Forum OpenACS Development: Re: OCT

Collapse
3: Re: OCT (response to 1)
Posted by Neophytos Demetriou on
Thank you Dave. I've also found http://www.openacs.org/oct-election/ which seems to be up to date.

May I ask what's the current stand on new contributions? I mean, the community seems to be shrinking and I understand that the resources are limited but is OpenACS still aiming to keep up with other projects such as Joomla, ROR, Zend, or to preserve the status quo (read do not break/disrupt running installations --- I see enough reasons that would justify that, this is no trick question, just asking)?

I have a few packages that I would like to contribute other than the Maps package that I have already uploaded. The Maps package could be turned into an OpenACS package more easily than some of the other packages. I understand some of those packages maybe/are a bit "radical" (read different presentation layer, different persistence layer, some use functional programming constructs in TCL, some are purely XOTCL) and it would help to know what the limitations are.

For instance, I understand it would take a considerable amount of energy and effort just to get the presentation and persistence layers into the core. Those packages are required from applications such as the MediaBox which I've already posted some screenshots here: http://www.openacs.org/forums/message-view?message_id=2551843 (more applications include blogs, personal&social bookmarks, etc)

What is the aim of the project now? I understand these are no longer the glorious days of ArsDigita/ACS but it's good to know what the situation is.

So that you know where I'm coming from... I've been asked more than a few times the question of X Vs OpenACS where X \in {Joomla, ROR, Zend, etc} and I'm obviously the last person that anyone needs to convince about TCL/XOTCL, PostgreSQL or AOLserver/NaviServer but I have to admit it is no longer a straightforward answer when it comes to the framework/suite/apps.

Collapse
4: Re: OCT (response to 3)
Posted by Gustaf Neumann on
Neophytos,

I can only give my personal view on the situation, i am not sure, how many people in the OCT have the same view on the matters.

In the very first beginning OpenACS went out to be something like a high level operating system for Web-Applications on a much higher level than aolserver: factor out common code, that has to be programmed all over in various applications into a common core, provide a complete persistent storage based on relational databases. In this respect, the goals are quite similar to e.g. the SAP kernel.

In this area, OpenACS was and is a success. We have various projects out there, ranging from e.g. DotLRN over ]po[, quest to various forks using the same kernel and the same data model. I would also count the xo* packages (xowiki, content flow, s5, request broker, the work at Gallileo and Viaro, etc.) as an own project. There are many other applications out there that share a certain part of the same kernel, you know them most likely better than me, since you are longer part of the community. Many OpenACS installations are essentially forks of the common code.

The various projects have different goals. For example ]po[ has the goal to be a web-based SAP on the longer range (my impression), while DotLRN has a strong emphasis on accessibility, and xo* has the goal to develop object oriented high level "end user development tools" to improve the code reuse and reduce the need to write code for various applications.

On top of this, there are various large and successful applications (e.g. like learn@wu) which have their own goals and needs. As another group of forces, we have several companies with different kind of goals. Frank made a nice presentation at the last OpenACS conference about different goals and needs of "service companies" vs. "product companies". While service companies have the goal to ship a clean and lean kernel (push everything dangerous out), and to provide services to make from this kernel useful software, the "product companies" need a certain featurism to prove the software is ready to compete now. They care much less about what's happening under the hood.

From this point of view, there is no single answer to your question, what "the aim of the project is now". Maybe this is a major difference to the "glorious days of AD" or to Joomla etc. We have various projects. Since at least 5.2, OpenACS is reduced to the core which is essentially maintained (bug fixes, speed and code improvements, adaptations for e.g. PostgreSQL 8.3, etc., accessibility). Several of the application packages (forums to filestore) are de facto part of DotLRN. OTOH more and more code is thrown out of DotLRN and is be replaced by standard kernel code, which is a good process.

In my point of view, one of the strengths of OpenACS/AD is that professional companies develop code and provide it as OpenACS packages. This strenght is up to a certain point as well a weakness. When this companies have different goals (or being sold etc.), the donated code is left without maintenance. On the longer term, an ongoing commitment to the provided code is essential (like e.g. for the xo* packages) to establish a good code and service quality.

So, if you contribute code (highly appreciated), it is quite early to ask, whether the code will be added to the core. A big question is, whether you can and will continue to work on the code, or whether you see this as an "own" project or part of some other projects.

In the last year, we had about 1200 commits to the OpenACS code repository. So we have some very active developments there. But there is likely to be nobody with the resources to take over the development and maintenance of your code. From my point of view, if you can't continue to work on, then this new code will have to wait for (maybe new) projects to continue, or until some volunteer steps up...

Collapse
5: Re: OCT (response to 4)
Posted by Neophytos Demetriou on
First and foremost I did not suggest that they should be included to the core and surely I have spent more time developing these packages than they need to be maintained. MediaBox is probably the only application which I haven't seen in any other project (there are ExtJS-based efforts that do similar things but they are not as complete, IMO)

Having said that, the presentation and persistence layer are a shift in the paradigm. If there is no chance to be considered then there is no real benefit than to keep developing on my own and just upload (like I did) for anyone to "steal ideas" (in the good sense, as described in the TAO of Programming).

So, in many ways, they were not published under any conditions. The question is whether the community/project is ready to consider something new (disruptive) or not. If one is going to take on its shoulders the extra overhead of standardizing/conforming to some rules, then it better be for some reason. My understanding is that the answer is no, that the project/community cannot withstand anything new/disruptive at the moment.