Firstly, I agree completly that it is moronic to set things up in this way... It's almost as if firaxis is trying to discredit the scientific method. 
It's the best thing that have ever happend in this universe, and that you don't get any serious bonus from it, but instead are beeing punished, is a gigantic fail and a mockery toward reality.
This obvious design flaw aside...
I almost never build monesteries, the 10% is too small to give a return of hammer investment for a forseable future.
If you are trying to be religious there is generally a period in which you can build all the misionaries you need, while in organized religion.
If you want to spread the religion everywhere, you have a monestery in a production focused city, so the 10% is even more irrelevant there.
The AP hammer bonus is nice, and if some AI builds it and I happen to belong to the same religion. I might go for some monesteries.
I don't normally build neither the parthelon nor the Glib neither, I am usually focusing more on either expanding or conquering a neighbour, so some AI is likely to beat me to it.
And if you follow rhinos advice about binary research, beeline through sci-meth toward biology.
The extra food will make it possible for your to hire 100 scientists to make up for the two you lost with the great library.

It's the best thing that have ever happend in this universe, and that you don't get any serious bonus from it, but instead are beeing punished, is a gigantic fail and a mockery toward reality.
This obvious design flaw aside...
I almost never build monesteries, the 10% is too small to give a return of hammer investment for a forseable future.
If you are trying to be religious there is generally a period in which you can build all the misionaries you need, while in organized religion.
If you want to spread the religion everywhere, you have a monestery in a production focused city, so the 10% is even more irrelevant there.
The AP hammer bonus is nice, and if some AI builds it and I happen to belong to the same religion. I might go for some monesteries.
I don't normally build neither the parthelon nor the Glib neither, I am usually focusing more on either expanding or conquering a neighbour, so some AI is likely to beat me to it.
And if you follow rhinos advice about binary research, beeline through sci-meth toward biology.
The extra food will make it possible for your to hire 100 scientists to make up for the two you lost with the great library.