremely hectic schedule, I should have some documentation on that by the end of March at the latest, with a book to follow. My advocating a methodology (SAM) as opposed to a framework is driven by a personal belief that by doing so, I can offer developers something more valuable and transferable - a better way to think about development.

What is important to me, for the good of the community, is not whether developers choose to use one framework or another or choose to use my methodology or another; it is that we not only accept that the way we solve problems and develop software could be better, but examine every alternative approach in our attempts to find better ways to work. This idea has become more important with the recent buzz about Web 2.0 and next-generation Web applications. Simply throwing a Flex or AJAX front end on our application isn't good enough - in order to develop a true next-generation interface we have to change the way we think about our users and the experiences they should have when they interact with what we create. This is something that the XD (extreme design) group at Macromedia has been talking about for several years. As applied to software architecture, these ideas have been discussed for several decades. My hope is that now, with the vast amount of available information and the mature tools and technologies available to developers, the ColdFusion community can begin to put these ideas into practice.

More Stories By Simon Horwith

Simon Horwith is the CIO at AboutWeb, LLC, a Washington, DC based company specializing in staff augmentation, consulting, and training. Simon is a Macromedia Certified Master Instructor and is a member of Team Macromedia. He has been using ColdFusion since version 1.5 and specializes in ColdFusion application architecture, including architecting applications that integrate with Java, Flash, Flex, and a myriad of other technologies. In addition to presenting at CFUGs and conferences around the world, he has also been a contributing author of several books and technical papers.

Comments (14) View Comments

Share your thoughts on this story.

Add your comment
You must be signed in to add a comment. Sign-in | Register

In accordance with our Comment Policy, we encourage comments that are on topic, relevant and to-the-point. We will remove comments that include profanity, personal attacks, racial slurs, threats of violence, or other inappropriate material that violates our Terms and Conditions, and will block users who make repeated violations. We ask all readers to expect diversity of opinion and to treat one another with dignity and respect.


Most Recent Comments
Dennis Muzza 12/11/06 08:07:30 PM EST

You shouldn't lose sight of the ultimate goal, which is producing good, robust, scalable systems. If this can be successfully accomplished by cheap, inexperienced programmers relying on a framework that does pretty much everything for them, more power to the business that employs them.

The problem I see with frameworks is that few of them are really good and widely usable, and those that are still call for a lot of custom coding which requires good developers. Oftentimes companies spend millions to develop frameworks loaded with every possible feature, and then it becomes the proverbial hammer to which everything looks like a nail. People begin using it because it's there and a lot was invested in creating it, not because it really solves a problem.

Serge Bureau 04/11/06 10:43:47 AM EDT

Sorry but I feel the opposite.

Framework is the reason for bad programmers.
EJB's and Struts for example are awful (I have not look at EJB 3)

Companies are pushing people to use those framework that are big and inflexible.

So while thre certainly is bad programmer's, I have still to see a good framework ?

jim 04/04/06 04:00:21 PM EDT

Poor developers using frameworks as a crutch is a symptom of them being poor developers in the first place. IMO there should be licensing for software engineers just like every other engineering profession. There are way to many script kiddies out there who have learned one framework and call themselves developers - There should be a bare minimum set of abilities you can count on a develper having, and so many don't. I can't tell you the number of "developers" I've dealt with who have degrees in philosophy or english or phys ed, who have no idea what a data structure is, but have mastered basic cold fusion, so they're now "senior programmers". Until we set a baseline bar that everyone has to hurdle to call themselves a software engineer, our industry will continue to be plauged by these frauds.

Ananthalakshmi 04/03/06 03:30:20 PM EDT

Hi, Your view may be true in the context of ColdFusion Development. But, in real application development scenarios in languages like Java, .NET, etc and in building enterprise applications where the problem solving methodoligies and the implementation needs reusable and prooved strategies for recurring problems(for ex, page iteration, adopter, MVC architectural patterns, etc) instead of developing the solution from scratch for each application that is being developed in the organization, instead to reuse existing components, build reusable services, adopt proven strategies(design patterns, etc) will certainly improve the productivity of the developer and also to build quality, reliable and scalable applications which is the most need of every application that we develop.

Yes, it is the responsibility of the developer to understand the underlying technology, logic that the framework uses and on which it is built on before started using it or understanding it while working on the project based on the framework.

In real world, frameworks make life simple and more simply makes application development more easier instead of building it from scratch.

Thanks.

Jos Warmer 04/03/06 06:45:54 AM EDT

You want controverse? Ok, I will help :-). I think you are absolutely wrong. The reality is that about 10% of the programmers is good, 10% is bad and the 80% remaining is somewhere in between (and do not have it in them to become good programmers). As with many professions my guess is that this will have a normal distribution.

Since the majority of the programmer is not excellent and never will be, we need ways of making mediocre programmers productive. Using a framework is one approach (although I have seen frameworks that are too complex for this). The main issue is that these programmers need less freedom and more prefabicated stuff to work with. This might mean using frameworks, using existing components, using MDA or DSL's, etc. etc. The challenge for those that are in the 10% good group is to find ways to help make the majority of less good programmers productive.

PS: I fully agree that we need more good programmers, I just do not agree that this will ever solve our problems. There simply isn't enough potential in the world.

Mark Walker 04/03/06 03:50:57 AM EDT

It's about time somebody said it. The multitude of frameworks, "evil" wizards and other sugar has led to a generation of inept programmers. I'm all in favour of good tools, but not a substitute for good people.

Just some programmer 04/03/06 12:22:25 AM EDT

Any idiot can write code a computer can understand. Frameworks, design patterns, OOP, and high level languages are all just strategies to mitigate the problem size / complexity curve - a primary goal of competent programmers. Why has the science of software development continued to inch forward when it should have leapt? Why is most "OO" software still written procedurally (and with less skill than procedural code written in 1979)? A lack of competent programmers is certainly at the heart of it but you have to look further than that. Most business stakeholders have no appreciation for what we do. They simply don't care how the work gets done. In fact, doing things 'right' (refactoring, writing automated tests, etc.) can get you in hot water. As a result, there's little motivation to improve and grow in skill. Managers focus on short term goals that cost big in the long run. Shops that value excellence in execution and invest in their employees produce much higher quality software - and programmers.

Wayne Cannon 04/01/06 02:21:03 PM EST

While I agree with your general sentiments, reality is that not every employer can afford to spend $100-150K for "good" programmers with the desired insight and depth of understanding. Looking at the market right now, there are lots of positions only willing to pay $40-80K. For most of them, it's this level or none at all -- and I would suggest that customers and our economy are better off with the current situation, even if some of the solutions are a little rough under the covers.

I've been developing software professionally for 40 years and have been fortunate to have been at the state of the art almost that entire time. The same issues occur in many other industries. Designers of electronics used to understand everything to the component level, but the sophistication, speeds, and associated complexity have increased to the point that only a few now have that level of insight. Many now use high-level (LSI) components without thorough insight -- and occasionally it results in glitches or poor maintainability. I say that's progress, and that the important point is to design architectures (and frameworks) that can guide those that flesh out implementations into better and more maintainable results -- and that "insightful programmers" should definitely be deployed for such architectural design.

Wayne Cannon 04/01/06 02:20:53 PM EST

While I agree with your general sentiments, reality is that not every employer can afford to spend $100-150K for "good" programmers with the desired insight and depth of understanding. Looking at the market right now, there are lots of positions only willing to pay $40-80K. For most of them, it's this level or none at all -- and I would suggest that customers and our economy are better off with the current situation, even if some of the solutions are a little rough under the covers.

I've been developing software professionally for 40 years and have been fortunate to have been at the state of the art almost that entire time. The same issues occur in many other industries. Designers of electronics used to understand everything to the component level, but the sophistication, speeds, and associated complexity have increased to the point that only a few now have that level of insight. Many now use high-level (LSI) components without thorough insight -- and occasionally it results in glitches or poor maintainability. I say that's progress, and that the important point is to design architectures (and frameworks) that can guide those that flesh out implementations into better and more maintainable results -- and that "insightful programmers" should definitely be deployed for such architectural design.

Manjunath Kustagi 03/31/06 09:19:54 PM EST

I absolutely agree, one sees many sundry god-knows who writing programs in industry nowadays

Ditto on Design Patterns being an absolute hogwash, but I disgree with your dislike of OOP.. I think it's based on a sound basis if comprehended and used properly

Sys-Con Australia News Desk 03/29/06 06:16:17 PM EST

It's no secret that I've been outspoken about not liking frameworks for quite some time now. The truth is, I believe that frameworks have a lot to offer. The most significant benefit that organizations stand to gain from using frameworks is a standardized way to code and an environment that is generally more conducive to allowing multiple developers to work on a project at the same time. If frameworks help to standardize how things are done and make it easier for many developers to work on a project, why have I been vocal about not liking them? Am I just trying to create controversy?

Kevin Roche 03/29/06 08:07:06 AM EST

Simon,

I know we will never agree on this. What would you do with all the guys who are never going to be able to develop a sensible architechture that will be suitable for each application. Would you ban them from programming altogether?

Most people are not as clever as you. Without the help of a framework their work would be much less useful.

Brian Cassell 03/15/06 12:34:25 PM EST

I agree with your bold and controversial comments, and will add a few as well.

Frameworks provide a neat way of chopping things into millions of little pieces with the assumption of reusability and granularity. Yes, you can reuse an include file that contains an HTML open tag on every page you create, but what is the sense, what is saved? There seems to be some greater value achieved by "simplifying" things to the smallest element in a framework based implementation. I prefer to open a page and be able to read from top to bottom to see what the heck is happening, rather than looking at the 10 included files that just overly complicate what probably could have been right on one page.

Separation between presentation and content is another popular idea that I disagree with. Presentation is nothing without content, and content is nothing without presentation. They add value to each other. What good is an over engineered application that supposedly can be updated independently of the interface if the front-end design sucks and has no integrated relationship with the content? I prefer to begin with a graphic design interface including content, and then work the code hooks into it, respecting the interface as designed, yet adding the desired functionality.

What do you think?

Edward Trudeau 03/15/06 10:33:14 AM EST

I don't know, Simon...I kind of want to say, "Waah." Some developers aren't where they "should be" after two years of development using frameworks; some are throwing around design pattern names without really "getting it." I'm just not feeling the pain. Are they forcing salaries down? Are they introducing security risks? Are they devaluing resumes with their empty-calorie experience? Are they giving CF a bad name? Are you offended on principle?

All of the above are probably true to a certain extent, but it seems like exhorting developers to grasp the basics, explaining design patterns more carefully, and building an authentic "Object Think" mentality is the more profitable route.

Historically, the technocracy has enjoyed privilege precisely because of the difficulty of grasping the technology. This has inspired a kind of obscurantism that I see in my colleagues; keep it complex so that I can keep my job, I can demand high salaries, I will get the ego boost of being the go-to guy. Talk technobabble to the management - our way to "stick it to the man." Frankly, I've been excited and encouraged by the free sharing of information that I've found in the CF community. I get the impression that people like you, Ben, Ray, Joe, and Sean honestly want all of us to become great programmers and revolutionize the world with fun, useful, productivity-enhancing applications. So I'd rather see the middle way between controversy and silence: dialogue.

Edward