Engineering Religion

If you use "engineering" in the title, you're going to have to be much more precise. What goals and measurements are you interested in? Why are you considering "religion" as a unitary thing, distinct from other belief and social conformity mechanisms?

What is success for a religion?

I'd say it seems pretty fundamental that Religion is a kind of meme. One measure of success for memes is their ability to spread (virulence). If that is your only measure, you're likely to have some ethically terrible things going on. It seems (to me) like an obvious constraint that the meme must spread without causing any obvious harm in its wake (except to related memes that may be in competition).

What useful purposes does religion serve?

I consider [create an in-group] to be a pretty central (usually unstated) optimization point (not exactly a purpose) in most religions. This has some well-studied psychological effects on said group that can benefit all its members socially, psychologically, and (further down the causal line) medically. Although it tends to lead to some unfortunate side-effects for the out-group.

How would you design a "rational religion", if such an entity is possible?

Problem is that I've seen more than one source define "religion" as something like "systematic belief in the supernatural". I'm not convinced that such belief can be "rational" (optimal for making more effective decisions). Perhaps as a stepping stone -- use a religious-type belief system as an infection vector, then strip the supernatural elements away bit by bit; but that seems awfully deceitful to me.

I'm new here and not sure exactly what you expect when someone links, but it seems like you guys are generally intelligent so:

http://www.enotes.com/research-starters/sociological-theories-religion-structural

It sounds like what you're asking (with regards to the function of religion) is something that has been covered a great deal by the structural-functionalist sociological approach. If you're willing to read up on it there's a lot of information out there on the topic. Hope that helps! If you'd prefer I answer your question on here more directly feel free to ask, I'm in the middle of finals and haven't read up too much on the topic myself so I'd have to do some research before getting back to you.

Religion serves numerous purposes, some of which have been mentioned already by other commenters. I want to add two others:

1) a mechanism for preserving Intersubjective Truths, that is, truths that it is not possible to re-derive from first principles in a reasonable amount of time.

2) a connection to the spiritual side of life and spiritual experiences.

Also, with so many different purposes it is tempting to design different religions to fill all these roles. I suspect that is harder than it seems. Since whatever fills at least some of these roles will attempt to expand to fill all of them.

As kithpendragon has already wrote, religions from a memetic point of view must contain tools for its propagation and defence, such as ways to infect other agents and fight aggressions.
I also want to propose another question: the purposes served by religions are better served by what other social constructs?

One way of defining things is when you inherently let signifcant parts of your life be dictated by trust. Normally there are requirements to earning trust such as being good in a field of knowledge to be able to function as a expert regarding such subject allowing others to skip details.

You can obvoiusly enter into dynamics that other dynamics are not allowed to move into such easily. Science or courts need some level of proof. Military power is dependent on weapon production economies. If you just trust it will be a good thing if you do it you don't need any of those complications.

For example a trust based approach to prisoners dilemma can just have the players cooperate while agents wanting to be game theorethically competent must make up all kinds of excuses to cooperate or suffer contraoperation.

I am really not a sociologist, so someone correct me if I what I'll say is totally wrong, but it seems to me that there are at least two quite distinct types of religion (and a continuum of possibilities in-between), the first one consisting of those religions where "religion" religion (gods, clergy, etc.) is almost one and the same thing as something like civil religion of a community (for example, if you found out that a tribe adds various religious chants to their local "judical process" which otherwise is very similar to a Western judical process you would not hesitate to call it a religious ritual, even though chanting part may be inessential), and another where those are two different things. In my mind the first type roughly corresponds to paganism, and the second one to religions similar to Christianity. I think that religions of the first type may be useful to "grease the wheels" of society (especially not a very sophisticated one), and its leaders may not even be that interested in spreading it, except for personal gains. However, note that this is a vague guess, I would need to read much more about pagan societies to understand if it is at least partially the case. It is unclear to me what does religions of the second type do, because greasing the wheels of society is covered by a civic religion.

Another possible benefit of some types of religion is providing some incentives to get some (although not all) things correct. If you believe that god judges you whether or not you your thinking about the world is correct, you might feel and behave as if you have "skin in the game" and thus you might end up with higher motivation to avoid deceiving yourself (and others) for personal gain [1]. In many contemporary societies personal belief in god seems isolated from most beliefs that have practical consequences. If your society practices trial by combat, and your belief in god makes you willing to fight against a stronger person believing that you will win just because you are in the right, your belief will have practical personal consequences. However, in a contemporary society belief in god is usually harmful only indirectly. Thus, a question is, does feeling that you have "skin in the game of believing the truth about god's creation" lead to enough correct beliefs so as to outweigh having incorrect beliefs about god? I think that it is likely that at least in some cases might do. In addition to that perhaps in some cases religious beliefs might be personally helpful if they are harmless and they displace potentially dangerous beliefs who are new and have not yet shed their most extreme parts.

I think that in both cases religion/belief in god is not strictly a necessity, but in some cases it may (or may not) turn out to be somewhat useful if there are no better alternatives available at that time.

[1] I don't like it. Also, I guess that it is probably not very good for society in the long run.

A successful religion must serve Moloch, and your designer choice is how much and what exactly are you going to sacrifice first.

Yeah, I realize this is a fully general argument against anything. Maybe the reasonable part is that I wouldn't want to compete at a religion market. I'd rather stay at a level of communities. Yeah, seems like a debate over terminology, but those two words have widely different connotations for me. In communities people stay freely; different communities are not necessarily enemies; etc.

Could you define what you mean with the word religion?