Thomas,
Please forgive me for jumping in impulsively (I've just read this now and
seconds later I am responding.. which could be dangerous =`,8>0)
My initial reaction to your 'patent' concern was to be reminded of my
introduction to JavaSpaces via the 'Sun' book "JavaSpaces(tm) Principles,
Patterns, and Practice".
Somewhere inside of 'Chapter 2' I became disenchanted with that technology
(and I use this word because I really think that I was 'enchanted' by the
whole concept that the JavaSpaces technology is built on) as a result of the
requirement that object properties be made 'public' for the convenience of
'the space', which relies on a 'free-for-all' approach in order to (easily?)
locate objects with the desired properties.
While I agree that it's is the 'easiest' way to accomplish that goal, does
it really make sense to expect a seriously object oriented developer to
suddenly give up private properties. Unless I am very misinformed about all
of these issues (which, because I only read to Chapter 2 and am impulsively
responding to your posting, is entirely possible =8>j) then it seems that
creating an implementation of the 'space' concept (which itself is not so
new -- or unique) that is so radically better in design (at least from an
'O/O' perspective) than the Sun JavaSpaces effort that it completely avoids
their patents (whatever they are) might be fairly simple.
I must admit that (again assuming that I am not way misinformed here, and
with all due respect to Sun) I am surprised and disappointed that Sun
decided that this naive approach to locating objects was 'the way to go'.
Regards,
Thomas J Lukasik
a/k/a [EMAIL PROTECTED]
-----Original Message-----
From: Thomas Down <[EMAIL PROTECTED]>
To: [EMAIL PROTECTED] <[EMAIL PROTECTED]>
Date: Friday, September 03, 1999 7:10 AM
Subject: Jini in a bottle?
>Not quite a classpath issue, but I was wondering if anyone here had any
>views on this:
>
>I'm likely to be doing some academic work for which the JavaSpaces APIs
>look like they could be useful. Eventually, I'd be hoping to release my
>code freely, but I don't think it's necessarily a Good Thing if everyone
>who wants to experiment with it is force to agree to the SCSL. Come to
>think of it, I'm not even sure I want to agree to it myself.
>
>Having looked through the freely downloadable specifications, it seems
>to me that doing a writing a simple cleanroom implementation of
>JavaSpaces shouldn't be too hard. But Sun are hinting that they have
>patents on parts of the technology, and that cleanroom implementations
>may violate these. Does anyone know what these patents might be? Does
>the fact that I'm in the UK make any difference?
>
>If Sun really are going for a `partial open source, plus patents to
>prevent cloning' approach, this seems rather worrying...
>
>Any thoughts?
>
> Thomas.
>
