Andrew's ramblings

New Atalanta takes the wrong view of my concern

Tags: OpenBD

Let me first say this to New Atlanta, the move to open source a product is an extreme move. And in this case it is one that I find more than going to be beneficial to the rest of the Coldfusion community, for reasons that I don't need to discuss here. Now I would like to add that it is not a complaint, pure and simple. On the surface it would appear that people have taken constructive criticism and not seen the true meaning behind the bigger picture.

What I blogged, was pure and simple as mentioned a constructive criticism that this product has been around for some years now. And what is generated from CFML is what I consider not very polished code, and I do use that term loosely.

The reason I made specific mention of " Please take this as it is, and not what it isn't.", should have been a clue that there are some serious issues that I have identified. Some are serious in terms of being able to code, and debug with confidence. The other is a not so critical fix, but is something that should also be taken into consideration.

My intent was to highlight something that I strongly believe, is a short fall in the implementation of this product. And as I also stated on my blog in a comment, that it is good that it is open source and has a steering committee, that will not only see the future of CFML expand and grow. But become a better product along the way.

So let's talk about a couple of examples.

The first was when I was debugging a cfloop, it took me longer to find the problem than it should have. And it was as stupid as 2 letters around the wrong way.

Now it might be very obvious, that the above code should have thrown an error along the lines of invalid attribute idnex. It didn't, and when looking at code one can very easily overlook the fact that idnex should have been index.

So if I go along those lines a little further, and unless you know CFML well. Then I see that as a big trap to fall into, because one day that code just might end up looking like this.

The above code should throw an error, something along the lines of " Attribute validation error for the CFLOOP tag ". However openBD, BlueDragon.Net and BueDragon J2EE are more than happy to process the code. This is actually 2 different uses of rolled into one statement, how is the developer going to ever debug that.

The above code is also very happy to run, even though an experienced CFML developer will know that this is wrong. Newer developers will be pulling their hair out wondering why it works, but doesn work as they expect it too.

I did not make this to be critical of the product, and be done with it. I was critical towards the product to ensure, that the product moves forward.

However I was stunned to find such problems have been around in such a comercial product for as long as it has.
Tweet