RTSJ

I thought it was hype 4 years ago, I still think it is hype, but apparently Timesys has a RTSJ VM. If 2 platforms support it, it
might actually go somewhere, however, this would essentially mean that QSS would be handing some market share to Timesys would it not
(since my app would essentially run the same on either platform). I think it would be good for the consumer (me), what do those who
read this NG think ?

“Rennie Allen” <rallen@csical.com> wrote in message
news:bmeojv$ga$1@inn.qnx.com

I thought it was hype 4 years ago, I still think it is hype, but
apparently Timesys has a RTSJ VM. If 2 platforms support it, it
might actually go somewhere, however, this would essentially mean that QSS
would be handing some market share to Timesys would it not
(since my app would essentially run the same on either platform). I think
it would be good for the consumer (me), what do those who
read this NG think ?

From where I sit, it looks like Timesys and QNX took somewhat different
approaches on this subject. They both have committed to using Eclipse as
primary deveopment environment. However Eclipse was originally designed as
Java development tool… so QNX went on to add CDT (so they can write
realtime apps in C) and Timesys went on to implement RTSJ (so they can write
realtime apps in Java).

I’ve heard a claim that IBM’s J9 essentially supports RTSJ, but I don’t
think it is a certified implementation. The lack of Swing makes it somewhat
‘crippled Java’ in eyes of Java developers too… OTOH, using SWT will give
you better performance anyway, so it is mostly an issue of
promotion/education/acceptance.

As of handing market down… compatibility of code does not make one to
switch platform. Performance, price and features do. As situation stands
now, QNX has better performance and comparable features and price. If they
lose this edge, RTSJ will be the least of their problems.

– igor

Rennie Allen wrote:

Igor Kovalenko wrote:


From where I sit, it looks like Timesys and QNX took somewhat different
approaches on this subject. They both have committed to using Eclipse as
primary deveopment environment. However Eclipse was originally
designed as
Java development tool… so QNX went on to add CDT (so they can write
realtime apps in C) and Timesys went on to implement RTSJ (so they can
write
realtime apps in Java).


Hmmm, from where I sit it looks like QSS might have killed RTSJ then. In
order for RTSJ to succeed it needs (at a minimum) two technically viable
implementations. It seems Timesys is one, I think QSS should step up to
the plate, and make another one, particularly since they were cracking
the whip on the horse that was pulling the bandwagon 3 years ago…

I’ve heard a claim that IBM’s J9 essentially supports RTSJ, but I don’t
think it is a certified implementation.


Yes, which is why I think that QSS should work with IBM to get it to
actually turn into something other than hype.

The lack of Swing makes it somewhat
‘crippled Java’ in eyes of Java developers too… OTOH, using SWT will
give
you better performance anyway, so it is mostly an issue of
promotion/education/acceptance.


Well, I certainly don’t think that having SWT and not having Swing
“cripples”
Java; quite the opposite in fact, IMO Java has been crippled by AWT/Swing.

I looked at Java and wrote it off as a GUI development platform several
years
ago when all that existed was AWT. Now that SWT is real, I have
"un"written
off Java as a GUI platform, I am hoping that I’ll be able to “un” write it
off as a real-time programming language also…

As of handing market down… compatibility of code does not make one to
switch platform. Performance, price and features do. As situation stands
now, QNX has better performance and comparable features and price. If
they
lose this edge, RTSJ will be the least of their problems.


Performance is definately an asset, but it is not the most important
asset. LynxOS outperformed QNX in many respects, and where are they now ?

What are those many respects in your opinion?

Timesys has a very big feature (that could be really important) RTSJ.
That’s what I mean by QSS handing some market share to Timesys (if QSS
implements RTSJ then the Timesys approach is validated - as is RTSJ -
and there will undoutably be some customers who might have viewed QNX
as their only viable alternative, to consider Timesys as a viable
alternative). IMO RTSJ becoming a “real” technology, would enlarge
the market sufficiently that QNX would experience a net gain, even if
losing a portion of that pie to Timesys.

Igor Kovalenko wrote:

From where I sit, it looks like Timesys and QNX took somewhat different
approaches on this subject. They both have committed to using Eclipse as
primary deveopment environment. However Eclipse was originally designed as
Java development tool… so QNX went on to add CDT (so they can write
realtime apps in C) and Timesys went on to implement RTSJ (so they can write
realtime apps in Java).

Hmmm, from where I sit it looks like QSS might have killed RTSJ then. In
order for RTSJ to succeed it needs (at a minimum) two technically viable
implementations. It seems Timesys is one, I think QSS should step up to
the plate, and make another one, particularly since they were cracking
the whip on the horse that was pulling the bandwagon 3 years ago…

I’ve heard a claim that IBM’s J9 essentially supports RTSJ, but I don’t
think it is a certified implementation.

Yes, which is why I think that QSS should work with IBM to get it to
actually turn into something other than hype.

The lack of Swing makes it somewhat
‘crippled Java’ in eyes of Java developers too… OTOH, using SWT will give
you better performance anyway, so it is mostly an issue of
promotion/education/acceptance.

Well, I certainly don’t think that having SWT and not having Swing “cripples”
Java; quite the opposite in fact, IMO Java has been crippled by AWT/Swing.

I looked at Java and wrote it off as a GUI development platform several years
ago when all that existed was AWT. Now that SWT is real, I have "un"written
off Java as a GUI platform, I am hoping that I’ll be able to “un” write it
off as a real-time programming language also…

As of handing market down… compatibility of code does not make one to
switch platform. Performance, price and features do. As situation stands
now, QNX has better performance and comparable features and price. If they
lose this edge, RTSJ will be the least of their problems.

Performance is definately an asset, but it is not the most important
asset. LynxOS outperformed QNX in many respects, and where are they now ?
Timesys has a very big feature (that could be really important) RTSJ.
That’s what I mean by QSS handing some market share to Timesys (if QSS
implements RTSJ then the Timesys approach is validated - as is RTSJ -
and there will undoutably be some customers who might have viewed QNX
as their only viable alternative, to consider Timesys as a viable
alternative). IMO RTSJ becoming a “real” technology, would enlarge
the market sufficiently that QNX would experience a net gain, even if
losing a portion of that pie to Timesys.

Yes, which is why I think that QSS should work with IBM to get it to
actually turn into something other than hype.

It exists today. jclRealtime has the API for j9. I have not checked
to see if it at 100% conformance, but it is there.

j9 -jcl:Realtime …

chris


Chris McKillop <cdm@qnx.com> “The faster I go, the behinder I get.”
Software Engineer, QSSL – Lewis Carroll –
http://qnx.wox.org/

Chris McKillop wrote:

Yes, which is why I think that QSS should work with IBM to get it to
actually turn into something other than hype.



It exists today. jclRealtime has the API for j9. I have not checked
to see if it at 100% conformance, but it is there.

j9 -jcl:Realtime …

Yes, I know. My understanding is that it is not 100% conformant though,
and without an actual commitment to it, I doubt it will fly. I realize
that IBM also needs to be committed as well, but I don’t really have an
avenue to voice my opinion with them :slight_smile: