[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: ADTPro beta available

mdj wrote:

Michael J. Mahon wrote:



The critical thing is: whose bug is it if a Java application fails to
run on a particular system?

If it runs on all other implementations of Java (and meets the specs
for a valid Java application), then the bug is in the implementation
of the Java platform on the non-functioning system.

If it does not run on other implementations of Java (and violates the
specs in some way), then the bug is in the application.  (BTW, if such
an invalid application *does* run on some Java implementation, there is
a likelihood that that implementation also contains a bug.)

*That's* the strong test for what is a "platform".



I'm not sure I like this definition either :-) Many platforms have only
a single implementation, yet the definition still stands.


*Successful* ones have multiple implementations.

I gave the definition as an "operational" definition.  Of course, there
must be a _de jure_ definition as well, to keep people marching in line.
For every platform, there is a council/forum to decide what is a valid
implementation and what isn't.  (Come to think of it, counting councils
may be the easiest way to count platforms.  ;-)


I think you have a much to restrictive notion of "platform".

A more useful definition of a platform is any layer of uniformity that
enables a common market in applications.  (Unix, for example, never
quite made it to this status, mostly for competitive reasons.)



It's interesting to watch the commercial Unix's being replaced bit by
bit with Linux, which provides a consistent API across essentially all
the hardware 'platforms' that commerical Unix's ran on, plus then some.


And in the process, the hardware vendors give up their proprietary
"look and feel", API, etc.  That was always the barrier to a common
Unix.  Too busy fighting among themselves dividing the pie to spend
any real effort on expanding the pie.  ;-)

They wouldn't even be converting to a standard Unix now unless they
were being beaten about the head and shoulders by Microsoft.

Say what you will about Microsoft, they have been a powerful force
of reality in a marketplace once addicted to fantasy.  The message
is simple and Darwinian:  do a better job (than we do) giving customers
what they want (a capable, affordable, unified platform), or die.


For this to be viable, the platform definition must be sufficiently
"complete" that one does not need to go outside it to write useful
applications, and it must be sufficiently "strong" that a single
testing regime is sufficient for deployment on all implementations
of the platform.



I agree in essence with this definition.

My point is, I currently have 3 Java applications open as I write this.
None of them feels like a different platform to me, they just behave
like any other Windows application. The net effect is more closely
aligned with the concept of a cross-platform framework than that of a
platform. In fact most people would not be able to tell the Java
programs from the Windows programs.

I guess you could say it fits the definition from the development
perspective, but not from the deployment or user perspectives.


So does each target system require special user training?  (For
installation, updates, operation, backups, etc.)


..NET does have one advantage over Java though: .NET is a multi-language
environment - there are at least a few languages available for .NET.
I suppose that's why that environment is called ".NET" rather than "C#"....


In fact, .NET and the CLI create a platform whose implementations span
Windows XP, Windows CE, and other embedded Windows environments--with
the appropriate resource and functionality subsetting that is required.

This is, in fact, the problem that it was meant to solve.



It does do this. One could call this a subset of the problem space Java
was meant to solve, as it also manages to solve those problems, and
extends to being able to solve server side problems as well, and isn't
limited to Windows environments.


As has been noted, that limitation is not intrinsic, but business
model motivated.  If you're little, you stick together; if you're
big, you don't play with mice.

-michael

Fast Sudoku solver for Apple II's!
Home page:  http://members.aol.com/MJMahon/

"The wastebasket is our most important design
tool--and it's seriously underused."