Forum .LRN Q&A: Some thoughts and question
I looked at our competitors (like Moodle, Ilias, iTutor, Typo3) and realized that there is alot of brilliant work out there that we could learn of.
Don't get me wrong. OpenACS/.LRN is already great and the platforms I looked at are not only typical LMS. Also I can see features that we are very strong at and others where we are behind. Anyway. I love our toolkit.
Three things that were quickly apparent to me and where I see great opportunities for improvement are
2. Organisational/hierarchy structures
3. Course Repositories
Integrating better UI with our toolkit shouldn't be too difficult and most probably even worth it from the marketing perspective.
2. Organisational/hierarchy structures
At present we have several decentralized installations. Each for one single institution. But there are many small institutions looking for a something that allows them to run a single installation for all of them. A system that reflect organisational, curricular structures and design but have a single repository for reusable content.
What do I mean by that?
Example 1: A Univesity - typical german universities have several independent schools. Each with more or less 1000 - 5000 students and a central library that provides services for the university and the surrounding region. A course centric platform won't work. Each single department within a school wants to participate in e-learning in a decentralized way. To solve that problem a bad but working solution is to define departments as communities and their courses as subgroups. But what happens after a couple of years? A department will have some thousand members since they one day became members to get access to the departments courses and never deregistered or are even not studying at the university anymore. The performance of the site will decrease with every single day.
2. Example - Several smaller institutions collaborate in e-learning. Like above there is a need to reflect their organisational structure. A new aspect is that some areas should be accessible for all and some areas only for members of one domain. Even more than the above example you will find the requirement for flexible UI.
I tested a community in dotLRN with 2000 students and several large subgroups. Waiting for the community page to render takes ages while small communities with only very few subgroups work well. From my experience currently the platform works very well for course centric but less good for organisation centric solutions.
Now my question is
- why not allow institutions to define their organisational structure (institutions, schools, departments), their subject and course structure or any other structure needed.
- why not allow organisational levels to map to domains (external authentication)
- why not allow each level to define their own master style sheet or even template
This would give us the means to navigation within th organisation from different perspectives and even restrict areas.
3. Course Repositories
Another aspect that needs to be adressed is the lors package. The idea behind learning object repositories is to have one repository for installation. At present each prof. can create his courses which makes sense if the course is class centric. Better would be to have lors as a single service with admin portlets on class level but there should be a reference to the repository not decentralized courses as default. Courses in the repository should be mapped to classes not the other way around (at least as a default).
Another aspect is the fact that courses and repositories are part of an iterative process. Those of you who worked with lors before know what I mean. A professor creates a new course and uploads the course. Some time later he realizes that a flash file or an avi would enhance his course as would adding some more subchapters or reorganising the course structure. Now what can he do? Basically nothing. He has to upload a new course and disable the last, right? I don't even want to go into the performance issues involved by that. It simply takes ages for a course to get uploaded to lors. Any chance to improve that? Other platforms upload courses fast as lightning. Only at item level he can upload a new revision to file storage but why would version control be important for such courses anyway? We need to find a way to allow incremental changes in a course or let's better call it course evolution without inflating the content repository with multitudes of dead content. Any ideas on that? My opinion is that the latest revision is what is important all other shouldn't even exist. How can we make lors faster on the one hand and less rigid on the other so that we allow course evolution?
Any ideas or thoughts on that?
1. completely agree
2. There's a brand new package, dotlrn-catalog which will help to address some of this issues, its independent of .LRN (so you can have you own course catalog), and can be tied to .LRN, a future enhancement will be to get rid of the rigid structure of deparments -> subject -> class, which in real life does not fit all the needs, and use the dotlrn-catalog, and from there address your suggestions.
3. Yes, LORS will change, we had the same difficulties and some work is going on to make it more flexible, I'll let Ernie to answer this question...
Thanks very much for this thoughtful post!
I've been thinking of you since upgrading to 2.1 and finding your "admin cockpit" in the new version: it is a splendid addition! It not only makes a bunch of admin things easier, but it makes admin work fun!
In my new job working for an outfit committed to Blackboard I do my best to avoid dealing with that beast, and when forced, have been driven to a new form of expressive exasperation involving the lifting of both hands to the forehead, a deep scratching of the frown lines with my fingernails, and the making of a simian-like screeching noise that both horrifies my new colleages and provides me with what is otherwise a graceful exit. Being sympathetic types, my colleagues sooner or later visit me in my office whereupon, often enough, I can show them how we do it far more elegantly in Dotlrn. It has gotten to the point that we are now engaged in a long running joke -- particularly as the costs of various add-ons and additional users in BB are like so many lampreys sucking out the institutional marrow -- such that, when faced with demonstrations of wonderful things Blackboard can do with only, say, about 20-40 clicks and the room turns cold blue and everyone suffers premonitions of carpel tunnel syndrome, the conversation inevitably turns to me and the question: "ok, Bruce, so now tell us how you solve it in Dotlrn."
Well, as I am not a real programmer, I don't pretend to say I have solved anything at all and consider myself a freeloader, but as I am interested in solutions and think with Dotlrn I've found a decent one, I do my best to describe how it works with strong (and uncharacteristic) undertones of understatement and humility.
All of this to say that I think Nima poses the absolutely right question among colleagues and friends: that we are having to suffer certain operations which, for their unnecessary and unexplainable duration, we characterize as "sticky".
Right next to the sluggishness of LORS I'd add the omission of an "unpack" feature, akin to what we have with the unpacking of .jpg's in the photo-album package, but for the unpacking of files and text for people who might want to build combinations like real graphic designers and using programs like Dreamweaver which, in the right hands, make it easy drag and drop and cut and paste and go back and edit and then automatically check to see if the links are dead.
The odd thing is, we've got webdav that does drag and drop, skilled html people can basically cut and paste, and in the bookmark portlet we've got a link checker, but we don't have them in anything as powerful as Dreamweaver because we don't do authoring. But most of our academic users are authors, already format their stuff in Word and many have learned to work with html or have course or department flatfile websites that work fine for them and have no yet gotten into forums, or are using listservs that serve their purposes, and are basically content to go with what they've got -- until we make it easier for them. Plus, for many, copyright is cramping their style, and though we might prefer to talk about other wonderful things, lms systems get a lot of juice for their support for fair use. We are addressing various niches here, and in this case, those building flatfile websites of limited size, with content that changes gradually from one term to another, and who need more than our wiki and less than a full content management system, or who are in transition.
I realized this when stuck feeding people to Blackboard and scheming for ways "not to inhale" and so discovered and started selling my clients on the "unpack" feature whereby you build a course website in Dreamweaver on your harddrive including, say, 15 weeks with 5-10 objects per week, use your favorite style sheet to make it legible and pretty, check the links and keep them up to date automatically, and the zip it and unpack it in Blackboard in about a minute. I've discovered that if you turn off as many features as possible in Blackboard in a certain way you can't, in a particular course, get them back and so you are left with basically the permissions shell, maybe acommunications device or two, can then fill it with what you've already got more or less in one fell swoop: you get the authoring and site maintenance power of Dreamweaver and the permissions and group spaces of a learning management system. So now I am dreaming of finding this feature in one of Sille's beautiful tabs, a start page to an easy-to-upload and replace Dreamweavered site, adding tremendous depth and allowing all those early adopter instructors who have already started to get their hands dirty to build on their investments and are ready to transition to a proper lms sytem with its powerful permissions and communications. Position ourselves right above that wonderful Moodle, make it easy for all those folks who have gotten used to it and building websites and now need more complexity and scale.
Solution Grove signed a contract last week with a high profile customer which included the LORS requirements you are proposing. Ernie working with a team lead by Vivian are working on the project. My contract requires I have written permission before posting the client's name and that is in the works. However, the team is free to discuss the requirements and our technical approach.
This is a summary of our current understanding of the requirments. I am traveling to DC to visit the client tomorrow for more detailed requirements gathering.
1. Ability to upload a course from Blackboard; then combine parts of that course into another course.
a. If you change the underlying resource, the change occurs in all courses that use that resource.
b. Ability to push changes to some courses and not to others (different courses have different revisions)
2. Ability to add a new resource (say a Word document) into a LORS course from the web interface.
3. Timed delivery of material. Publish a new revision of a resource (or an entire section of a course) at a specific date automatically.
4. Expire items: unpublish a resource on a certain date.
As Rocael mentioned, LORS is about to change to address most of the concerns that you have brought up. So let me just go one by one.
> 3. Course Repositories
> Another aspect that needs to be adressed is the lors package. The idea
> behind learning object repositories is to have one repository for
> installation. At present each prof. can create his courses which makes
> sense if the course is class centric. Better would be to have lors as a
> single service with admin portlets on class level but there should be a
> reference to the repository not decentralized courses as default.
> Courses in the repository should be mapped to classes not the other way
> around (at least as a default).
As you well mentioned, LORS works in a decentralized sort of way. Basically when you install LORS in a class and upload a new course, the course is now part of that class. Yes, you can share it with other classes, but still it resides within that original class.
Although that has its advantages and disadvantages, I see the point of having a centralized repository per installation. On top of my head, I think that'll make easier for people to create customizable courses reusing "chucks" of courses to create a new course. At the moment you can only share a course as a whole but not its underlying learning objects. I think in the centralized case, the administration of all these learning object would be, to certain degree, simplified.
> A professor creates a new course and uploads the course. Some time
> later he realizes that a flash file or an avi would enhance his course
> as would adding some more subchapters or reorganising the course
> structure. Now what can he do? Basically nothing. He has to upload a new
> course and disable the last, right?
Well, yes and no.
If all the teacher wants to do is change existing content, then she can upload a new revision of the file that she wants to modify and that'll do. Now, if what she needs to do is to modify the learning object radically (changing the files and addition more and all), yes: LORS at the moment can't quite handle that.
However, and here's the good news, we've been working on versioning of learning objects. That basically translates into: Random Teacher Striker upload a course on Biology 101 to LORS. Then he realizes that he wants to modify one chapter of his course as it doesn't quite convey his learning objectives, so he can modify the course in using Reload in his desktop and then uploads the course as a new revision of the existing course. LORS will be smart in of to say "Oh, alright, Random Teacher Stricker wants to modify his existing course, let's see what the deltas are with the existing learning objects and create new versions of these learning objects". Analyzing the changes, LORS then creates a new revision of the learning objects and the course as well. So following as you mentioned before, if the teacher realizes that an AVI does better than his old HTML pages, then the new course revision will have the AVI file and the teacher won't have to disable his existing course and upload a whole new one.
> I don't even want to go into the
> performance issues involved by that. It simply takes ages for a course
> to get uploaded to lors.
Really? I thought it was quite quick. Is that because you are using large AVI files? All it does basically when you upload a course into LORS is adding it to the CR (via file-storage).
> Any chance to improve that? Other platforms
> upload courses fast as lightning.
I guess that is because they just dump the files into a www directory . That will take a second for us to do it as well, but then we will loose all the beauties we get by using the CR (access control, versioning, etc).
> How can we make lors faster on the one hand
> and less rigid on the other so that we allow course evolution?
Out of all the repositories that I have seen, LORS is probably the most flexible one. It gives you not only a repository, access control, IMS MD editor (soon we'll have an online IMS CP editor too), internationalization, full text search, etc. You can see a comparison in features between LORS and 3-to-5 other learning object repositories here http://weg.ee.usyd.edu.au/people/ernieg/thesis/final.
Once we manage to do versioning of learning objects and course we would be a far more superior repository than even the commercial ones (intrallect and the like).
PS: I would love to work on repository interoperability (ie: one instance of .LRN-LORS sharing learning objects and courses with other .LRN-LORS instances). Hopefully that will be not too long in the future.
thank you for your responses.
Bruce - sorry for my bad englisch but I had difficulties understanding your post. You are too poetic ;) Could you kindly sum it up in form of requirements or needs that you see?
Caroline, Roc & Ernie - I am happy to read about these developments. And yes Ernie - I think LORS is great. My comments were from the perspective of a power user.
Ernie - as you have noted correctly they do extract the stuff in a www folder. At present I don't see any benefits CR could give since I believe that learning objects have a very short lifespan in the sence that noone will care about old revisions of objects. What people care of are complete course revisions. So why not extract the stuff to a www folder first and let a scheduled proc upload the content to CR behind the scenes? Just a thought. Or maybe always upload to a www until the author agrees that the required maturity degree is achieved. Anyway, I don't have a good for that solution at present. All I know is that at the moment we end up with a lot of dead disk space and time consuming uploads at present which doesn't go well with an evolving course paradigma. And to be honest I think that is mostly because of the SCORM and LOM standard. The standards regard learning objects as static content that you simply deploy. They don't care much about evolution.
Caroline - I think lors should be a central repository of learning objects and the manifest file with it's organisations and items only a view to that repository (course structure). We shouldn't care too much about the way resources are organized to a single course since that information is too transient and each course user (prof) has a different demands regarding the course structure. So what we need is a more flexible way to redesign course structures. I am really interested to see what you all are developing on that field. But one solution could be: If you want to change a course structure (removing, adding, reorganising) you first download that course. LORS makes sure all objects have unique Ids which it can recognize later on. Even better would be if LORS would respect Ids that were passed from the package the first time the course was uploaded. This would allow authoring tools to reuse that course. Now you can rearrange your course let's say with reload. Finally when uploading the course LORS can figure out by the Ids which objects were deleted and added and wether the structure has changed or not and offer (and this is what I think is important) an new revision of that course structure. So each prof could reuse a course and rearrange it the way needed. Does this make sense?
Sorry for waxing poetic. What I'd like to see is an "unpack" function for zip files composed of websites such as one might make with Dreamweaver. This method may not be as sophisticated in data terms as Lors, but for those having built simpler flatfile sites and just starting to use an lms, being able to "unpack" an existing site in the middle of Dotlrn, I am thinking, this feature would be a plus. The advantage of Dreamweaver for authoring, as I understand it, is the site maintenance features, including automatic link checking -- something my clients are very interested in as they are making course websites including texts and links. The rest, as they say, was fill. B
As long as I have understood in your post, you would like to upload a simple collection of HTML files packaged into a "zip" one, and deploy the whole site into dotLRN. Well, I think this doesn't means any improvement for the platform.
My reasons are the following:
- An SCORM package is just a zip file containing all the html files and the manifest, written in XML. Tools like Reload allow you to create this manifest easily, and the package can be deployed into LORS as a simple web-site.
- So, this feature (your proposal) is covered by the platform, so if another module is developed for unzipping web-sites it would be redundant features.
- SCORM includes a lot of features very usefull in education, much better than a simple plain web-site, so i'm sure that the effort done for creating the manifest is not too much.
I apologize for my English, i hope my ideas are clear despite my lacks on writing.
Based on my requirement gathering next week I think our client has the same requirement to be able to upload any zipped collection of files and have it put into LORS and become exportable in IMS format.
On a related note, if you are working on SCORM/LORS and get paid for it, would you mind making the interface as slick as the one from Ilias (http://www.ilias.de/ios/index-e.html):
This is a (albeit german) PDF document and I'm more than willing to translate on chat if you'd like to know which slide shows what.
In general, if you are marketing .LRN in the US, you might want to take a very close (read: copy/paste) look at the PHP code of Ilias to save you quite some time and get a lot of ideas which you can sell in the future.