The Apache MyFaces team is pleased to announce the release of Apache MyFaces Core 1.2.0. MyFaces Core 1.2.x is a JavaServer(tm) Faces 1.2 implementation as specified by JSR-252. MyFaces Core has passed Sun's JSR-252 TCK and is 100% compliant with the JSR-252 specification. MyFaces Core 1.2.0 is available in both binary and source distributions. * http://myfaces.apache.org/download.html MyFaces Core is also available in the central Maven repository under Group ID "org.apache.myfaces.core". Release notes are available. From the editor: It's as yet unclear whether Tomahawk, which extends JSF with new components and converters, is compatible with MyFaces Core 1.2.0, as Tomahawk is specified as a JSF 1.1 library. Likewise with the other MyFaces subprojects, such as Tobago (more user interface components and theming), Trinidad (dialog management and page flow scope), and Orchestra (which combines JSF with a persistence layer). They should be compatible, however: is anyone willing to confirm?

The Tomahawk components work with MyFaces 1.2.x, there was no overhaul on the JSF specific API in tomahawk yet, but this will happen soon. The same is true for Tobago. Orchestra can also be used with JSF 1.2 (MyFaces 1.2.x). Trinidad 1.2.1 works fine with MyFaces 1.2.0, where Trinidad 1.2.1 is already "optimized" for the JSF 1.2 API. MyFaces 1.2 currently lives on a branch and the JSF 1.2 stuff will become trunk, to have it the main JSF-IMPL. MyFaces 1.1.x will go to a branch, for "maintenance" releases.

Congrats guys!

Can someone please compare JSF 1.2 SUN RI vs JSF 1.2 MyFaces? --Mark

That's also what I want to know.

Can someone please compare JSF 1.2 SUN RI vs JSF 1.2 MyFaces?

--Mark

Different license, different implementation same api... ;-)

Congratulations for getting this out! :) It must have been an incredibly amount of work. I'm going to set up a test project using this very soon. :)

Can someone please compare JSF 1.2 SUN RI vs JSF 1.2 MyFaces?

--Mark

Different license, different implementation same api...
;-)

Indeed, of course JSF 1.2 SUN RI has been out for more than a year, so it also has seen quite a lot of bug fixes already. MyFaces 1.2 is brand new, so we might not -really- be able to use this in production for yet another year (depending on your company's policy of course). I do wonder, are any of the new Java EE 5 servers going to use this? At least Geronimo will (of course), but could Jboss 5 AS also use this? I know that in the past they said they were going for SUN RI. However since Jboss 5 AS is probably still something like a year away, any chances they would make up their minds now MyFaces 1.2 is finally here?

I do wonder, are any of the new Java EE 5 servers going to use this? At least Geronimo will (of course), but could Jboss 5 AS also use this? I know that in the past they said they were going for SUN RI. However since Jboss 5 AS is probably still something like a year away, any chances they would make up their minds now MyFaces 1.2 is finally here?

I don't see us moving back to MyFaces unless we are presented with a compelling reason to do so. It's a bit ironic since I was the one who started the MyFaces 1.2 project in the first place. (getting pulled off was beyond my control) And, BTW, congratulations to the guys who took over and finished the job! Well done. As was previously pointed out, MyFaces 1.2 is new and the Sun version has been out for over a year. JBoss 5 is certainly NOT a whole year away. So, I just don't see why we would switch back. But if you still want to run MyFaces Core on JBoss 5 you will be able to do that. We have a switch to let you run MyFaces 1.1 with it today. It hasn't been tested with MyFaces 1.2 yet, but it probably works: http://wiki.jboss.org/wiki/Wiki.jsp?page=JBoss5AndMyFaces Stan Silvert JSF Lead Red Hat/JBoss

Can someone please compare JSF 1.2 SUN RI vs JSF 1.2 MyFaces?

--Mark

My experiences are with RI 1.2 vs. Myfaces 1.1.x, but i expect that much of this still holds true: - RI 1.2 has been ready for over 12 months, and is really solid and bug free. I can't make any hard claims about MyFaces 1.2 quality, but since it's recent release i'd assume that RI has definite edge in bugs and stability. - RI 1.2 has paid-for developers, and their support and turn-around time for bug fixes has been perhaps the best i'v ever experienced with any open source software. I have no experiences with Myfaces in this regard. - RI 1.2 source code is really clean, and extremely well documented. The last time i looked at MyFaces code (>12 months ago) it was not particularly well documented, and had strange coding practices (iirc the instance variables were prefixed with underscore etc.). - Myfaces can go 'beyong the spec' and has added previously handy extra features (for example keeping the scroll position of window between requests). RI is just about the spec, and nothing but the spec. - Myfaces has traditionally had broader developer community, where people with great knowledge help, innovate and share ideas. RI doesn't have that kind of community, at it's core its run by paid-for developers. - Myfaces has more of an ecosystem around it; tomahawk, tobago, trinidad, and now orchestra. RI has some, but it's not at the same level. /Henri Karapuu

Can someone please compare JSF 1.2 SUN RI vs JSF 1.2 MyFaces?

--Mark

I can't make any hard claims about MyFaces 1.2 quality, ...

I have no experiences with Myfaces in this regard.

The last time i looked at MyFaces code (>12 months ago)

.... why did you even reply to his question?

RI doesn't have that kind of community, at it's core its run by paid-for developers.

While both teams have excellent developers, Ryan Lubke, from Sun, does steal a bit of the spotlight from the RI with his quick turn around; but there are other developers that make commits. Because of the nature of being the RI, there's a large vendor community which isn't openly obvious to everyone else. The MyFaces guys have always been able to pull together genius ideas and I think the competition between the two implementations is fantastic! With MyFaces now at 1.2, both implementations are only going to get much better.

The last time i looked at MyFaces code (>12 months ago)

.... why did you even reply to his question?

Because the question was asked, it was important question, and nobody else was offering their advice. I feel that all of the points i raised are caused more by the philosophies and backgrounds behind the implementations, than exact versions numbers. And as you noticed, i made my experience level explicitly clear so that others can evaluate whether to trust my opinion. I know you are RI contributor, and really didn't mean to put down your efforts or RI community, sorry. It's just that i feel MyFaces is stronger in this respect, while RI is in others. And i call them as i see them, and try to share information with others. /Henri Karapuu

Can someone please compare JSF 1.2 SUN RI vs JSF 1.2 MyFaces?

--Mark

My experiences are with RI 1.2 vs. Myfaces 1.1.x, but i expect that much of this still holds true:

- RI 1.2 has been ready for over 12 months, and is really solid and bug free. I can't make any hard claims about MyFaces 1.2 quality, but since it's recent release i'd assume that RI has definite edge in bugs and stability.

- RI 1.2 has paid-for developers, and their support and turn-around time for bug fixes has been perhaps the best i'v ever experienced with any open source software. I have no experiences with Myfaces in this regard.

- RI 1.2 source code is really clean, and extremely well documented. The last time i looked at MyFaces code (>12 months ago) it was not particularly well documented, and had strange coding practices (iirc the instance variables were prefixed with underscore etc.).

- Myfaces can go 'beyong the spec' and has added previously handy extra features (for example keeping the scroll position of window between requests). RI is just about the spec, and nothing but the spec.

- Myfaces has traditionally had broader developer community, where people with great knowledge help, innovate and share ideas. RI doesn't have that kind of community, at it's core its run by paid-for developers.

- Myfaces has more of an ecosystem around it; tomahawk, tobago, trinidad, and now orchestra. RI has some, but it's not at the same level.

/Henri Karapuu

Thanks for comprehensive response. Do you think JSF benefits from several implementations of specification? IMHO JSF looses its ground because of different implementation of same thing. The result several "not so good" implementation. Regards, Mark

Thanks for comprehensive response.

Do you think JSF benefits from several implementations of specification?
IMHO JSF looses its ground because of different implementation of same thing. The result several "not so good" implementation.

Regards,
Mark

I am not the original poster, but I personally think as long as there is a limited number of implementations it is better than just having one. I have asked myself the question also. So far the competition between myfaces and the RI has benefitted both frameworks to a big degree, MyFaces used to be the only implementation being opensource, now the RI is as well, on the other hand MyFaces got some ideas of "hidden" RI functionality like the server side state history, while the RI borrowed other ideas of MyFaces. If we had only one implementation a significant aspect of driving the quality and ideas forward would go away, monopolies are always bad, even codewise. While most aspects in the licensing are gone there still are some aspects which probably will never go away. MyFaces is under the apache license while the RI is under the CDDL, for the user it does not really make a difference but for projects which stay on a pure BSDish approach this is a significant difference. As for now to my knowledge Apache software cannot integrate CDDL code on a source level (same goes for most other BSDish projects as well I assume) So there still are licensing incentives to have a second slightly more free implementation, even if it does not really make a difference for the end user, who can read both codebases and can contribute to both. I think the entire aspect if it makes sense to have two implementation comes down to the fact if it makes sense to have more than one app server, more than one servlet implementation etc... It makes sense as long as it does not cause a fracturing of the "market" which thankfully the specification and the TCK prevent!

MyFaces is under the apache license while the RI is under the CDDL, for the user it does not really make a difference but for projects which stay on a pure BSDish approach this is a significant difference. As for now to my knowledge Apache software cannot integrate CDDL code on a source level (same goes for most other BSDish projects as well I assume)

Actually, the JSF RI is dual licensed under CDDL and GPL v2 with the classpath exception, as is the entirety of the JavaEE RI (GlassFish). With that being the case, many of the licensing issues are now "fixed." Unfortunately, there's still the GPL vs BSD battle, which some people find REALLY important, though I would think the CP excpetion would mitigate that somewhat if not entirely. I am not a lawyer, though, and have had no opportunity to play one on TV. :)

MyFaces is under the apache license while the RI is under the CDDL, for the user it does not really make a difference but for projects which stay on a pure BSDish approach this is a significant difference. As for now to my knowledge Apache software cannot integrate CDDL code on a source level (same goes for most other BSDish projects as well I assume)

Actually, the JSF RI is dual licensed under CDDL and GPL v2 with the classpath exception, as is the entirety of the JavaEE RI (GlassFish). With that being the case, many of the licensing issues are now "fixed." Unfortunately, there's still the GPL vs BSD battle, which some people find REALLY important, though I would think the CP excpetion would mitigate that somewhat if not entirely. I am not a lawyer, though, and have had no opportunity to play one on TV. :)

I am not a laywer either, but it does not really matter for the endusers anyway, the main thing as I understand it is, that BSD is probably the most liberal license except for do anything with the code without giving credits. The main problem even with GPL and classpath exception (even more for the CDDL is) that code which is under bsd, should not be mixed in the bsd projects repository with licenses which basically say, do anything with the code bug open and give back the changes. So the classpath Exception in GPL2 is basically useless for the authors of other libraries who want to integrate original code under more liberal licenses, the same goes for the CDDL which is sort of a realtered mozille license, it simply is not possible without altering the underlying BSD license into something less liberal. Well all this nitpicking does not make any difference for the enduser, but you really have to be careful of what you commit and use if you work on bsd licenselevel in one of those projects (one unnoticed wrong commit, and gone is the bsdish license of the original codebase, unless reverted). On the other hand GPL or CDDL code authors can use code from more liberal licenses and integrate it without too much fuzz. Sort of a one way street in licensing. This is nothing big, but it explains also why it is probably good to have yet another jsf implementation from a licensing standpoint as well.

This is nothing big, but it explains also why it is probably good to have yet another jsf implementation from a licensing standpoint as well.

I can buy that. Personally, I'm not all that excited about licenses or debates surrounding them. :) I'm happy with the licensing of the RI and feel comfortable contributing code to it, as well as using it in apps. But, all that aside, congrats to the MyFaces team on all their hard work. "Competing" impls will keep us all honest on our toes. :)

This is nothing big, but it explains also why it is probably good to have yet another jsf implementation from a licensing standpoint as well.

I can buy that. Personally, I'm not all that excited about licenses or debates surrounding them.

Same here as a programmer there seems to be some kind of natural aversion against any legal mumbo jumbo. As it seems the laywers always seem to be the natural enemys of engineers. ;-)

how about if it compared against ICEFaces 1.6 ???

Mhh comparing a JSF implementation against a UI-component framework for JSF seems odd to me!?

Congratulations ! Thanks for the hard work.

Hi I'm using the JSF1.2 RI (1.2_02-b03-FCS) and tomahawk + tomahawk sandbox and facelets without any problems!

What are the benefits over SUN RI? --Mark

I've posted my thoughts about Sun's RI vs. MyFaces 1.2 on my blog. From what I can tell, Sun's RI seems to be a bit better because it allows me to run JSF 1.2 in a Servlet 2.4/JSP 2.0 container. Regardless, it's great to see the MyFaces 1.2 release - well done gents!