Nothing quite like instant gratification I suppose. Of course you get what you pay for.
I read a lot of SciFi. Actually I read pretty much only SciFi and books that I hope might teach me how to build better systems. Anyway, I read this story once. In it kids would be educated at home by computers and, at some point, they would take a test. The outcome of this test told the kid what they were best suited for. And, in the story, this test was infallible. If the test said you would be best as a garbage man, you would not only be a great garbage man, you would enjoy being a great garbage man.
So in the story there are two boys, friends, and they wish and hope and pray that the test tells them they will be widget fixers (I forget what the widget was). These, of course, were the coolest widgets in the world so working on them is seriously attractive to these guys. They do everything they can to influence the test, reading up on the subject, and various ridiculous things kids do to control something they can't control. Like hanging by your arms so you become taller.
The day of the test arrives and, no surprise, one kid gets what he wanted (widget fixer) and the other one doesn't. In fact the one that doesn't discovers the test says he's really no suited for anything at all. Suffice to say that the first kid, the widget fixer, goes on to be a great widget fixer after having a 10 minute or so bout with RNA shots or something. Until the next model comes along and he discovers he's not changed. Turns out the kid who wasn't suited for anything invented the new and better model.
I'm waiting for some moderately long running tests to complete and I'm surfing, as you do, and I come across this piece of brilliance. Wonderful. If something can't be made to deliver satisfaction in 10 minutes it's pointless. Your software better appeal to the point-and-click set or it won't be adopted and, from the tone I gather doesn't deserve to be either.
Someone, in comments, mentions that the simple things should be simple and the hard things possible. If you've ever actually built something designed to stay up for much more than a week, to do so without a cadre of admins, something that knows enough to know that the world changes and can somewhat adapt, then guess what? That's hard. And Jini makes it possible. But this isn't really about Jini, it's about something much larger. This entry opened my eyes to the apparent fact that there are technology consumers at many levels. Those who require satisfaction in the requisite 10 minutes will, perhaps, end up building yet another RSS feed reader or something equally well defined. It'll probably be nice and have some slightly cool feature that seems useful at the moment. But the 10 minute crowd (hmm, the MTV crowd?) won't be inventing anything new. None of this group will invent a Java, or substantially do much of anything to change the technology landscape. And I suppose that's fine. As that guy in Caddy Shack said, "How about a Fresca?". No, that's not it. Ah, "The world needs ditch diggers too"
I hope the world is not overly influenced by the 10 minute crowd. I hope that having a blog is not all that's needed to be taken seriously. I hope I can recognize these sorts before they start making far reaching design decisions in systems I work on.
It's funny, I read the story Rick related and I took a different message from it. There are people who will spend a long time to understand and use a complex system, and there are others who, by temperament, may not be suited for that. Who knows? One of them may design a simpler system.
> I'm waiting for some moderately long running tests to > complete and I'm surfing, as you do, and I come across > this piece of brilliance. > Wonderful. If something can't be made to deliver > satisfaction in 10 minutes it's pointless. Your software > better appeal to the point-and-click set or it won't be > adopted and, from the tone I gather doesn't deserve to be > either.
That's my piece of brilliance - thanks. But you're taking what I said the wrong way.
I half expected this sort of reaction; and I want to talk to talk a bit about some implicit assumptions, such as; a) I'm stupid, b) Jini is dealing in a domain that is hard, c) no-one has the right to expect anything hard to be usable.
The thing is that my rant was about *starting* Jini; not about network computing - you're not the first person to decide to get those two matters mixed up. Go back and read what I actually said. The fact is I already know and use Jini. The fact is I know quite a bit about networked computing. But - starting Jini 2.0 is arbitrarily hard. I don't believe for a second this represents something neccessary about the problem domain Jini works in. I think it's just sloppy in much the same way a lot UIs are sloppy - their creators aren't able or willing to to hide what doesn't need to be seen.
> Someone, in comments, mentions that the simple things > should be simple and the hard things possible. If you've > ever actually built something designed to stay up for much > more than a week, to do so without a cadre of admins, > something that knows enough to know that the world changes > and can somewhat adapt, then guess what? That's hard.
Yes that difficult. That in fact is what I do for a living. But I've managed to make a distinction between inherent and arbitrary difficulty - and I have no truck with arbitrary difficulty. if you want to talk about what is genuinely difficult that's fine but starting something like Jini isn't remotely close to that.
> Jini makes it possible. But this isn't really about Jini, > it's about something much larger. This entry opened my > eyes to the apparent fact that there are technology > consumers at many levels.
No, it's about Jini startup files.
> Those who require satisfaction > in the requisite 10 minutes will, perhaps, end up building > yet another RSS feed reader or something equally well > defined. It'll probably be nice and have some slightly > cool feature that seems useful at the moment. But the 10 > minute crowd (hmm, the MTV crowd?) won't be inventing > anything new.
I wonder how much you know about RSS if you think writing an aggregator is a well-defined task? As for RSS, I can monitor a heck of lot of servers with an RSS reader and I can probably get that up and running a heck of lot faster than a Jini backed framework. Never mind that Jini isn't suitable for Internet scale work or that RSS is currently turning the journalism/media/portal industries inside out. RSS is likely to become this generation's Syslog. It is *exactly* this kind of good enough technology that changes the world. Not the over-architected stuff that delights in exposing how clever it's creators are. OSI stack, anyone? CORBA? DCOM? No?
> None of this group will invent a Java, or > substantially do much of anything to change the technology > landscape.
To which I counter - those who find excuses in complexity and does not care to excuse their users from complexity will never invent anything worth using. They might invent C++ or COM or some ill-advised junk that rejoices in being hard, but not Ruby or Python or Lisp.
> I hope the world is not overly influenced by the 10 minute > crowd.
:) If you want to say I'm stupid, stop mincing your words and say it - forget about making up some strawmen to argue about. I'm not in that category - it doesn't exist.
> I hope that having a blog is not all that's needed > to be taken seriously. I hope I can recognize these sorts > before they start making far reaching design decisions in > systems I work on.
Look, my point has been that the 10 minute test is something anyone who has technology they want adopted should aim for - *especially* when the problem domain itelf present some complexities that are not going away. Some people have decided to take that quite badly and imply I'm a clown; but I haven't heard a good counter-argument yet. All by-proxy ad-hominem arguments tells me is that I'm probably onto something. No-one seems to want to talk about the actually issue - starting Jini is more work than it should be. Perhaps it's easier to sling some mud around - this is blogland after all.
I still claim it is no wonder Jini isn't getting used.
Jini... aaahh :-) I liked Jini in the early versions. I too found 2.0 a bit odd in the descriptions already. Maybe I just didn't take the time to get it running with 2.0. And still I can't see the need for the Jini lookup service. It's a single point of failure, and doesn't really contribute anything. Clients could still find services over a UDP protocol without the lookup service, and services still send "I'm up again" notification on that same UDP protocol. Am I missing some extremely important function of the lookup service apart from this?
10 minute crowd... I think what was originally meant was that someone has to do some hard work to make something complicated simple to use for others, right? Or no? The 10 minute crowd goes for what is easy. The 10+ minute crowd designs these easy thingys. I find myself in the 10 minute crowd a lot when building systems. Easy or out! I don't have time to spend days only to discover something can't do what I want it to.