You have to like the way, in his latest piece on metadirectory, Dave Kearns summons Lewis Carroll to chide me for using the word “metadirectory” to mean whatever I want:
“When I use a word,” Humpty Dumpty said, in rather a scornful tone, “it means just what I choose it to mean—neither more nor less.”
“The question is, ” said Alice, “whether you can make words mean so many different things.”
“The question is,” said Humpty Dumpty. “which is to be master—that's all.
Kim talks about a “second generation” metadirectory. Metadirectory 2.0 if you will. First time I've heard about it. First time anyone has heard about it, for that matter. There is no such animal. Every metadirectory on the market meets the definition which Kim provides as “first generation”.
It's time to move on away from the huge silo that sucks up data, disk space, RAM and bandwidth and move on to a more lithe, agile, ubiquitous and pervasive identity layer. Organized as an identity hub which sees all of the authoritative sources and delivers, via the developer's chosen protocol, the data the application needs when and where it's needed.
It's funny. I remember sitting around in Craig Burton's office in 1995 while he, Jamie Lewis and I tried to figure out what we should call the new kind of multi-centered logical directory that each of us had come to understand was needed for distributed computing.
After a while, Craig threw out the word “metadirectory”. I was completely amazed. My colleagues and I had also come up with the word “metadirectory”, but we figured the name would be way too “intellectual” – even though the idea of a “directory of directories” was exactly right.
Craig just laughed the way he always does when you say something naive. Anyone who knows Craig will be able to hear him saying, “Kim, we can call it whatever we want. If we call it what it really is, how can that be wrong?”
So guess what? The thing we were calling a metadirectory was a logical directory, not a physical one. We figured that the output of one instance was the input to the next – there was no center. The metadirectory would speak all protocols, support different technologies and schemas, support referral to specific application directories, and preserve the application-related characteristics of the constituent data stores. I'll come back to these ideas going forward because I think they are still super important.
My message to Dave is that I haven't changed what I mean by metadirectory one iota since the term was first introduced in 1995. I've always seen what is now called virtual directory as an aspect of metadirectory. In fact, I shipped a product that included virtual directory in 1996. It not only synchronized, but it did what we used to call “chaining” and “referral” in order to create composite views across multiple physical directories. It did this not only at the server, but optionally on the client.
Of course, there were implementations of metadirectory that were “a bit more focussed”. Customers put specific things at the top of their list of “must-haves”, and that is what everyone in the industry tried to build.
But though certain features predominated in the early days of metadirectory, that doesn't mean that those features ARE metadirectory. We still live in the age of the logical directory, and ALL the aspects of the metadirectory that address that fact will continue to be important.
[Read the rest of Dave's post here.]