Forum OpenACS Q&A: Summary of the Sloan - Berklee dotLRN meeting

The Sloanspace V-2 team and the Berkleemusic.com team met on May 7. Berklee and Sloan are both using and expanding .LRN (dotlrn) to build educational websites. We know there is much interest in the OpenACS community so we are posting some of the highlights of the meeting.

.LRN will benefit most from the difference in our projects.

  • Sloan is using Oracle, Berklee is using Postgres.
  • Berklee's Phase I is supporting distance learning and creating a tool for structured course delivery. Sloan’s Phase I is supporting a brick-and-mortar university with online community tools.
Both sites are launching by August 2002, insuring that .LRN will support a broad spectrum of functionality by fall.

We both want to keep the general OpenACS community informed about specific areas where Sloan and Berklee will be focusing in the next few months.

  • Berklee and Sloan will be working together expanding the Survey Package.
  • Berklee is developing a basic LMS capability to deliver existing HTML course content.
  • Berklee is currently evaluating chat solutions.
  • Sloan will be creating a Photo Book Package to display and manage students' photos and profile info.
  • Both groups will be doing extensive customization as well as working on general .LRN UI and functionality.
We welcome input and collaboration from the OpenACS community. The contact people are:

Sloan: Caroline Meeks cmeeks@mit.edu
Berklee: Michael T. Serio - mserio@berklee.edu

Collapse
Posted by Don Baccus on
Are there any plans to, say, actively coordinate this work with the development efforts of the OpenACS community? Or are you folks going to continue to plug along in the "we'll gift you when it's done" mode?

Just the way you state the relationship

We want to keep the OpenACS community informed
and
We welcome input and collaboration from the OpenACS community.
indicates that your mindset is that you're not part of the OpenACS community. This feels a lot like the old aD days, except for the minor difference that without actively coordinating with the community minor issues like "hey, the core's changed and our new stuff doesn't run with it" might crop up.

We've got a lot of aggressive plans, and while it isn't clear what the timeframe for implementation is, there's going to be overlap.

And I plan to try to work to set down some standards for how to use the basic architectural components like the object system. As far as permissions go, at minimum we will want to come up with best practices, so to speak.

And I want new development to use the form builder whenever possible and am working on a new tool that will help simplify its use (using it means forms will be templatable).

Open Force, while criticized by the community for keeping things under wraps, did at least communicate with me - the OpenACS project leader - and did solicit opinions from a few others.

I really dislike the standoffish approach that was taken with dotLRN development. I think that this newly announced effort is a great opportunity to increase cooperative efforts. You already have two groups - Sloan and Berklee - cooperating, why isn't the rest of the community invited?

And, no, this isn't about Don being on a power trip. During the last several months several of our best developers have hounded me over the closed nature of the original dotLRN effort. I forsee myself being hounded over and over again if this is the model that's followed in the future.

-Note the original post was from both the Berklee and Sloan teams and was seen and approved by everyone on the teams, hence the formal language. The following is only my personal opinion.

Dear Don,

I'm very sorry you feel you have not been consulted.  I was under the impression that Sloan was contracting you do the very first non-Open Force .LRN package (Homework Drop-Box) specifically so you could use it to model best practices.  My understanding was that you had already received the spec. Sounds like there has been some confusion.  Please call me voice tomorrow.

I think I do speak for both institutions and the programmers when I say we are 100% behind and wish to be a part of OpenACS, however, the "We" in the above post was not ".LRN", but the two specific teams of developers who are using .LRN to create specific client sites, with concrete defined requirements,  and very strict deadlines. If you read this announcement closely, you'll notice that despite the fact we are less the 2 miles apart, both doing web sites that support classes for universities, the only piece we found to really collaborate on in the short term is Survey.  Basically we are at the point of saying. "We both need survey and neither team has started on it yet."  "We" are not the OpenACS community. "We" are not .LRN. "We" are not ArsDigita.  "We" are two separate teams focused on our deadlines and deliverables.

After we meet these deadlines, these two specific websites will "gift" this code back to .LRN and then we as individuals and community members will certainly hope to work with you and the rest of the community in building and improving .LRN and OpenACS.

When I was at ArsDigita, I tried and completely failed to explain to my management that what we had really built for Sloan was an intranet.  As programmers it's obvious to us that these packages have extensive general usefulness across industries.  I learned the hard way that this is not obvious to business people.  They think in terms of Industries and Sectors.

The way I see it is .LRN is a "brand". It is a distinction we can make while marketing to help us communicate with educational customers that we understand their problem and have a proven tool to help them solve it.  From a technical point of view it is my individual opinion that it is best for all of us for .LRN and OpenACS to be as close as possible.

Again, these are my personal opinions and vision, they do not necessarily reflect Sloan, Berklee, Open Force or .LRN!

Collapse
Posted by Alfred Essa on
Don, I would like to take a crack at your concerns from MIT-Sloan perspective. First, please understand that this is our first venture into the opensource community and it's a learning experience. Opensource is a weird admixture of anarchy and tyranny.

Second, both OpenForce and we are trying to strike a balance between coming up with code, on the one hand, that will be benefit the community and, on the other hand, will meet Sloan's needs. An important reason why the project has been delayed is that OpenForce is trying to do it's best cleaning up a lot of crud code in ACES and trying to make the application less Sloan-specific. At the same time, we have been putting a lot of pressure on Ben and OF to make sure that Sloan's needs and timetable are met. While in the best of all possible worlds everything would be transparent to everyone, in this the actual world we don't have the luxury of maintaining n conversations.

Third, we began collaborating with Berklee because they are just across the river and recently committed to dotLRN. They also have a "real deadline" to meet. As Samuel Johnson said, "Depend upon it, sir, when a man knows he is to be hanged in a fortnight, it concentrates his mind wonderfully." At the meeting we concluded that it would be a good idea to let the community know where we are and also what further work we envision. Our intention is to establish a mechanism for RFCs and that was the spirit of Caroline's posting. But please recognize that we also have deadlines to meet and may not be able to keep the process as transparent as we would like.

Let me conclude by saying that the intention here is not to keep the development closed and inaccessible. We are all just in a mad dash to meet the needs of our clients. Our first priority is to the folks who the write the checks...

Collapse
Posted by Jon Griffin on
In support of Don (and Sloan/Berklee i.e. S/B) I agree with all sides!

I understand the business perspective and also the history of ARSDigita/OpenACS. I truly believe that OpenACS will be a better product eventually due to the work of the S/B teams.

I also completely understand the concerns of Don and the OACS community. Hopefully any core changes you are about to make will be made public before they are committed in stone (of course, you don't have to do that). Then we won't have the reinvent the wheel problem that happens (and happened at AD) frequently when great ideas get implemented by multiple parties independently.

There are certainly some very weak aspects in OACS as inherited by AD, but there are many strengths also and certainly OACS is a much more stable product in the community than it ever was at AD.

I look forward to any enhancements S/B creates and also any new packages. Thanks for keeping us in the loop at all.

Collapse
Posted by Don Baccus on
First of all, my post probably comes across as sounding more harsh than I meant it to be.  I appreciate the information in Caroline's original note and the fact that Sloan has hired her to, in part, communicate with the community.

Let me think some more about the actual message I'm trying to convey and I'll post more tomorrow.  The basic issue boils down to the fact that we've got a bunch of companies, Sloan, Berklee who are all starting to run with the code base, implementing important projects right-and-left.  I'm in the midst of such an effort myself.

Last year it felt like there was real focus on getting our code base solid enough that folks could build industrial-strength sites on it.  The community of hackers here were marching in one direction.  Now it's like Brownian motion ... things just seem ... relatively uncoordinated.

The basic question, then, is how do we focus on making sure the underlying code infrastructure moves in a well-defined direction that dovetails with the plans of others?  Or at least move to a space where those of us involved with building pieces that are generalized enough to fold back into the toolkit are able to coordinate with each other?

Anyway ... let me sleep on this and I'll try to coalesce my somewhat vague thoughts into something more concrete.

I very much appreciate the work that Sloan's sponsored, I think you folks know that.  I hope there's no doubt about that.

Collapse
Posted by Rafael Calvo on
First, I want to congratulate Sloan and Berklee for their commitment to innovation. As Al said, opensource  "is a weird admixture of anarchy and tyranny", and I am very aware that this means risks (and opportunities) for their clients. Universities as well as many other large organizations need to manage risks, together with budgets and delivery dates and be strict about them.

These two universities are always at the forefront of technology (MIT) and arts (Berklee), and this represents a great opportunity for the openacs community. If they are successfull the whole opensource community will be successfull.

The OACS community will soon have one of the top Business Schools and one of the top music schools in the world using dotLRN/OACS. We will have the largest enviromental movement in the world (Greenpeace) using OACS and we have companies like Siemens that use ACS and could eventually use OACS as well. These are great news!

I believe the community is very aware and happy for this.

I also agree with Caroline and we must think of dotLRN as a "brand" of OACS. It is the killer application that the project needs and will help develop.

On the other hand, the developers I have talked with are trying to find ways of collaborating with openforce, or at least of not reinventing the wheel. Since OACS is very modular the best way I found of doing this is building things that other people is not (previous consultation with Don and Ben). I think that is a good way to work for all. If people working in openforce and Berklee tell Don (or post it in the bboard) what they are working on, we can try not to overlap.

Finally, there are other universities that would like to be more involved in the dotLRN project, and would be able to add value to it. Al and Caroline had planned to be more open about the dotLRN governance, but I haven't heard much about it. How can other universities participate more in the "strategy" aspect of the "brand"? I am consciously trying to separate this from the purely Software engineering aspects, since I don't think that IMS compliance, learning objects, etc,.. are issues that the OACS cares much about. How can we collaborate so other universities start using it?

Collapse
Posted by David Kuczek on
I am very happy that Berklee is tackling the chat solution!

Some comments from our evaluation experience:

1. The solution should run on a dedicated port AND through http-tunneling on port 80 (corporate firewalls)

2. The chat should be scopable: groups/private/public

3. The chat should be moderateable: A moderator/expert can accept/deny posts.

FOR THE FUTURE:

4. The chat could support voice, video capability

5. The chat could support whiteboard capability

6. The chat could support co-browsing

-------------------

It looks like Jabber could be the answer:

1. There is a JEP for http-tunneling (http://www.jabber.org/jeps/jep-0025.html)

2. Jabber supports scoping

3. "The existing open-source conferencing implementation does not include this functionality. Dave Waite spent a bunch of time trying to come up with a new draft for conferencing protocol which included all of this type of functionality, but that effort has been pretty "stagnant" for a while. Lots of people are interested in having better/more-robust conferencing stuff, but no one has really stepped up to the plate and put together a JEP yet."

(http://mailman.jabber.org/pipermail/jdev/2002-April/011545.html)

4-6: We didn't really check on this yet, but you can check out http://www.groove.net to see this functionality at work...

One really import point for us in Europe is the support for multiple languages. Please note that most (if not all) major commercial learn platforms made the mistake of not being localizable from the start, and are paying for this now (figuratively and literally). I just wanted to mention this, because I doubt there are many foreign language courses at Sloan or Berklee, internationalization is easy to forget about when on a tight schedule, and working on it after the fact is a lot harder than when it is considered from the start (not to mention that making something support multiple languages at a later stage requires more coordination and communication than if it is done from the start).

This is important.

Finding out that OpenACS is the platform Greenpeace has chosen, eases my multilingual concerns on the OpenACS side (although it does not eliminate them). I just hope that the multilingual features of OpenACS and dotLRN will "dovetail" as so elegantly stated by Don. Obviously, the best solution would be one that is independent of dotLRN and built into OpenACS, making multilingual support a snap.

Can somebody that knows both systems and has experience in internationalization please comment on this? Who is working on this?

Collapse
10: OACS and .LRN (response to 1)
Posted by Caroline Meeks on
I actually think there is very little that can be done for the education market that doesn't have applicability in other markets.

Take for instance the little Homework Drop-Box Package.  All it is file-storage with specific permissions and flow so that students can upload homework, but can't see other's homework and teachers can download homework and upload comments/graded papers.  Seems pretty education specific?

Commonly in manufacturing facilities production workers are encourage to submit suggestions and if those suggestions save the company money they get a percentage of the savings.  Seems like pretty much the same code to me!

Now try to go into a big manufacturing company and tell them you can solve their problems because you have a "Homework Drop-Box".

Good programming sees the similarities between problems. Good sales tells the customer they are special and you understand their problems.

They are 180 degrees apart.  No wonder sometimes the "Tone" of marketing material rubs the technical people the wrong way.

From a technical point of view we absolutely have to be sure .LRN and OACS are compatible, hopefully to the point of being almost the same thing. But please be tolerant of the fact that we will be trying to create a separate image that says education.  This .LRN centric language will probably get worse not better in the next year or two. This separation does not mean that technically we don't consider ourselves part of the community.

I personally am committed to reducing duplication of work and keeping people informed.  I've been emailing all the people doing any work for educational institutions. I attended the last OpenACS social.  I have visited the Open Force and Ybos offies and I spent a bunch of time with the Musea people at the social.  I've been AIMing the engineering team at AD.  I'm having lunch with Philip next week.  I posted about which major packages we are about to start work on.  I'd love to go out and meet Dan if anyone wants to buy me a plane ticket J.  Believe me, the last thing in the world I want is to have the OACS and .LRN cores not be compatible.

Two other things:

- Survey Builder:

https://openacs.org/bboard/q-and-a-fetch-msg.tcl?msg_id=0002Qz

- Chat:

The chat should be an applet and not a standalone program that users would have to download and install...

Collapse
12: Internationalization (response to 1)
Posted by Caroline Meeks on
Hi Carl,

We have two parties that are actively planning non-english .LRN educational sites.  Malte you have spoken to the other is Rocael Hernandez who did one of the first ACES implementations.

Summer 2001 - Viaro launches pilot in Spanish with ACES at UNIS, Universidad del Istmo in Guatamala.  Contact: Rocael Hernández mailto:roc@viaro.net

January, 2002  - Vario launches a full system for UNIS, Universidad del Istmo in Guatamala, serving 1000+ students in Spanish.

I agree that Greenpeace may end up doing some of the best work in this area and I'll send the person I've been talking to there a copy of your post.

I totally agree with you that multi-lingual support and a less US centric approach will be strong selling points for .LRN.  It will be extremely helpful for the project if some of the next phase of adopters are from different continents and I am very excited that we already have interest from South America, Europe and the Philippines. True internationalization takes into account not only language but cultural differences, such as does a week start on Sunday or Monday.

One of the Sloan team members, Tracy Adams, will be in Germany later this month. I hope that you can meet while she is there.

Collapse
Posted by Don Baccus on
I'm going to talk about one very specific issue which needs addressing.  Perhaps in raising it I can help Caroline and Al understand where I'm coming from.

dotLRN will not run on the OpenACS 4.5 release that will be issued in the next week or so.  I know that Sloan's goal is to release dotLRN "sometime" (when?).  I assume that it would be in Sloan's best interest if dotLRN, when released, will run atop a named OpenACS 4 release.  "Grab the nightly tarball built on July 4th, 2002" probably doesn't cut it as a named release.  Just seems too unprofessional ...

Sloan could cut their own release of OpenACS 4 and bundle it together with dotLRN.  That risks a unintentional fork in the codebase, though, due to the temptation to fix bugs and all that.  Note that ACES was a fork of ACS 3.x and the codebase wasn't 100% merged until recently, by ex-aDer and OpenACS community member Malte Sussedorf.

Now ... if Sloan does see value in releasing dotLRN to work on a named release ("4.6" for convenience), how does Sloan see that becoming a reality unless Sloan actively coordinates with the community?

In particular with me, the person the community expects to more or less coordinate production of that release?

Ben has started talking to me about this issue (I told him we should start floating 4.6 ideas as soon as I get 4.5 out, for those of you who think we plan to hold this discussion in private).  My impression, though, is that he's doing this on his own initiative.

As part of this, I expect that Ben or someone will inform me of Sloan's planned schedule through alpha, beta and final releases.  I hope that *someone* on the Sloan side - and if Sloan's appointed Ben to do this, fine, but how about saying so explicitly? - will inform me of that schedule so we in the community can decide what a 4.6 release should look like.

Also ... I think it's clear that there are going to be some minimal community expectations for an OpenACS 4.6 that can serve as a named foundation for dotLRN.  While dotLRN proper won't support Postgres right off, if the new portals package is going to be part of OpenACS 4.6 - and that's where it belongs, without doubt - it will have to support both Oracle and Postgres.  Code that doesn't won't make it into an OpenACS 4 release, pure and simple.

So we need a final determination if OF or the community will do that work.  This isn't Sloan's responsibility, but because it impacts "OpenACS 4.6" it has the potential to impact dotLRN's release.

I'm really talking nuts-and-bolts stuff here.  I wonder, for instance, if Caroline was fully aware that dotLRN won't run on OpenACS 4.5?  Being "in the know" at this level would be a consequence of closer coordination between the OpenACS project proper and projects like dotLRN.

They aren't the same project, and shouldn't be.  I'm certainly not lobbying that the OpenACS project, rather than Sloan, drive dotLRN.  Sloan has specific needs that must be met and must have a deployed solution that fits those needs.  Fall Term (Sloan's planned deployment) is just around the corner, in software engineering time.

But Sloan also has a broader vision than internal deployment.  And *that's* where some level of coordination at a nuts-and-bolts level seems very necessary.

OK that's post #1.  I've got one more :)

Collapse
Posted by Don Baccus on
* Berklee and Sloan will be working together expanding the Survey Package.
Let me take this as a starting point ...

Recently one of Michael Feldstein's acquaintences, a fellow named Dean Des Rosiers, e-mailed me to tell me he has broad experience developing web-based knowledge management systems, and that he wanted to start an effort to provide such tools for the OpenACS 4 community. He said he felt that a good place to start would be a "general ratings" package.

I suggested that he post to the forum, as I know there's interest here in such things.

In short order he

  • Got pointed to a couple of examples of specific things folks had done with either OpenACS 3 or 4 - a source of ideas for Dean
  • Got a tarball from Lars Pinds of a simple one-dimensional rating package he'd put together very recently for his own site, which Dean may very well be able to use as a base for his more general "general ratings" package
  • Some answers to questions about how to structure the package to best fit within the OpenACS 4 framework
  • etc

Now contrast this with how Sloan typically approaches module design. Sloan's needs are spec'd internally, people like myself (the Homework dropbox, as Caroline mentions, and it's really Furfly not "me"), OF or others are invited to make a proposal, and off we go.

There's nothing intrinsically wrong with this last approach, and certainly since Sloan's spending the money and is deploying the result in a real-world situation the institution must ensure that its needs are being met (and since I've worked with Sloan in the past, I think Sloan understands I am very serious about making sure their needs are met when I do contract with them).

But think about the survey package for a moment. Berklee's involved so the above scenario is slightly modified. Two institutions must be satisified.

But imagine for a moment that Sloan and Berklee raised questions about the survey package here first, rather than later on. "It falls short in the following areas, we have these needs that aren't being met" etc.

What might happen?

Well, what might happen could be that

  • Others in the community have been working on similar enhancements that could be shared and serve as a starting point
  • Others may point out other shortcomings that Sloan/Berklee haven't thought of
  • Folks may have suggestions for improvements in the user interface, etc
  • It might be met with total and stunning silence on the part of the community. It might be that no one else cares about the survey module.

What would be in implications of floating ideas for comment in the beginning of the process, then? Certainly no obligations on Sloan's part. People might float additional interesting ideas that Sloan/Berklee have no interest in funding. On the other hand, you might just see an idea pop up taht would make you folks say, "hey, wish we'd thought of that, let's do it!"

Being a more active participant in the community, then, has potential benefits for both sides. It seems that every few weeks, someone pops up here asking about "OpenACS 4 vs. ACS 4.2", or "OpenACS 4 vs. Java ACS", etc. Each time, at least one respondant says something like "one of the most important things about OpenACS is the dynamic and helpful community"

Sloan already gets some of the benefit from such feedback, because Sloan has contracted with folks like OF and Furfly. Speaking for myself - and Dee Dee would confirm this, I'm sure - specs you guys send my way generally come back with suggested changes. I make those suggestions with a view towards making the feature work better for you folks. For instance, my response to the Homework package will include two or three suggestions.

Why not get the benefit of even wider community feedback? Why limit it to folks who might be interested in formally contracting to do the work?

Of course, there are all sorts of questions about "process", in other words I personally wouldn't want to see more open discussion about design and plans result in Sloan (or Berklee or anyone else) getting bogged down in endless overhead.

Collapse
Posted by Don Baccus on
Finding out that OpenACS is the platform Greenpeace has chosen, eases my multilingual concerns on the OpenACS side (although it does not eliminate them). I just hope that the multilingual features of OpenACS and dotLRN will "dovetail" as so elegantly stated by Don. Obviously, the best solution would be one that is independent of dotLRN and built into OpenACS, making multilingual support a snap.

Can somebody that knows both systems and has experience in internationalization please comment on this? Who is working on this?

If you're interested in a lengthy response to this, how about re-posting it as a new thread? It's a good subject for discussion.
Collapse
Posted by Don Baccus on
I agree that Greenpeace may end up doing some of the best work in this area and I'll send the person I've been talking to there a copy of your post.
Even more reason to start a new thread. Greenpeace has contracted with me to finish up the basic site, and I've been working closely with the folks there over the last couple of months. I'm sure Bruno Mattarollo (the tech contact at GP) would be interested in joining myself and others in a discussion about this topic. Lars Pind has been involved, too, and if he sees the thread he may join in.

So ... would one of you two like to start a thread?

Don,

I guess I don't understand in what way we haven't done exactly as you suggest about Survey.

We posted our intentions to work on it before we even have a spec written.  The community knows about this project before a single line of code is written, before the requirements are even done and certainly before a contractor is picked.

I've been talking with the Survey Builder group out of UCLA, I've been talking to Ybos, I've been talking to Open Force.  I've emailed my draft, which is hardly a spec at this point, to someone from UMass Donahue who emailed me in response to this post.  I talk frequently to Michael Feldman. I've also gotten really good ideas on this subject in the past from the Berkman Center at Harvard.

The community is being involved before the draft of the spec is even done.  Your posts are not exactly acting as positive reinforcement.  Exactly what more do you want me to do?

Berklee adopted dotLRN as soon as it became available as an alternative to a Cold Fusion product that has failed to meet our needs. We're grateful to MIT for providing the dotLRN platform and the OpenACS community for a solid foundation to build the next generation of our web services. We bring great enthusiasm to this project and a commitment to open source software.

We will do our best to be good OpenACS community members but at the same time must weigh this against real deadlines - launch in less than 3 months. Our goal is to figure out an effective way to collaborate so that we can avoid duplicating our efforts as much as possible.

Collapse
Posted by Don Baccus on
I've been talking with the Survey Builder group out of UCLA, I've been talking to Ybos, I've been talking to Open Force. I've emailed my draft, which is hardly a spec at this point, to someone from UMass Donahue who emailed me in response to this post. I talk frequently to Michael Feldman. I've also gotten really good ideas on this subject in the past from the Berkman Center at Harvard.

The community is being involved before the draft of the spec is even done.

Well, this is news to me. First I heard of it. It's welcome news. That's a start. If you folks post a list of new features that you intend to implement, once you get that far, then that would be absolutely fantastic. We could even give you your own dotLRN forum if it would help.
Your posts are not exactly acting as positive reinforcement.
How can I give positive reinforcement when I'm kept in the dark?
Exactly what more do you want me to do?
A personal heads-up to the OpenACS project manager - me - would be useful. Keep in mind that folks ask me questions. "what's happpening?" "what's the plan?"

My stock answer - "I dunno, no one bothers to tell me" - gives folks little satisfaction.

So maybe a commitment to keep the OpenACS Project Leader informed in a broad, general way as to what the future looks like from Sloan's perspective? Not me, personally, but whoever fills that role, which happens to be me at the moment?

From a technical point of view we absolutely have to be sure .LRN and OACS are compatible, hopefully to the point of being almost the same thing. But please be tolerant of the fact that we will be trying to create a separate image that says education. This .LRN centric language will probably get worse not better in the next year or two. This separation does not mean that technically we don't consider ourselves part of the community
There's nothing wrong with this at all. In fact, the community has anticipated this in discussions held last year. We've talked about the notion of "vertical apps" and the fact that dotLRN is our first instance built on the OACS 4 framework. We've talked about branding as being important in the "vertical app" space. People have talked about other branded vertical apps that might evolve in the future. A replacement for Raiser's Edge, for instance. The follow-on to e-base is currently planned to be such an OpenACS vertical app, and if this happens you can bet your last dollar that they'll brand and promote the result in much the same way Sloan will do with .LRN.

So I doubt you're going to see opposition to your efforts to market .LRN as a branded app.

We could even give you your own dotLRN forum if it would help

Given the interest-level and expected high-level of discussion activity with regard to .LRN, a separate bboard would likely be very helpful.

Caroline, Michael, and Al, it might be possible for the Sloan
and Berklee to realize some more synergies with the community
(and make folks on the outside feel more involved and
energized) with a relatively small effort. Here's an example of
how things could go with something like the survey
enhancements:

  1. Berklee and Sloan decide they are going to enhance the
    survey package.
  2. You spec out your requirements.
  3. You do a quick search of the bboards to see what
    conversations have taken place that might be relevent to their
    goals (like
    https://openacs.org/bboard/q-and-a-fetch-msg.tcl?msg_id=0002Qz&topic_id=11&topic=OpenACS, for example) and incorporate
    any ideas that you like
  4. You post a message to bboard saying, "We're going to
    enhance surveys. Here are our high-level functional
    requirements and rough spec. Obviously, we're going to have to
    do this in a way that meets our needs, but before we start
    building, does anybody have any ideas/suggestions/requests for
    us to keep in mind?"
  5. You read the suggestions and take the ideas you like.
  6. You post a message to the thread saying, "Here's what we've
    decided to do. We'll let you know when we're ready to release the
    code.

If you want to see a good example of this process in action,
then check out this thread:
https://openacs.org/bboard/q-and-a-fetch-msg.tcl?msg_id=0004HO&topic_id=12&topic=OpenACS%204%2e0%20Design

I think you'd find the return on your investment pretty
substantial. For a couple of hours' worth of effort, you'd get some
good ideas, substantial good will from the community, and very
likely some folks who are willing to enhance what you're building
and return those enhancements to the community. You're not
giving up control of your own process, priorities, and schedule,
and you're not investing a large amount of time in writing up docs
that you wouldn't have to write up anyway. You're asking for input
but you're not promising that you'll do anything unless it meets
your needs. You might get increased email from excited folks
who are eager to get their hands on your code, and you might get
a little grumbling from folks who aren't happy with your
design/feature choices, but you'll probably get that no matter
what you do anyway. Unless I'm missing something, there's
almost no downside.

Just a thought.

a dotLRN discussion board is a great idea. And BTW, thanks to
Caroline, Al, and Michael for their update and quick responses to
the thread.
A personal heads-up to the OpenACS project manager - me - would be useful. Keep in mind that folks ask me questions. "what's happpening?" "what's the plan?"

My stock answer - "I dunno, no one bothers to tell me" - gives folks little satisfaction.

So maybe a commitment to keep the OpenACS Project Leader informed in a broad, general way as to what the future looks like from Sloan's perspective? Not me, personally, but whoever fills that role, which happens to be me at the moment?

I think this is a wonderful idea Don. People in the educational vertical come to me with the same questions. I had no idea you were doing the globalization work for Greenpeace and I'm very interested.

Why don't we commit to a voice meeting at least once every two weeks, after which we can jointly write minutes to the Bboard, perhaps even starting a few different threads to keep discussions organized.

As other Vertical Apps start to have multiple organizations collaborating, we can include one representative from each vertical in a conference call.

  1. Berklee and Sloan decide they are going to enhance the
    survey package.
  2. You spec out your requirements.
  3. You do a quick search of the bboards to see what
    conversations have taken place that might be relevant to their
    goals (like
    https://openacs.org/bboard/q-and-a-fetch-msg.tcl?msg_id=0002Qz&topic_id=11&topic=OpenACS,
    for example) and incorporate any ideas that you like
  4. You post a message to bboard saying, "We're going to
    enhance surveys. Here are our high-level functional
    requirements and rough spec. Obviously, we're going to have to
    do this in a way that meets our needs, but before we start
    building, does anybody have any ideas/suggestions/requests for
    us to keep in mind?"
  5. You read the suggestions and take the ideas you like.
  6. You post a message to the thread saying, "Here's what we've
    decided to do. We'll let you know when we're ready to release the
    code.

The original post was intended to inform the community we were at Step 1.

The orginal post was intended to inform the community we were at Step 1.

That's great, then. I look forward to step 4 with great enthusiasm.

Collapse
Posted by Don Baccus on
Michael - thanks for succinctly summarizing a process, you've laid it out better than I did.

Caroline - what do think about putting up a forum here? I see two "yes" votes from "stakeholders" above (I put Michael in that category), how do you feel?

Why don't we commit to a voice meeting at least once every two weeks, after which we can jointly write minutes to the Bboard, perhaps even starting a few different threads to keep discussions organized.
The short answer? Time. As you know, my OpenACS role is strictly voluntary and uncompensated. Unlike Philip, I don't have any $7M targets to sue so am stuck working for a living :)

So I can't really commit to the time. In fact, I shouldn't be spending time on any of this today, the Greenpeace schedule is extremely tight.

On the other hand, it doesn't have to be me. What we need is a conduit of information who understands both the community here and Sloan's internal needs and workings. I would put Michael Feldstein on the spot but if I do, I may not live to see the morning :) As dotLRN rolls out, usage grows, it learns to speak Postgres ... I think chances are excellent that someone will step forward to take on the role of trying to coordinate OpenACS releases and changes with dotLRN extension needs.

Of course even better will be the day when the toolkit core stabilizes to the point where dotLRN releases don't get tied to specific OpenACS 4 core releases. But I think we're a couple of toolkit releases from that point. Check out Lars Pind's post today on reorganizing the object system, over in the 4.x design forum, new post to an old thread. I think that will remove any doubt that there's a technical need to loosely couple our future schedules.

And if, as you say, you look at the survey package being at step one of Michael's process ... then perhaps we can take this package as an experiment in involving the community in the planning process? We' d all learn from it. I'm aware that open planning of this sort may not work for Sloan, i.e. be too unwieldy. I don't expect that to be the case but I'm not a starry-eyed idealist totally detached from reality. Let's try it and see what happens ...

Don wrote:

On the other hand, it doesn't have to be me. What we need is a conduit of information who understands both the community here and Sloan's internal needs and workings. I would put Michael Feldstein on the spot but if I do, I may not live to see the morning :) As dotLRN rolls out, usage grows, it learns to speak Postgres ... I think chances are excellent that someone will step forward to take on the role of trying to coordinate OpenACS releases and changes with dotLRN extension needs.

As long as everyone understands that I am *not* an OpenACS project leader and that my understanding of the programming issues are limited at best, I'd be happy to help out in whatever capacity would be useful. Caroline, if you would find it useful to speak with me instead of Don for the updates, then I'd be happy to take responsibility for helping post the minutes to the community and for pinging Don whenever it feels like we're onto an issue that might need his direct attention.

I'm not sure if I'm the right person for the job, but I'm happy to do it to the best of my abilities.

Collapse
Posted by Jon Griffin on
This thread has me interested in getting my meeting minutes module ready for OACS. It currently is only ACS4.2 but shouldn't be to hard to update. I just haven't had the time. It has been in use in the educational sector for about 1 year so it is stable.

Then instead of using bboard for something it is not good for, you can use a module made for meeting minutes.

Just to go back to an earlier (and v. v. good point) about the difference between sales and technical presentaqtion of a product. Its absolutely our experience that the OACS loses out in competitive situations through a lack of understanding of how the average middle-manager-come-software-buyer thinks.

Even the simplest connection between seemingly disparate problem spaces can be difficult for commercial managers to see.

Branding is possibly as important as functioning in the commercial world (well perhaps not that important) and therefore is this a time to suggest a general theme/brand for vertical applications etc..

Seems to me the Sloan folks have the start of a good branding pitch there that could be quite naturally evolved into the overall OACS consciousness (did I just use that word??).

I.e.

dotLRN

dotWRK

dotMRKT

and so on......

Although entirely off this topic, is there any case for essentially packaging OACS distributions in these formss, perhaps with more pre-configuration. In this way we present the same excellent package to a variety of industries whilst seemingly offeringa specific solution...

Simon

Collapse
Posted by Don Baccus on
is there any case for essentially packaging OACS distributions in these formss, perhaps with more pre-configuration.
Yes. During earlier discussions (maybe more like 18 months ago! Long long ago!) about "vertical apps", the notion included packaging as well as branding.

In practice I think this can be handled by building a release tarball that includes the acs-core and needed packages, and the custom "vertical app" packages (dotLRN stuff, in this case). The package(s) providing the framework for the app would be flagged as "load at install time".

At that point, depending on the app, various approaches might be taken. For dotLRN you'd probably load all the dotLRN packages automatically at install time, and automount the basic framework packages (dotLRN itself, for instance) after restart. Then after a second restart you could jump right in, defining terms, departments, classes etc. Off you go! That's how I'd package it, anyway ... it can all be done with the APM in its current state without touching the core at all.

At least I think it can :)

It would make a hell of an onsite demo ... assuming you could keep your audience entertained during the lengthy "insert circus" called "ref-timezones" ...

On the issue of marketing and packaging, I recently asked
somebody I trust with experience in enterprise IT which of the
following two statements would go over better:

a) We have the perfect solution that runs on top of a great web
server you may not have heard of

b) We have a great fully-packaged solution

The answer was (b), even though (a) actually contains (a little)
more technical information. The truth is that many buyers of
polished vertical apps *don't want to have to know what's under
the hood.* They want to know that they can install it and have it
work. They may occasionally want to know about data
interchange standards for integration (usually in XML, at least in
the eLearning world). They very rarely want to know how it works.
The technical stuff scares them. If you can tell them that they can
install it, run it, and have it play nice with their other apps without
too much fuss, then that's all they need or want to know.

This is the polar opposite from the current OpenACS community,
which is used to thinking of it from the inside out. We need these
verticals to feel like simple, monolithic apps from the perspective
of the buyer and the typical admin.

Hello all,

Is there any update on when the PostgreSQL version of dotLRN will be available to the public?

Collapse
Posted by Ben Adida on
The PG version will be available on July 1. There are still some
changes going into the core so we want to take our time to do
things right first on Oracle, and then quickly port.