I appreciate your enthusiasm, it's a good thing and I'm not against
the discussion...

What qualifies as "success"?

Adoption of Jini everywhere, using it the right way for the right
things, making some money from consulting?

"Jini is full of excess".

Yep for sure there's too much for most to cope with. Question is
though, can you reduce it down and still tackle the distributed
systems challenges properly?

Maybe you can make service building a little easier, sugar coat
configuration and packaging but will that bring you "success" (see
question above)?

Jini I reckon is an embodiment of practices and principles for doing
distributed systems well. Is it the sweet spot? Maybe not, the Java
and mobility thing is a double edged sword. Powerful but brings some
pain and doesn't play so nice with other platforms (sure you can use
surrogate but is that sugar coated enough?)

Let's say you manage some level of sugar coating - can you do enough
to hide the challenges of understanding distribution and coding
properly for it? Don't think so. That is seriously audience limiting
so if "success" is mass adoption, I don't think you wanna get your
hopes up.

Or we could rebuild Jini around the patterns and principles and lose a
few bits to make it more acceptable to a broader audience. How much
broader though?

If you add annotations and such you can give it a character somewhat
like J2EE but I reckon you'll attract an audience that expects
something like actors and will run away screaming once they start
dealing in leases and the fact that, oh heck, things can go AWOL.

For me at least, there's the rub. You can put a lot of work in and not
vastly expand your user base. I'm not saying don't bother putting in
some work but pick the sweet spot which mightn't be sugar coating
(80/20 rules). I'd also say that if "success" is mass adoption you're
looking at boiling the ocean, you'll need to undo the mess that is the
system that produces low-grade developers and persuade business to
consider aptitude in their hiring. And if ocean boiling is what you
want to do (and I'm not saying don't just that it's a long hard road),
I reckon kindle'ing a couple of books or similar isn't the place to
start.


On 27 January 2012 16:55, Gregg Wonderly <[email protected]> wrote:
> Here's what I'm thinking about.  First, there is the "early visibility"
> issue, that can "increase" the traffic just because people are looking at
> "what's available", rather than looking for something specific.  Second, the
> straglers, might be the "voices" that are listened to, once they discover
> Jini and start talking about.
>
> Believe me, when I say that Jini is full of excess.  There are just too many
> things that are too much effort to use or do, because it's still a platform,
> and not actually much of a toolset.  This is why I've spent so much of my
> time trying to create tools, asking about unifying "platform" views to
> create a "hierarchy" of interfaces and classes for "defining" as server that
> is "Jini enabled".
>
> There is just so much to do, but there is never any agreement on what to do,
> because there is, of course, so many things that Jini can do, and so many
> ways to do it.
>
> There are literally 100's of thousands of lines of code in the collective
> Jini community that is going to waste, because it's not been brought
> together, collected and assembled into a toolset that we can at least tryout
> to see what else we might need.
>
> There is always the argument about whether Jini should be made to run inside
> of a JavaEE container.  There is the counter position that Jini can run a
> servlet container as a service so why even get into JavaEE.
>
> There are lots of standards around related transactional processing that
> Jini can't do natively.  There are always the "It's Just RPC and Corba
> proved that was too painful" comments/arguments running around.
>
> Practically, we just don't have anything easily demonstrable to hand out and
> show that Jini can be a part of a large number of different modes of system
> build out, and it is an asset, not a liability to use it.
>
> So, I'm looking for ways to get to people "earlier" in their exposure to
> "computer science".  All the old farts and all the "Rest" crowd just spew
> stuff out that we don't have demonstrable arguments against.
>
> We need new, powerful experiences for a large number of "fresh minded"
> people...
>
> Gregg
>
>
>
> On 1/27/2012 10:28 AM, Luis Matta wrote:
>>
>> Furthermore a true distributed architecture, well
>> thought/designed/implemented and freely available to students is unheard
>> by
>> me.
>> (I could say here: fault tolerant, transactional, performant, protocol
>> agnostic, but that is way we care about Jini anyway)
>>
>> I believe Jini has to focus on its strengths, because it maybe that on
>> distributed systems one should only exchange data
>> (so the structures that Jini create are an excess), it maybe that the java
>> type system is to rigid for distributed systems,
>>  yada yada yada.       There is no other Jini, and it rocks.
>>
>> The interest is there!
>>
>> On Fri, Jan 27, 2012 at 2:08 PM, Dan
>> Creswell<[email protected]>wrote:
>>
>>> On 27 January 2012 15:53, Gregg Wonderly<[email protected]>  wrote:
>>>>
>>>>
>>>> On Jan 26, 2012, at 3:28 AM, Dan Creswell wrote:
>>>>
>>>>> On 25 January 2012 23:52, Luis Matta<[email protected]>  wrote:
>>>>>>
>>>>>> Wonderful idea,  but I do not believe very much in its success (so
>>>
>>> ignore
>>>>>>
>>>>>> me).
>>>>>>
>>>>>
>>>>> I'm with you on that.
>>>>
>>>>
>>>> Is that a "Not a wonderful idea", "with you", or a "do not believe",
>>>
>>> "with you", or both?
>>>
>>> Both.
>>>
>>> Because I don't believe we have an accessibility problem to worry
>>> about rather we have an appreciation problem. For example, if a book
>>> on Python and a book on Jini are both available on iPad and we present
>>> an average student with the chance to buy either and they're both at
>>> the same price which one will they pick?
>>>
>>> I reckon they'll pick the Python one - more likely to have heard about
>>> it, it's a programming language and that's a subject all geeks readily
>>> relate to. i.e. A book on Python fits the expectations and habits of
>>> most. A book on Jini though doesn't fit expectations and habits.
>>>
>>> You may scoop up a few stragglers but there's a more general
>>> educational thing that needs taking care of first. Create demand, then
>>> you can sell something (unless the demand already exists).
>>>
>>>>
>>>> I'm just wondering where some of the "we teach web programming"
>>>
>>> "schools" are going to go with their programs, considering how much
>>> effort
>>> and interest there is on mobile devices.  It seems interesting to me,
>>> that
>>> it might be possible to "expose" people to a different path for
>>> application
>>> development just by raising the visibility of Jini.
>>>>
>>>>
>>>> I'm not sure it's an awesome idea myself, but just curious what others
>>>
>>> views are.
>>>>
>>>>
>>>> Gregg
>>>
>>>
>>
>

Reply via email to