You can rest assured that none of my judgements are made at speed. Particularly when it comes to technology. I have had a look at CSLA, as I have previously mentioned. I have looked deep enough to be soundly disturbed. Yes, there is a Data Portal (which is what the Message Portal) was built on top of. In my years as a software developer I have never ever seen something like that make it into production. In my opinion, it is for good reason.

Yes, I do know enough about CSLA to know that even the most simple mechanisms are made quite complicated. Your argument suggests that having 30+ frames added to a call stack for a simple function call is acceptable given that the rest of CSLA does the same. In my view, the fact that it’s common or atypical is another indication that CSLA is overengineered and bloated. There’s no excuse for having a call stack that deep to make a simple function call. Yes, I know it’s there “in case you need to do stuff”, but it’s another example of YAGNI, and a very bloated and slow one at that.

To answer your other question, I most certainly did try to influence what was going on. Unfortunately, due to the nature of the management structure and the positions that certain ignorant people were in, I was unable to make any changed. Keef is indeed right, he knows me very well. I am vocal about my views and am quite happy to have discussions about this kind of stuff with those involved. I had attempted to highglight the issues with the framework, but I was simply ignored and basically shut down. Ignorance is bliss isn’t it? So I was in the same boat as you described. As a contractor who came in rather late to the project I wasn’t in any position to make the decision. I did, however, do my best to influence it. My best wasn’t good enough.

That last part is quite interesting! You’re right, there does seem to be a fear out there over being tied to a framework that isn’t supported or owned. If you use the example in my post, you can’t hardly blame them! The framework was based on the open source CSLA. It was hacked to pieces by an inexeperienced developer who made it worse than it was in the first place. This is quite common of open source forks.

Don’t get me wrong, I love open source, and I see great value it in. Unfortunately, it seems to be abused on a fairly regular basis and people end up creating thousands of bastardised branches of code that they should have left alone.

As far as picking a framework and sticking to it goes, I do agree that it’s stupid that people are scared to take it on. NIH really is an issue, particularly in the world of bespoke software development.

@Keef: Thanks mate, you’re right! I only gave up when my contract ended.

You can rest assured that none of my judgements are made at speed. Particularly when it comes to technology. I have had a look at CSLA, as I have previously mentioned. I have looked deep enough to be soundly disturbed. Yes, there is a Data Portal (which is what the Message Portal) was built on top of. In my years as a software developer I have never ever seen something like that make it into production. In my opinion, it is for good reason.

Yes, I do know enough about CSLA to know that even the most simple mechanisms are made quite complicated. Your argument suggests that having 30+ frames added to a call stack for a simple function call is acceptable given that the rest of CSLA does the same. In my view, the fact that it’s common or atypical is another indication that CSLA is overengineered and bloated. There’s no excuse for having a call stack that deep to make a simple function call. Yes, I know it’s there “in case you need to do stuff”, but it’s another example of YAGNI, and a very bloated and slow one at that.

To answer your other question, I most certainly did try to influence what was going on. Unfortunately, due to the nature of the management structure and the positions that certain ignorant people were in, I was unable to make any changed. Keef is indeed right, he knows me very well. I am vocal about my views and am quite happy to have discussions about this kind of stuff with those involved. I had attempted to highglight the issues with the framework, but I was simply ignored and basically shut down. Ignorance is bliss isn’t it? So I was in the same boat as you described. As a contractor who came in rather late to the project I wasn’t in any position to make the decision. I did, however, do my best to influence it. My best wasn’t good enough.

That last part is quite interesting! You’re right, there does seem to be a fear out there over being tied to a framework that isn’t supported or owned. If you use the example in my post, you can’t hardly blame them! The framework was based on the open source CSLA. It was hacked to pieces by an inexeperienced developer who made it worse than it was in the first place. This is quite common of open source forks.

Don’t get me wrong, I love open source, and I see great value it in. Unfortunately, it seems to be abused on a fairly regular basis and people end up creating thousands of bastardised branches of code that they should have left alone.

As far as picking a framework and sticking to it goes, I do agree that it’s stupid that people are scared to take it on. NIH really is an issue, particularly in the world of bespoke software development.

@Keef: Thanks mate, you’re right! I only gave up when my contract ended.

In the case you have described I personally wouldn’t be so quick to judge. For instance, you mentioned CSLA, if you have a look at CSLA the concept of the messageportal is a core part of CSLA (except that it is called a data portal) and uses a lot of reflection. Also you mention the stack being fairly large (I think you mentioned 30+), again if you know CSLA you would know that this sort of stack is not a-typical depending on the scenario you have.

From the sounds of what you have described, did you try and influence what was going on and make them aware of the problems (i.e. the database connections, the object graph, etc). When I am faced with a problem like this, I usually try and help show them what is wrong with what they are doing and that reinventing the wheel isn’t helpful – lets assess what’s really going on here. In some cases, I am not in a position to make the decision, but I always feel I should put my 2c worth in at the time and not just after the fact when it fails. But I am sure this is what you have done and who knows what else was going on at the time.

Also in my experience the decisions surrounding whether to build a framework or not usually comes down to “some” businesses being unwilling to have a hard dependency on open source software. Which I think is for the most part completely absurd. I think what we as an industry are lacking in a general standard by which open source projects can be rated so that when we go into bat and try and convince the people who don’t want to have a “dependency” on open source projects, we can say that this project is better than another one, not just because I or you like it but because the community has given it the following – quality, maintained, reliability, etc rating.

In the case you have described I personally wouldn’t be so quick to judge. For instance, you mentioned CSLA, if you have a look at CSLA the concept of the messageportal is a core part of CSLA (except that it is called a data portal) and uses a lot of reflection. Also you mention the stack being fairly large (I think you mentioned 30+), again if you know CSLA you would know that this sort of stack is not a-typical depending on the scenario you have.

From the sounds of what you have described, did you try and influence what was going on and make them aware of the problems (i.e. the database connections, the object graph, etc). When I am faced with a problem like this, I usually try and help show them what is wrong with what they are doing and that reinventing the wheel isn’t helpful – lets assess what’s really going on here. In some cases, I am not in a position to make the decision, but I always feel I should put my 2c worth in at the time and not just after the fact when it fails. But I am sure this is what you have done and who knows what else was going on at the time.

Also in my experience the decisions surrounding whether to build a framework or not usually comes down to “some” businesses being unwilling to have a hard dependency on open source software. Which I think is for the most part completely absurd. I think what we as an industry are lacking in a general standard by which open source projects can be rated so that when we go into bat and try and convince the people who don’t want to have a “dependency” on open source projects, we can say that this project is better than another one, not just because I or you like it but because the community has given it the following – quality, maintained, reliability, etc rating.

I’ve been around long enough to witness the dubious ethos of some developers (with tendency towards contractors that is to deliberately create an overly complicated solution with the sole purpose of fulfilling their coding fantasies, extending their own contract, and fattening their wallets … and this can be excruciatingly frustrating.

From a business point of view you are right … but the one thing that concerns me is that you may be stomping a little on creativity and innovation.  I have to agree with @tarn and his comment about the people who ‘show us a better way to roll’, people with the engineering spirit and can’t help themselves but to pull things apart, see how they work, and make them better. Yes, they may make mistakes along the way (at the expense of other people’s time and money) but that is how we learn isn’t it?

I think every good developer probably needs to cut their teeth on their version of a framework, a search engine, or write their own blog from scratch, to help with the learning process and to appreciate what goes into making a decent solution.  I wouldn’t want you to send the wrong message and create an army of brain-washed code monkeys who only have a hammer in their toolbox and make every problem look like a nail.

Hmmm, although, an army of brain-washed code monkeys could actually come in handy sometimes.

Keep up the good work!

I’ve been around long enough to witness the dubious ethos of some developers (with tendency towards contractors that is to deliberately create an overly complicated solution with the sole purpose of fulfilling their coding fantasies, extending their own contract, and fattening their wallets … and this can be excruciatingly frustrating.

From a business point of view you are right … but the one thing that concerns me is that you may be stomping a little on creativity and innovation.  I have to agree with @tarn and his comment about the people who ‘show us a better way to roll’, people with the engineering spirit and can’t help themselves but to pull things apart, see how they work, and make them better. Yes, they may make mistakes along the way (at the expense of other people’s time and money) but that is how we learn isn’t it?

I think every good developer probably needs to cut their teeth on their version of a framework, a search engine, or write their own blog from scratch, to help with the learning process and to appreciate what goes into making a decent solution.  I wouldn’t want you to send the wrong message and create an army of brain-washed code monkeys who only have a hammer in their toolbox and make every problem look like a nail.

Hmmm, although, an army of brain-washed code monkeys could actually come in handy sometimes.

Keep up the good work!

There will always be some need for frameworks. I have indeed used ASP.NET Web Forms as well as ASP.NET MVC and WinForms too. They can all be branded as frameworks. Same goes for W?F. Am I happy to use those? Of course I am.

RoR is indeed genius (though made much better thanks to Merb). Django is not something that I’ve played with but I have heard good things about it.

The crux of my issue is this: stop writing bespoke EA frameworks. They’re an epic fail

There will always be some need for frameworks. I have indeed used ASP.NET Web Forms as well as ASP.NET MVC and WinForms too. They can all be branded as frameworks. Same goes for W?F. Am I happy to use those? Of course I am.

RoR is indeed genius (though made much better thanks to Merb). Django is not something that I’ve played with but I have heard good things about it.

The crux of my issue is this: stop writing bespoke EA frameworks. They’re an epic fail