Ask HN: Why is there no good open-source LMS?

[seestraw]

I want to start a teaching business. I'm struggling to find good open-source software on which I can start building my platform. The state-of-the-art seems to be Moodle or Canvas which are designed for universities - they use archaic UI and features.

Is there something where I can host an online course, also have live classes and perhaps add on features like a community for students, assignments, etc? The only options I am seeing are things like Kajabi and Teachable which are very restrictive in their feature set, and not customizable.

Anyone running an edtech company here? What do you use? Or do you build everything custom?

[claviska]

I haven’t worked in academia for over 10 years, but back in the day Blackboard would either purchase or squash it’s competitors with litigation.

https://en.wikipedia.org/wiki/ANGEL_Learning

https://www.zdnet.com/article/blackboard-wins-e-learning-pat...

They even sued the government to prevent reexamination of their patents.

https://www.bizjournals.com/washington/stories/2008/12/01/da...

As an open source author who was once interested in producing an LMS, this was a major concern that forced me to explore other fields.

[fabbari]

Their patent is currently expired [0] in the US - but yes: it is a great example of overly generic patents.

[0] https://patents.google.com/patent/US6988138B1/en

[streamofdigits]

The paucity of open source LMS likely reflects the uneasy relation of the public sector with open source more broadly. You'd think its a marriage made in heaven: public funds, procurement etc leveraged optimally for common good infrastucture - where that makes sense. But in practice this mode seems entirely marginal. It may be corporate capture, lack of savvy people among decision makers or other factors.

In any case, given the huge upfront investment required for a quality platform this doesn't seem like something an edtech startup can bootstrap. An education platform is not a CMS and its not a social media platform. I think especially now with the pandemic experience it has become very clear how rich, complicated and demanding the educational process actually is. The "archaic UI and features" comment hints maybe at too narrow and technical view. It may be a very relevant aspect (eg if young students puke at the UX it is not of much help). But from an education perspective what matters are not smooth appearances and gimmicks but "educational outcomes".

If you dive into the Dougiamas/Moodle team's thinking you'd see what permeates the design/architecture is to be able to translate the huge body of educator experience and infuse it into software. Somehow we need to move to the next chapter of the book they started writing.

[xmcqdpt2]

>The paucity of open source LMS likely reflects the uneasy relation of the public sector with open source more broadly. You'd think its a marriage made in heaven: public funds, procurement etc leveraged optimally for common good infrastucture - where that makes sense. But in practice this mode seems entirely marginal. It may be corporate capture, lack of savvy people among decision makers or other factors.

This is interesting. I had the exact same experience working in a major bank. Corporate would usually rather buy anything than "invest" in open source.

Partly it's a question of "support" (getting RHEL instead of CentOS) even though in practice support is often rather poor and distant. More frustrating is when we buy what are clearly simple reskins of OSS with terrible support from eg Oracle. You get all the disadvantages of using OSS (sometimes poor documentation, too many configuration options etc) while also not having code access or control over the platform.

I've come to the conclusion that corporate is sketched out by OSS because there is no one to sue of something goes awry, and that they just don't trust in-house expertise over basically any vendors. To be honest, they are not entirely wrong about the second one: this kind of corporate culture makes good engineers leave.

[blablabla123]

> In any case, given the huge upfront investment required for a quality platform this doesn't seem like something an edtech startup can bootstrap.

Exactly, especially if it's supposed to be really generic. I've co-founded 2 ed-tech startups targeting niche areas. In both cases there were dozens of competitors in these niches and some specializing in some sub niches. Those niches are most of the time good enough, usually you can find a mobile or web app for anything.

But as mentioned, working with public or private education orgs brings its own challenges. I expect getting any leads there to take years. It's much easier to target consumers directly.

That said, if anyone came up with a generic open-source LMS, on the long-term it could be quite disruptive because everyone in that area is just constantly reinventing the wheel. It's just a ton of incompatible solutions.

[codegeek]

One reason is that is is really tough to define what an LMS does since it can serve various use cases. Most people who are not familiar with LMSs think of Blackboard/Moodle/Canvas/EdX and mostly for academic/schools.

There is a whole set of industries for LMSs like B2B training, HR/Compliance/Continuing Ed/Healthcare etc. Then you have Product companies who just need to train their customers, create brand recognition etc. I jus spoke this AM to a startup who works with Kids and they need a "Kid Friendly LMS".I regularly see companies try and twist Moodle/Canvas/EdX for these but usually they end up with in-house duct taped solutions.

The challenge is that each industry can have its own needs and requirements and building something that can truly cater to all audiences in one single monolith LMS is impossible.

I am thinking about building an API first platform where you could build your own interfaces on top while the API does the heavy lifting. For example, you want to start a webinar ? Just call "/api/v1/webinar" and build your own frontend for it. Think of it like "Stripe for eLearning".

Disclosure: I run an LMS company so I know the challenges :) If you are interested in this space, hit me up. I am looking for people to join us.

[zerkten]

There is a lot of untapped potential here. I was close to early enterprise adopters of Articulate which was a much more limited product in the mid-2000s. Pharma had HQ training (yearly compliance plus some specialist courses), but field sales has regular training and updates which was a full time production effort. A single LMS could have worked across the org as long as it had enough of the nudging and reporting which would have served the industry needs.

It seems that many people on HN struggle with understanding the needs of industries because their formative experience is as developers in startups and tech. This really inhibits the ability of folks to attack opportunities later when they want to create a startup. The number of times I've seen enterprises play with startup tools while they wait for the industry incumbent to catchup is very high.

[cors-fls]

Check out Richie. It is a modern MIT-licensed LMS built with Django and React.

It is used by FUN (France Université Numérique) a public MOOC platform, and EduLib, which looks like the equivalent in Canada.

Richie : https://richie.education/

FUN : https://www.fun-mooc.fr/fr/

[seeekr]

Thanks for the link. IIUC Richie is more of a CMS for courses and related info that can integrate with an LMS like edX, but it's not an LMS itself.

[billpg]

What's an LMS?

(I do understand that my asking this question almost certainly means I can't answer your question.)

[HatchedLake721]

I don’t know why but it always gets on my nerves when people decide to address a multinational forum with people from every corner of the world and ask for their time and attention, but at the same time decide to save 2 seconds of their own time to avoid typing out “Learning Management System (LMS)”.

We all are from different backgrounds and industries and same acronyms can mean different things.

I worked in finance, for me LMS is Loan Management System.

[mkl]

Learning Management System: https://en.wikipedia.org/wiki/Learning_management_system

[callmeal]

I thought it was a Laboratory Management System (for tracking samples).

[Megha11]

I am wondering the same thing. Thanks for posting your comment.

[nick88msn]

Learning Management System. An online tool to manage physical and online courses students wiki test etc.

[raesene9]

Learning Management System - manages training, students etc

[zecg]

Moodle is actually great, IMO. You can customize the UI to your liking and integrate it with bigbluebutton for a complete solution that includes remote learning classrooms instances as need arises, with full recording of lectures.

[BudaDude]

Moodle is good until you realize that the documentation on it is garbage.

[indirectlylit]

At Learning Equality we build and maintain Kolibri. It's an open-source, offline-first, self-hosted learning platform.

https://learningequality.org/kolibri/

https://github.com/learningequality/kolibri/

[sails]

How does it compare to Moodle/Canvas in terms of intended use?

I have a contact who is looking for an LMS for teacher education, and kolibri looks probably appropriate.

[endymi0n]

I've been wondering the same and came to this conclusion:

- Existing learning management systems are a mirror (and a victim) of the education system itself, as that's where most of the developers come from: Academic, underfunded and people-focused.

- Academica leads to overflowing complexity. In the whole system, simplicity is punished and complexity is cherished, so you end up with confusing UX.

- Underfunded is pretty self-explanatory

- People-focused is where like in any bureaucracy, nobody really wants to make anyone else replaceable. So instead of with a hard focus on users and learning tracks, you end up with an old system of classes, teachers and students, where of course you have the 10% great teachers that _should_ have run the class for everyone and the 90% that starves their class of any legit info.

From the uni and publisher sides, it's similar, but not completely equal. Both universities and book publishers would never make anything truly great because it would all cannibalize their business.

If you come from the other side and think that the whole learning system is rife for revolution with first-principle thinking, tree-trunk learning and a standard of "Intuition isn't optional" (massive props to https://betterexplained.com/articles/intuition-isnt-optional... ), then you quickly see the other side isn't interested in any fundamental change in their approach (see above for the why).

Furthermore, even if you try, those are the people in charge deciding where their (very tiny) funds go. They will go towards the solution that prioritizes system survival above quality and radical change.

So you see these people rather going towards the content publishers.

For investors, the field is mostly uninteresting for the reasons above, so you don't see any quality invest.

As a parent, I'd love to see someone really cracking edtech, but unfortunately what it would take would be a pretty massive initial investment into a seriously great solution that then proceeds to tackle funding, education system and self-reliance as well. It'd be a philantrophic invest of a few dozen million and then go for a very long road of paying back in small rates.

I'm still up for it, maybe in my next startup :)

[em-bee]

Existing learning management systems are a mirror (and a victim) of the education system itself, as that's where most of the developers come from: Academic, underfunded and people-focused.

this.

part of the problem is that an LMS is used by several different types of people:

    - students
    - teacher
    - parents
    - administrators
    - other staff

everyone has different needs and expectations.

also, the users who use the system the most are not the ones paying for it. so they don't get much of a say.

every school has different priorities.

a good system for you may not be a good system for anyone else.

I'm still up for it, maybe in my next startup :)

i have been looking into doing that for a long time, but i had to shelve the idea. i am still interested in approaching this, but i do not believe that a large project without any users right from the start will be successful.

rather the approach should be to find a school, build a custom system for them and expand from there. it is the only way to build a system that actually has users and has a chance of getting funded (by those users)

without those users you'll build something that noone else will want.

[rvanlaar]

I wondered the same as well. My idea was to change this and build one that worked....

A few thoughts:

All LMSes need to be seen as a result from their surroundings. Education is a very bureaucratic sector with lots of money, just not in the places to make a teachers' life easier.

Current applications are in this equilibrium where they are good enough for the field as a whole. Yes, there needs to be innovation and it will happen, just very slowly.

My 2 cts, it can be done by lobbying with politicians, deans and school boards. Getting them on board, creating a pilot program and buzz. That part will take more time than actually developing the software.

Disclaimer, I build such a system for a customer. It was very specific and trimmed down for their use case.

[akie]

Even if you have great world-class software, you will get bogged down so easily in university politics and the slooooooow moving oil tankers that they are that I would strongly recommend you to spend your time and energy elsewhere.

[Jacqued]

I work on a team where we build around and run Open edX. It's incredibly complicated to run because it's not built as an open source product. It's just that edX open sources their software after the fact, so they're not investing in making it easy for anyone to run, of to contribute features.

On the team I'm a part of, our strategy is to build open source tools around Open edX, using it as a site builder that stitches together all our external services (video hosting, live conferencing, course content, quizzes, web-facing portal, forum, etc.). That is, until we cover enough scope that we can start to replace it with a flexible and lightweight LMS that would basically just do the aggregation.

IMO it's not necessarily a good fit for the LMS category to try and build a gigantic software codebase that handles everything the way every single learning institution wants it. You end up with a glorified site builder with a bunch of specialized features tacked on.

[sails]

> incredibly complicated to run because it's not built as an open source product

Would you suggest it as an option to a team with limited devops budget?

The documentation seems reasonable comprehensive. Do you have a brief summary of issues?

https://edx.readthedocs.io/projects/edx-installing-configuri...

[jrm4]

Because there are no good LMS, period. The entire concept of a "Learning Management System" is stupid and conclusory, but makes for a good business model that you can sell to unsuspecting administrators.

You just cannot "all-in-one" a system of learning without real modularity, and/or non-exclusivity. -- but the LMS always HAS to do some of this in order to keep its percieved edge; this creates very bad incentives of lock-in, like Canvas' idiotic and horrible "Conversations," which is just email but worse.

Source: I work in one everyday, and something always suffers as the money-maker tries to keep their competitive edge. In my experience the best shot at this is probably Google Classrooms, precisely because they're NOT desperate in this business.

(This is very analogous to how Wordpress became the best CMS precisely because it wasn't trying to be one.)

[TrispusAttucks]

The underlying protocol most [1] LMS use, [2] LTI, is horrible (IMHO). That may have something to do with the quality of platforms built in top of it.

[1] https://en.m.wikipedia.org/wiki/Learning_management_system

[2] https://en.m.wikipedia.org/wiki/Learning_Tools_Interoperabil...

[Uehreka]

Ah LTI, it’s been a while.

The tough thing about LTI is that it assumes the piece of content you’re trying to serve is a single resource. So if you’re serving a video or a single HTML page, then cool. If your page is some sort of JS player that loads the content from an API after the page loads, you basically have to create a relay race of auth systems to tie it all together. This is a super common format (a JS lesson player that lazy loads each page/slide as needed) so a ton of people have had to solve this problem in different ways.

[PeterisP]

While LMS's support LTI, I have rarely seen it actually used; usually you would have all or almost all content hosted "natively" on the LMS you use.

[seestraw]

Hmm do LMSes really get built for the LTI protocol in mind? Or do they add it later for compatibility?

[ashwinm]

You should try this out - https://www.pupilfirst.com/

[seestraw]

This looks pretty cool. Thanks, I'll take a look.

[2Gkashmiri]

looks pretty nice

[major505]

I would really take a second look to moodle. It's very customizable, have some good mobile apps, and a great api, so you can extend it to your heart's desire.

[geocrasher]

If you're a developer, Moodle is fine, because it needs a developer to become usable for most businesses. It might be better for academia, but for business use it's garbage.

[acbart]

Built on PHP, over-engineered, and terrible out-of-the-box UI. I wouldn't personally recommend Moodle, unless you had no other option. It takes all of my energy to resist suggesting you just write your own LMS from scratch instead of using Moodle...

[brodock]

The problem most LMS face is that they have to cover a variety of personas and automate their worflows. What schools and universities pay for is something that can work more like a CRM than a CMS, and those have to be flexible/configurable etc, which makes it complex to setup.

[immnn]

Good question and very relative nowadays. You should have a look at ILIAS. It’s been developed by German universities. However, it’s very versatile and pluggable.

As a company we are offering several integrations.

ILIAS is very strong when it comes to collaboration, but it doesn’t offer video conferences out of the box. Actually, I also don’t see ILIAS in that position. For such requirements you should definitely have a look at other solutions like MS Teams, Zoom etc. just because of scalability. Here in Germany mentioned solutions aren’t allowed because of privacy concerns. However, there are open source solutions followed by a system offered by the state for schools and other educational instances.

[ktpsns]

Can second that. Have worked 10 years with ILIAS (in all roles, from coding extensions up to editing content). It is a very solid platform with a rapidly moving team and an ever increasing set of features.

[chkgk]

Check out OpenOLAT. Many universities use it and if offers great integration for assignments, self-tests, video-calls (big blue button), etc.

https://www.openolat.com

[trabant00]

I find commercial LMSs to be full of dark patterns that substitute learning and value with gaming-like interfaces with achievements, medals, bars to be filled, check-boxes to be checked, Netflix-like hard to navigate to give the appearance of endless choice, lack of human curation so you can only find the most popular but not the most valuable and so on.

If your interest is in selling then yes, nobody has done the work for you for free. Why would they? Open source is geared towards sharing, not selling.

[867-5309]

having worked for a company which made LMS software I can vouch for this. buzzwords like "gamification" and "user stories" were bandied about daily with next-to-nil privacy concerns. want to implicitly record every student's interaction? there are a handful of APIs for that..

"The Experience API (or xAPI) is a new specification for learning technology that makes it possible to collect data about the wide range of experiences a person has (online and offline)."

[jrochkind1]

The open source solutions that exist are not very good/difficult to work with. The proprietary-commercial solutions that exist are not very good/difficult to work with.

This may mean that it's a hard space to do something "very good" in; compare and contrast to "ERP" systems. Or "enterprise" systems generally -- an LMS is definitely an enterprise system. (Meaning: purchased for an entire organization; those with the most power in purchasing decisions are for the most part not those with roles as core users; often purchased based on "feature checklists", or "what are my famous peers using"; needs to support a kind of 'workflow', which can vary drastically among different organizational customers or even within a single customer).

Consider when it's a pitfall to actually "give the customers what they are asking for". (pitfall to quality/ease-of-use but not always to sales)

Those making procurement decisions need to have someone to call and complain and ask them to fix things -- even if they don't fix them. It's important for the careers of those making procurement decisions to have someone else to blame -- and "nobody ever got fired for buying IBM." This entity to blame hypothetically could be a vendor offering installation/hosting/support for an open source solution... but it's a risky business to be in, when those doing procuring would rather take the safe/familiar route, and the product you are offering installation/hosting/support for, any competitor can too.

[kayodelycaon]

> "ERP" systems

If there is a good one, I don't know about it. :)

[bcx]

I once TA-ed for the guy who runs https://www.sakailms.org/, I know his vision is to make a good opensource LMS, so you might be able to tell him what you are finding lacking in other projects.

Dr Chuck is one of the top professors on Coursera, and I understand he's mostly using the money he makes on coursera to fund Sakai as a passion project.

[michelb]

I worked for an e-learning company where they built a custom solution for various reasons, but mostly for flexibility.

Some of those reasons: - integration with other (client)systems - advanced payment integrations, like bulk sales to schools, free accounts for teachers, combination packages with books (separate ISBN's) etc. - ability to introduce way more types of content presentations and learning systems - work really well on mobile - make the UX way, way better - different ways of tracking progress, Xapi or something country specific

It was and is quite the investment, but they have complete control and they can offer something nobody else can.

We've looked at other solutions both open-source and commercial and most of them truly suck. They either contain too much or too little or require heavy customisation, or the UX is appalling.

Once you need to onboard organisations or schools, each with multiple departments and their demands/regulations, you're in for a world of hurt unless you can adapt your platform.

It really depends on what a 'teaching business' will be for you.