Core Auth Library

[eluser]n0xie[/eluser]
I too think it's a lost cause to try and make a generic standard Auth Library. For our in-house projects I have written such a Library and the problem is that if you don't assume anything, your Library becomes a mess. So you have to assume 'some' things. And for every detail you assume, you will make the Library less useful for others. So our Library assumes quite a bit, and we are content with that because it fills our needs, but I bet it would be horrible for other developers. They would either think the assumptions were impractical, restricting, or downright bad design. Which is why I haven't released it to the community at large: it is perfect for our needs, but would need a lot of hacking for someone else, which means they'd be better off with one of the Auth Libraries already available.

These assumptions is why there are so many Auth Libraries out there and I still decided to write my own. They were either too restrictive or too generic for my taste. This is in no way saying they are bad libraries: on the contrary most are pretty well written and pretty clean. But for every assumption the Library made, it made it less useful for me.

So think about that before you put too much energy into this project: for every assumption you make, you will lose part of it's interest to a part of the community. Make it too generic, and it becomes useless and bloated.

Just my 2 cnts.