Feedback wanted. Would you choose JB Portal again?

[wheezer]

We've been evaluating Jboss vs Liferay and Jetspeed. It's hard to find really good reviews or success stories and evangelists. It seems that Red Hat should give JBP a big lift over Liferay and others, but that's speculative. We liked Liferay, but the community seems very quiet, and the liferay company seems more focused on their commercial support, than really pumping their community.

All of the portals have fewer open source portlets than we imagined they would have. Logic says that Red hat's involvement might rev up the portlet development, and probably here at JBP first, but logic has been wrong before. It's been a year and a half. Is that happening, or....?

We could really profit from some feedback from JBoss Portal users who have serious experience with it, and perhaps the other portals.

So if you were given a second chance, would you still choose JB Portal, or go another route? If not, what might you choose instead? We've learned on other sites that questions like this get a lot of response. They can be VERY very helpful to we seekers looking to find the right solution.

Thanks in advance.

[claprun]

"wheezer" wrote:
All of the portals have fewer open source portlets than we imagined they would have. Logic says that Red hat's involvement might rev up the portlet development, and probably here at JBP first, but logic has been wrong before. It's been a year and a half. Is that happening, or....?

Just a personal opinion, not necessarily representing my employer's ^_^: why does it stand to logic that Red Hat would rev up portlet development?

When considering a Portal, should you be looking at the gadget portlets that are bundled and that you will probably not use in your actual portal or at the capabilities of the portal server? Should portal developers spend their time developing portlets or make sure that the actual portal server works properly, scales, is well tested, etc.? To me, the portlet market is completely separate from the portal one. "Interesting" (i.e. portlets that you would use in a business environment) portlets are products with their own lifecycle, independent of portals.

Any JSR 168 portlets should work on any JSR-168 compliant portal. If they don't, then either the portlet is not conforming or the portal is wrongly advertising compliance. So my opinion is that you should choose a portal on its technical merits, not on the set of bundled portlets. But then, that's just me! :)

[wheezer]

Chris,

I complete respect what you're saying about the far more important design, stability and usability of the portal, independent of the portlets. Let me clarify my post and perspective a bit. I suppose my post was geared more toward a view of the portal as a ready-to-run production environment for a large (social) web site. As more of a portal user than a core developer, I am very interested in the existing applications that a good portlet catalog could provide (from any author or vendor).

In most reviews and remarks about many portals, and the promotional materials of some of the portal sites themselves, the issue of off-the-shelf portlet quantity and quality receives far more than a trivial mention. Like us, not everyone choosing an enterprise portal is a large enterprise, able to just crank out the code for a custom portlet to meet their needs. More often, they want to modify something that comes close to their requirements, and go from there.

The jsr168 approach provides a huge step up from the wild, wild, west that is the open source portal world outside the JavaEE sphere. But the advantage of some of those products is a wide array of third party plug ins that can jump start many kinds of projects. I don't think it's a stretch to wonder if such a community of portlets could be fostered in the JavaEE world too.

Liferay gets a lot of attention, mainly because of these available portlets, and it's what attracted us to it. From just a few weeks of experience in their forums, I can comfortably claim that this is what has motivated hundreds of other users, as well. My sense, however, is that JBoss Portal is a better architecture for our longer term goals, so I am eager to know much more about it. Presumably, the portlets we develop there will be reasonably portable.

[claprun]

"wheezer" wrote:

Liferay gets a lot of attention, mainly because of these available portlets, and it's what attracted us to it. From just a few weeks of experience in their forums, I can comfortably claim that this is what has motivated hundreds of other users, as well. My sense, however, is that JBoss Portal is a better architecture for our longer term goals, so I am eager to know much more about it. Presumably, the portlets we develop there will be reasonably portable.

Your points are valid. Then again, if you like Liferay's portlets (and presuming that they are JSR 168 compliant), what is preventing you from using these portlets in JBoss Portal?

Another aspect is, with JBoss Portal 2.6, we are starting to integrate with non-JEE portals. You can for example very easily use Google widgets with a Portal page with support for other (NetVibes comes to mind) probably coming up...

Also, I often see this argument where people want to benefit from an open source community of portlets to modify and use in their projects. How is it though that these people rarely contribute their portlets back to the community? Hint: useful, reusable portlets take time and attention to develop. Most people write a custom solution and don't care about making it reusable. From my perspective, and that's only mine, the return on investment on open-source portlets is very low. First, it's not clear what portlets the community wants. Second, we would probably spend a significant amount of time developing them but they would probably not fit all users needs and the more generic we make them (to try to make them easier to reuse), the more time we would have to spend on them, also running the risk of making them more complex. What do we have to gain from doing this? Increased visibility in the community (which is great) and market share. But how much of this would translate to paying customers? Would it be enough to justify the time we spent on them as opposed to working on the core product (which IS making us money)? Hard questions to answer...

A community doesn't build itself. It's a two way street... We provide an open repository http://labs.jboss.com/portletswap/ for portlets... It is not as great as it could be, and we are aware of that, but like most things, this takes time and, from my point of view, with seemingly low interest from the community, I'd currently rather spend that time on improving the Portal server.

[wheezer]

All true.

I suppose though, in my head, I am thinking about why it is that so many open source modules get built in the non-java, non-jsr168 communities. Naturally, the universe of developers is much smaller here, and their professional demands greater, but still, it seems there should be more interest. I mean hell, it's not just open source portlets I am interested in. I'd happily pay reasonable prices for useful portlets. I just don't see any around--at all. SyncEx advertises all over the place, and I can't get a demo out of them. I'm not expecting much, when I do.

If Jboss had a vibrant marketplace for portlets, I can't imagine independent coders and developers not seeing a relatively easy path to some real revenue without all that much work. I wouldn't even care how generic they were. I'd bet there's a large market of enterprises and vendors who wouldn't either. Sure, generic configurations are nice, but ANYTHING pre-built can be a huge jump-start over building from scratch.

Incidentally, we do plan on probably running the Liferay portlets in Jboss. Still just getting our bearings.

Thanks again for all the feedback. This has been very helpful. (Hope others feel free to toss in some cents :)

[claprun]

"wheezer" wrote:

I suppose though, in my head, I am thinking about why it is that so many open source modules get built in the non-java, non-jsr168 communities. Naturally, the universe of developers is much smaller here, and their professional demands greater, but still, it seems there should be more interest. I mean hell, it's not just open source portlets I am interested in. I'd happily pay reasonable prices for useful portlets. I just don't see any around--at all. SyncEx advertises all over the place, and I can't get a demo out of them. I'm not expecting much, when I do.

Maybe it's because JEE developers spend their time in business integration of complex enterprise services and it doesn't make much sense to provide the portlets that were developed in that context to a larger audience? Maybe it's because non-jsr168 portlets are usually not much more than a bunch of javascript/html bundled together and which doesn't connect to enterprise services, hence can be developed fairly quickly? I don't really know. I'm not a portlet developer. :)

"wheezer" wrote:
If Jboss had a vibrant marketplace for portlets, I can't imagine independent coders and developers not seeing a relatively easy path to some real revenue without all that much work. I wouldn't even care how generic they were. I'd bet there's a large market of enterprises and vendors who wouldn't either. Sure, generic configurations are nice, but ANYTHING pre-built can be a huge jump-start over building from scratch.

Don't get me wrong. We would love to have a vibrant marketplace for portlets. My guess as to why it's not there yet is that most portal deployments occur in an enterprise context, with custom developed portlets that cannot be reused outside of the context they were developed for.

What kind of portlets would you like to see?

"wheezer" wrote:
Incidentally, we do plan on probably running the Liferay portlets in Jboss. Still just getting our bearings.

Thanks again for all the feedback. This has been very helpful. (Hope others feel free to toss in some cents :)

Glad I could help! Once again, though, this is my personal opinion so don't construe it as JBoss' stance on open source portlets! ^_^

[theute]

I really hope we can get portletswap to a *2 ways* interaction website for portlets.

Everything i read in this topic are very valid :)

[ebrennan]

I am very much in the same boat. My organization is pushing hard for a portal to be up and running. Most of the portlets will be custom in house stuff. I am focusing on getting Active Directory up and running correct. I have been able to connect and load the users but not map the groups over.

I came across Liferay last week and decided to give JPortal a break. I like the path that Liferay is moving in. They support a wide mix of enviroments. They have a wide base of wikis and documents. There admin portlets are a little more developed than JBP.

I am going to continue to develop in both until I have one up and running to our standard.

I also looked at Plone which had some great reviews but did not like the fact that not based on Java.

[claprun]

There admin portlets are a little more developed than JBP.

Care to elaborate?

[bvogt]

to answer as short as possible? ...yes!

Nearly one year ago we compared JBP against jetspeed2 and liferay on behalf of our customer. Disappointed about liferays design and (professional support - they didn't find answer within one month) in contrast to the clearly structured design of JBP we decided to customize JBP for our customers needs.

Yes, a group feature is currently missing but it's being discussed already by the portal team...

For our needs we have improved JBP 2.4 by that feature and provided it in jira. In contrast to liferay extending the portals core was more or less straight-forward...!

Attaching AD to the portal has been discussed in detail in this forum, a wiki page should be available... just search for it.

Our decision for JBP wasn't driven by a quick setup only but for a sophisticated core design, extensibility, intuitive understanding and last but not least the support of the portal team within this forum...