LDAP is a protocol. NoSQL is a shitload of software that works extremely different and has no protocol or common data scheme.
LDAP is probably most used for organizing a companies data (employees) and works quite well for that. It's proven, it has major software products that are stable and used in worldwide deployments.
Just exchanging LDAP with NoSQL makes no sense and is a move only driven by "LDAP sounds soo old. Let's use that NoSQL everyone is talking about!". And maybe slap XML, Web2.0 and HTML5 on top of it. Just, you know.. it's state of the art!
I've yet to see a comment on this page that relates to the problems of the protocol. Mostly vague points without objective data.
Basing all those criticism on the fact that this protocol has been around or a while and thus has more than one RFC is just stupid.
No, more just idly wondering if it'd be possible to build something that is a less shitty experience than LDAP - see above/below. I'm not particularly interested in the NoSQL stuff, by and large, because for what I do, Postgres is more than enough, however, I'm curious enough to wonder "what could be", and think that with all the different new systems out there, whether it'd be possible to build something new and better.
I don't have the answers, I just didn't really like my experience with LDAP, and was idly wondering whether it'd be possible to build something better these days.
It's a particularly ugly protocol, if you ask me. I hacked and slashed and swore and got it working for a company I worked for several years ago, but it was not a pleasant experience.
I know my way around Unix, and am a fast learner, so "no experience" is not something that generally scares me. I know I'll make some mistakes and waste a bit of time.
That said, there are plenty of systems that I have encountered in my years that are much friendlier to a new user. Specifically, I recall awful command line queries, kind of wonky tools in general, and, generally, a fiddly feeling to the whole thing. I did get things working, but I was never really happy with the whole setup. We ended up calling in a consultant to check over my work and see if there were ways to improve it, and aside from a few things, there really weren't.
Yes, but honestly, in my world "a strange feeling" doesn't qualify for ojective proof that a system is bad.
There are plenty of powerful tools, and if you don't like the commandline (although you say you use unix a lot, ldap commands are straight forward), you probably could've used GUIs in the most cases. My "feeling" of this thread is that people were unsatisfied with openldap and now blame a protocol.
The major enterprise directories have a lot of additional commands/features/web interfaces to fiddle with.
It's not the protocols problem when a specific implementation is hard to use, isn't it?
LDAP is a protocol. NoSQL is a shitload of software that works extremely different and has no protocol or common data scheme. LDAP is probably most used for organizing a companies data (employees) and works quite well for that. It's proven, it has major software products that are stable and used in worldwide deployments.
Just exchanging LDAP with NoSQL makes no sense and is a move only driven by "LDAP sounds soo old. Let's use that NoSQL everyone is talking about!". And maybe slap XML, Web2.0 and HTML5 on top of it. Just, you know.. it's state of the art!
I've yet to see a comment on this page that relates to the problems of the protocol. Mostly vague points without objective data.
Basing all those criticism on the fact that this protocol has been around or a while and thus has more than one RFC is just stupid.