Ask HN: Why is there no good open-source LMS? | Hacker News Reader

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

157 pointsseestraw159 comments

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?

View on HN

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.

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

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.

>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.

Maybe its just a phase towards maturity. Functioning / reliable OSS support models are important. "Suing" is just a metaphor for needing reliable partners: running a banking institution or an educational institution or a local government is not a hobby.

But the outsourcing to proprietary vendors mentality might be a far bigger risk. If all you are doing is processing information those vendors will eventually eat your lunch and you will have financed them every step of the way.

> that they just don't trust in-house expertise over basically any vendors.

Another way to look at this: they recognize that they aren't software firms. If you bring in OSS, you have to manage it like a product. You have to recruit, train and manage competent developers. You have to balance competing priorities from "customers" across the firm. That's hard for firms that focus on software, and often impossible for big enterprises.

There's also a cost distribution problem. If you buy commercial software, probably other people are buying it to. Thus you share the development and maintenance costs. You can (hopefully) competent software management too.

I think "open core" is probably best balance here. Yes, we need better funding/pricing models here.

> there is no one to sue of something goes awry

Hum... There's no way to sue any commercial software distributor, it doesn't matter if it's proprietary.

This is the correct answer for sure. I managed an LMS department for years and you nailed it.

I'd also like to add that most education IT departments tend not to have many people capable of building or maintaining something as complex as an LMS. That means you'd need to hire at least a couple new software devs at software dev salaries and software dev benefits to keep them...which ends up costing MORE per year than it costs to shove 99% of the problem off onto Blackboard support and remove a large source of risk while saving money.

this does not address the OSS / proprietary choice though. e.g., there are companies providing support for Moodle. it is service that must be paid somehow, whether through internal staff, external partners or a proprietary software vendor as an additional offering.

The mystery is why they wouldn't amortize the development costs of the platform across the vast number of institutions, creating an ecosystem where smaller edtechs could provide niche customizations via plugins etc.

> 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.

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.

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.

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/

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.

Hey! I'm one of the maintainers of Richie. That's exactly right, it's meant to be a web portal in front of one or more LMS.

Thanks for the correction. My mistake.

What's an LMS?

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

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

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

Learning Management System - manages training, students etc

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

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.

I have the same issue with numerous TLAs, most frequently around here: "NLP".

My usual thought sequence is "'Neuro-Linguistic Programming'? Uhhh, no... it's not that, not in this space. 'Near Letter Perfect'? Hmmm... don't think so. 'New Lunar Passage'? Has Elon been smoking that stuff again?..."

Eventually I figure it out, but for some weird reason my brain doesn't really want to hang a hook into "Non-Linear Processing"

;)

TLA?

Three letter acronym

Oh, not TLA+?

No, that would be an ETLA - "extended three letter acronym".

Errr isn’t it natural language processing?

It's also a branch of pickup artist pseudoscience.

Yes, yes it is.

Aks my brain to pick that one first, though.... ehhhh.

You just did the same with TLA ... Thanks to another commenter who explained it

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

I've usually see that as a Laboratory Information Management System (LIMS)

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.

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

Moodle Is Great! It's Open Source! If you need more functionality, just add a plugin.

WAIT. HIT THE BRAKES RIGHT THERE.

Yes. It's open source. But that isn't helping anything. When you find a plugin that does The Thing You Need, you find that it's for a version of Moodle that's several years old and was written by one guy who hasn't touched it since 2014.

There are some plugins that are great and up to date. They are not Open Source. They cost real dollars and aren't terrible.

But if you want Moodle because it's FOSS, then your motivation is wrong.

The one largest problem with Moodle is its dependency on plugins and the consequent decentralization of core functionality. The core UX itself is shit because it needs to open space for plugins to change it and because it was created as a set of independent plugins itself.

But yeah, if you need more functionality, just write a plugin.

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/

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.

Unlike most other learning platforms, Kolibri supports users that don't necessarily have fast or consistent internet access.

While it can be hosted in the cloud, it is designed to run on a local network so that the experience is fast and responsive, and is specific for the needs in places with limited internet. Content and data can be synced to other locations as-needed.

Thanks. If the content is likely to be entirely cloud based, would this then be inappropriate due to the platform being built around local and location syncing?

If you mean there's a server-side component to the content, then no Kolibri would not work.

Kolibri supports a range of offline-friendly, self-contained formats such as videos, exercises, HTML-based apps, PDFs, and e-books. These can be uploaded to the cloud-based Kolibri Studio [1] and then synced to the learning platform. We also maintain a large collection of openly-licensed educational resources [2] freely available for immediate use.

[1] https://kolibri-studio.readthedocs.io/ [2] https://catalog.learningequality.org/

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 :)

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.

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.

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.

I’ve worked in academia as a SDE or Product Manager for 3 different universities ( Columbia, MIT, and an online school ). From my experience the LMS is always a COTs product with very little customization by IT. I’ve never heard anyone complain about it ( from the college side at least ).

The real need is not the LMS but rather the SIS ( student information system ). The available LMS’ are adequate but none of the Student Information Systems are flexible enough to suit the need of most large universities.

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.

> 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...

I'm not an ops guy, but I know it was a constant source of trouble in our team and a large challenge to keep it running smoothly. For us it took 2-3 experienced engineers something like 2 years to have a stable and smoothly running production environment at scale.

If you start from scratch today things may have gotten better. You might want to look at https://github.com/overhangio/tutor, I know Régis has been hard at work making it easier to run Open edX.

Thanks, that is exactly what I need to avoid. Will have a look at that.

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.)

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...

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

LTI is an afterthought for LMSes.

The standard is a classic patchwork of whatever is used by IMS members currently.

LTI 1.3 is the current active version, which deprecated 2.0.

But wait LTI is actually a collection of standards. AGS (assignment and grade services), DL (deep linking), dynamic registration, etc.

And they are optional, and LMSes implement them in slightly different ways.

So good times.

But wait, there's also the afterthought of 1.3 compatibility jury rigged on top of 1.0/1.1 implementations. Which leads to comical implementation details poking their ugly heads out :)

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.

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.

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

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

looks pretty nice

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.

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...

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.

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.

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.

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.

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

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.

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)."

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.

> "ERP" systems

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

The best you'll get is NetSuite.

exactly

Depends on what you mean by good but there are many that certain enterprises couldn't do without that well outperform competitors and employees would melt into bubbly goo if they couldn't use them anymore -- but they tend to be very custom made for either a particular enterprise or a very niche set of very similar enterprises.

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.

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.

I've thought many times of building one but ended up using something similar to Teachable (for now). I don't even want to mention their name because their service isn't that good and I wouldn't recommend them but moving tens of thousands of folks to a new platform is a big burden for students so the only time I ever plan to make a move is when I release my custom platform.

But, as for open sourcing the platform. I think it's tricky because if you're in tech chances are you'll want to build your own custom platform, if you're not into tech you'll use an existing SAAS service or outsource someone to build it for you in which case there's almost a 0% chance it'll end up being open source because the goal there is to hire someone on demand to build and lightly maintain it, not champion an open source project for years to come.

Plus the decisions you make when building the platform are super custom based on what you want.

Are you going to support the idea of buying multiple packages for 1 course, such as package A and package B at different prices? How will you handle upgrades? What about subscriptions vs single purchases? Can students comment on things? Will it track and remember progress? Will you have searchable captions? What about handling time sensitive discounts? Which payment providers are you going to support?

Then there's the whole idea of optimizing the course consumption experience for your audience, for example a tech course could be displayed and have features that are much different than a Yoga course.

There's about 100 other questions to ask here and chances are your combination of features isn't going to be the same as someone else so then you end up with limited platforms that try to do everything and create an average at best experience, or you drill down and build a super niche platform that's great for what you want to do but maybe not everyone else.

All of this makes it very hard to have a successful open source project IMO.

It sounds like Appsembler's Tahoe product is what you're looking for, come checkout our free trail!: https://www.appsembler.com/tahoe/ FYI - I'm a software engineer and I work on the Tahoe/Open edX team.

Tahoe is a user-friendly platform that makes it really easy to spin up an open edX LMS. It aims to alleviate a lot of the issues that are mentioned on this thread about Open edX. Tahoe comes with out of the box tools to customize your LMS and provides great user support (docs/highly responsive customer success team).

Would also suggest Moodle. My remote university is using it and while the interface is a bit archaic (= looking and feeling outdated) and many functionalities are almost never being used because of better alternatives (like the chat function or messages), it does have its advantages. You can set up a good work space where students can use the forum, you can interact with them, you can set up learning assignments, upload documents, videos and so on. I think it does its job quite well in the area that it is meant to be used. No live classes, afaik, but those are easily managable via jitsi, for example.

Part of it is probably the cross compatibility requirements. SCORM and LTIs and all that stuff mean that content has to be able to render the same way for the textbook publishers and everyone else to sell content to schools running on any LMS.

Canvas may seem a little old school but I really like it. It’s not that hard to extend, API is good (they even expose graphql to students) and it’s fast. It feels like it was made by humans, as opposed to every other LMS that feels like it was made by a bunch of people in suits for school administrators. (Much like GitHub in that respect.)

I agree with you canvas is actually very good at what it is supposed to do. I had to work with blackboard previously and I still have nightmares. Also regarding innovation in this space. I think we've actually seen quite a lot of movement in this space. Over the last 5-10 years many university moved from proprietary solutions like blackboard or custom solutions on top of some cms to open source solutions with open interfaces. There is a whole community of plugins and extensions and consultants writing oss extensions.

https://www.odoo.com/app/elearning

Not exactly a LMS but the open source ERP system Odoo offers a lot you need to sell and provide courses. You can also do live trainings using the Jitsi integration.

Odoo is solid web-software built with Python (and HTML, CSS, JS of course). It's quite easy to extend.

There's a demo at https://demo.odoo.com

(Disclaimer: I'm an official Odoo partner)

Hello!

1. Can you tell me more about the scalability of Odoo LMS for 100-1000 simultan users. What server resources is needed?

2. In the context of using the Odoo LMS, it is a good idea to have the ERP and LMS sharing same web server resouces? In production you use the LMS as a separate Odoo instance, but sharing same database as the main ERP?

Thank you in advance!

Drupal has some distributions aimed on this. You can look here: https://www.drupal.org/project/project_distribution?f%5B0%5D...

Or build yourself using modules on a vanilla drupal.

Because there are no good LMSes. It doesn’t matter if it’s open or not. The concept is broken. The platforms people actually like learning with aren’t LMSes.

Kind of agree. I learn all the time, through writing posts, using an mooc platform, coding sites, I search in scholarly search engines, read books.

There is not one "learning" platform that covers all these cases, let alone cover well. The activities and the learning modes are too diverse, I would not trust any single entity to cover all of them.

Apart from underfunding, I think fundamentally the problem is that a good LMS needs to be extremely flexible and customisable to support different learning styles of students, and also pedagogic techniques which vary a lot across subjects and disciplines. It's a complex problem to solve IMHO, with the greatest risk being an overly simplified and constraining system that ends up getting in the way of learning.

Since you list other non-open source solutions as well in your question, I’ll mention one more, which might be the closest to what you are looking for > maven.com.

We started collaborating with them right after they launched, and just finished our first course on their platform.

I also wish there was a good open source alternative but the revenue sharing with maven is fair, the team is a joy to work with, and the product is improving rapidly.

Is (online) education is an attractive market? Before Covid, the answer was: Meh. Lot of competition, little funding in most countries. And Moodle (or Canvas) is a good enough solution, both also enable kind of third-party marketplaces. We looked at different LMS solutions and some better ones could best be described as "Moodle distributions", centered around a theme and some plugins.

Moodle seems like Joomla. The support forums are also full of posts by what seems like "the least bad IT person" in the education dept wondering how the hell this thing works, posting all kinds of confusing statements and code.

There's so many excellent SAAS platforms ... there should be something better than Moodle.

I worked in this space for a loooooong time (mostly in higher ed; I have less knowledge about the corporate or k-12 segments.)

The short answer is: It's surprisingly complex to build a good LMS at all; it verges on impossible to build a good generic LMS that covers all or even many use cases.

Everyone teaches differently; everyone has their own particular method for grading; everyone has their own opinions about how much or how little classroom communication should be handled through the LMS vs in person; everyone has their own preexisting tools and data they need to integrate with. All of those differences conflict with one another to some extent; class sizes differ, class durations differ, etc etc etc. Supporting all of these would guarantee an overcomplicated and difficult to use interface.

This isn't the sort of thing that open source tends to do well. The best open-source work is "scratch your own itch" kind of stuff; this pattern in LMS development has historically been that they start out that way, as a home-rolled solution built by a tech-savvy instructor, then accrete layers of contradictory features in pursuit of a larger audience, and either become too unwieldy to use in their own right or else get purchased by one of the already-bloated large competitors.

(I find it personally hilarious that you cite Canvas as the "archaic UI and features" state of the art: Canvas found its original success as the clean, modern alternative to BlackBoard, which at the time was the bloated-to-the-point-of-unusability standard tool. The cycle continues.)

There are also a lot of regulatory hurdles around accessibility, privacy, and pedagogical standards -- particularly in K-12 ed -- that, while very necessary and useful, take a lot of work and knowledge to accomplish, and serve as a barrier to entry to new products in the space.

https://www.sakailms.org/

Sakai was what we used before moving to canvas.

In my limited experience (mostly hearing complaints from others who have to use them, rather than personal experience) there appears to be a lack of non-open-source options too!

I think the problem is that everyone wants something slightly different from an LMS. If they are lucky there is something out there that works the way they want out-of-the-box, otherwise they end up trying to bend something else into the target shape or using something that tries to please everyone so ends up overly complex and unwieldy.

> which are very restrictive in their feature set, and not customizable

What you want to customise could be key. Can you name a non-open option that you think would be at least close to ideal for your plans, without too much panel-beating to change its shape, and why that one in particular is attractive? This might help with getting better suggestions for alternatives.

Have you looked at https://github.com/edx/edx-platform? It's one platform that's established and open source. It might not be lightweight enough for your needs though.

What's so bad about Moodle?

I mean it's not shiny, but it worked fine, what else should app like that do?

Having worked in this industry for a number of years I feel the core issue is that an LMS is trying to organize and standardize an industry that is fundamentally disorganized.

There is an appearance of structure to content, courses, and curriculum but it's constantly evolving and changing. Often mid-class! The reality is the Eng 101 class you took last year is not the same someone will take this year.

Every school/class/teacher has exceptions to the way they run and evaluate things. So the software always evolves into more of a CMS+CRM with a million checkboxes and settings.

Source: I'm developing an LMS + remote teacher management platform at a startup right now.

I deployed the open-source version of Canvas for use at a summer course at a university in Beijing (it was done remotely). Just getting it installed was the hardest part. We hosted on Alibaba in Hong Kong (no web site license required there). Support for the open-source version comes via IRC and was actually quite good. Show-stopper was a rapidly growing log file (due to a persistent error I couldn't figure out). Eventually had to create a cron job to regularly truncate it. Apart from that it was pretty solid. The other tricky thing was getting outbound email to work (not very well documented). Persistence pays.

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

We created a LMS from scratch at Primerlabs.(https://primerlabs.io) Though it had different requirements, because we create conversational courses.

If you are starting with video based platform, I recommend not focusing on creating a LMS from scratch and use tools (zoom, google sheet/form) etc. based on your requirement and create small tools to fit as your grow. For community you can use discord or discourse.

Creating an LMS from the get go is time consuming and does little for your business. My 2 cents.

I teach 15-30 sized courses to adults but I also teach at a university.

My solution - I installed Moodle on a cheap Hetzner VPS.

Added Let's encrypt for SSL and it seems to work pretty well for 100ish students.

I looked at alternatives but there really wasn't anything better that is open source.

The critical factor - all the universities that I've taught at and also know of at use Moodle.

Thus preserving the lessons, quizzes, exams, pages, etc. was quite important.

I was able to migrate my courses from older Moodle versions(3.7) at university to my own latest version (3.11) with no issues.

One thing I have not done yet is set up mail.

How well does e-mail from Hetzner IPs get through to gmail (what most students use)?

LMS stands for Learning Management System in case anyone else didn’t know.

And it’s a misnomer as far as I can tell, because it’s a system for managing the teaching, or the administration maybe, but is not actually used by the learners.

IDK, in our local universities Moodle-based systems are very much used by the learners as that's the primary way of receiving and looking up all learning material other the lectures themselves (and now with covid, also the recordings of remote lectures), the organizational information e.g. schedule changes and other comunication, feedback, and of receiving and submitting assignments and, in some cases, tests; it's essentially the primary "portal" used by students every day.

Thanks for explaining, I stand corrected!

Not a misnomer. Features of an LMS are used by the learners

From someone who's spent 15 years in industry (from time to time being subjected to corporate training) and a dozen years as a prof where I've used Blackboard and Canvas:

An LMS falls into the class of software where the purchasing decision is made several layers of management above the actual users. At least in the case of LMSs sold by for-profit companies (I don't know about Moodle, but Canvas is basically Instructure Inc) there are zero incentives to provide a product which actual does the intended job, rather than ticking all the check boxes in an evaluation process.

Moodle is very configurable, took me 40 minutes to create a basic theme so that it sort of followed my company’s style guide.

I also found it a bit complex at first. But after an hour or so it started to make sense.

Just don’t try to hack on its internals or extend it. At least back when I was tasked with doing so.

I was well versed in PHP, but my gosh was that idiosyncratic even then…

I would really recommend Eduflow here. Have tried several LMS's and it has an intuitive platform that resembles notion. We're experimenting with its peer learning components at the moment.

That doesn't seem to be open source? Their home page [1] has lots of stuff about pricing, sales, requesting quotes, etc., and their GitHub [2] has lots of forks of open source projects but not a main LMS project as far as I could see. It also isn't clear it's an LMS itself, as it has a use case example "Extending your LMS's capabilities", and the pricing page has an "LMS Integration" enterprise feature: "Integration to the Learning Management System (LMS) you're using. We support all major LMS's."

[1] https://www.eduflow.com/

[2] https://github.com/orgs/eduflow/repositories

Correct, we are not open source (I am the CEO). But if that is not a hard requirement, then I do actually think we offer a lot of the things mentioned in the post.

I thought that some of the MOOC companies open-sourced their LMS. Have you looked at https://open.edx.org/ ?

I am doing the odd contracting job for a company that deploys and manages edX for clients. It's not fun. Bloated and outdated..

Question for people who seem to have experience in the space: why are corporate HR LMS / annual certification tools so universally bad?

I've never worked at a company where they've even been easy to use (for common use cases) or cross-browser compatible. Statistically, as a consultant who went through a lot of them, I feel like I should have.

Is it that HR is a cost center and so the market is price-dominated? Or that there's so much regulatory minutiae that change is glacial? Or something else entirely?

most simply, the decision makers are not users. enterprise software is about 50% money laundering

Another open source option: https://atutor.github.io/atutor/

Atutor was my favorite when i researched the LMS space a decade+ ago

Agree with this sentiment having used Moodle and Canvas with school edtech companies (not what they are designed for), and they are really frustrating.

The UX for students is just, sadly, terrible - "old" feeling and difficult to navigate. The UX for learning/content designers is just as bad - a nightmare to create effective content, with poorly considered UI features and constraints.

Hoping to find some great alternatives in the comments.

I work for an HR compliance training company and we use a heavily modified version of Moodle in combination with our own back end. Moodle was originally chosen as an MVP for the company and it is good at playing industry-standard formats, like SCORM and AICC. But Moodle's code base is around 20 years old and very cumbersome to modify so we do plan on moving to 100% custom in the future.

I toured a upper elementary school a couple weeks ago where they were building a remote learning system inside Roblox. I thought that was pretty cool (a little over ambitious for Dallas ISD though).

My son was all about it obvious haha.

Salesforce LMS (trailheads) is really good and although not open source they are planning or already have released it as a product.

I wrote an open source scorn sequencing engine which is really at the center of an LMS and my co-worker wrote a really nice looking content editor. But we were working as government contractors in the ADL office and it got squashed because there was legal concern that the government would be seen as competing with private industry.

I work for another govt agency that would like to hear more about this - been banging my head against the wall trying to convince people that powerpoint with a scorm manifest is not the way to go...

Open source or not they pretty much all such. and I generally don't blame users but In this case it is flat out the users/admins

They are going to contort straight forward concepts into crazy mangled mutant plans

No matter the quality of the product,. The documentation available, the training, the handholding is all for nothing really quickly

You can try this - https://open.edx.org/

I tried Open EdX once, it was unbelievably difficult to get it up and running, despite it being a Django app and me being pretty experienced with Django.

I work for a company that customizes and runs Open edX and you're not wrong.

I did eventually give up!

Don't overthink it. WordPress + LMS plugin + Zoom could get the job done.

Lifterlms is my current favorite amongst plugins.

The thing about edtech is its B2B nature with a mostly public sector that drives it.

Its hard to make a neat product because you have to accommodate the whims of as many institutions as possible.

You might have to bite the bullet with moodle. Its a learning curve but you can eventually beat it into submission to do what you want.

Check out LearnWorlds, they are not open-source but they are pretty good with customization and branding, they've offer community feature, assignments and stuff https://www.learnworlds.com/

GP was asking about open source software

We're offering a number of courses on top of Moodle, and it's not great. UX/UI is terrible, various bugs are haunting us and our users regularly. Moodle seems to be extremely complex, and far from modern in any regard. I have the same question as OP.

It's been a little while since I used it, but I remember Canvas having a relatively nice UI compared to Blackboard, back when I did. Has it fallen behind? What strikes you as archaic in its design? What's nice about Blackboard these days?

Canvas is open source and well documented on GitHub. Both the frontend and the backend are available. https://github.com/instructure

Moodle has adoption by a lot of companies for integration. Having the text book companies have plug-ins that allow battery quizzes and such.

I will agree with others that Blackboard has made it very difficult for entry.

Check edisonos.com! It precisely being built for the reasons mentioned above. We haven’t open sourced it but it’s designed for launching new edtech businesses in no time.

another one:

https://www.sakailms.org/

https://sakai.screenstepslive.com/s/sakai_help

https://github.com/sakaiproject/sakai

Moodle was open I believe? Not great UI, too much config.

Not open source, but Northpass has a pretty robust API so you can white-label your LMS. I've seen it used to power Shopify's academy.

We're working on something related.

https://joinlesson.com

Because there are no good LMSes. It doesn’t matter if they are open or not. The learning platforms people like aren’t LMSes.

Ah, got interested because I thought it was about Lab Management Systems ;)

Learning Management System... Meh. I never managed my learning.

having tried selling a product into education myself (not the college circuit but high school/middle school) i greatly underestimated the difficulty of sales into education which is probably one of the reasons there are not many choices in terms of software made by startups or small teams, open source or not.

Odoo recently added a LMS feature.[0]

0. https://odoo.com

Start writing a good design document, and somebody might implement an open source version of it.

You can use a wordpress based lms, such as Tutor LMS, Learnpress, Sensei lms, etc..

Do you mean these?

https://senseilms.com/ https://www.themeum.com/product/tutor-lms/

Are there any disadvantages of building WordPress based LMS sites?

What’s your issue with Moodle?

> The state-of-the-art seems to be Moodle or Canvas which are designed for universities - they use archaic UI and features.

Let me add 'Blackboard'. What a joke of a piece of software. If I agreed with people who think blackboards are bad and should (have) be(en) replaced with whiteboards, the joke might be 'ha ha it's old and crusty it's even in the name'. I'm a few years out of university so I wouldn't know where to start on its problems - though one thing in particular I remember was it's weekly offline for maintenance window on Wednesday afternoons - but the university paid a disgusting amount of money for it. Lecturers thought it was crap too. The Computing dept. had its own system (built and maintained as a Summer project for undergrads I believe) that did mostly the same (well, the bits it needed of course) that worked great and must have cost ~nothing. They were starting to be forced on to Blackboard as I left.

tl;dr: It's like SAP for universities.

Ah, Blackboard. https://www.lawyersgunsmoneyblog.com/2014/01/christ-i-hate-b...

Not open source, but have you checked Google Classrooms ?

Oh gosh. I could write a whole book on this topic… “SCORM” gives me PTSD, and I’m only barely exaggerating.

Have you looked at Open EdX?

You might want to reconsider Moodle as an option. There's a new version (4.0) coming up early 2022 with a major UI overhaul. It looks like a big improvement, with the UX design at least ten times better than the current version (though admittedly the bar as set by the current version is not very high).

That said, end users (students / learners) need not be bothered by the sad state of the current backend: you can configure Moodle so they'll only get to see the easy parts (I've seen Moodle used in primary education on two different occasions).

As for your requirements, it looks as though Moodle's got them all covered (depending on the details of course). Except maybe for the live classes, which will probably require a 3rd party plugin.

Other things to consider:

- Cost of customization: Moodle developers are harder find than, say, WordPress developers. Putting something together based on WordPress might also get you there, for less money.

- Edtech standards: Moodle supports H5P, SCORM & LTI out of the box (WordPress has limited support for SCORM & LTI, full support for H5P). This means you can integrate your Moodle based platform with a lot of other systems, offering your courses to customers on their own platforms, when they might otherwise not be willing to do business with you.

- Community support: Moodle's the largest open source LMS out there, and it has excellent support forums. There's also commercial support available through Moodle partners (and, shameless plug, parties such as myself, who offer Moodle customization services).

- Shopping cart / payment features: you can put individual courses behind a paywall, but overall Moodle does not really support e-commerce features (except through integrations such as Edwiser Bridge).

- Installation & deployment: since you're looking for an open source solution, I'm going to assume you want to host the platform yourself. In that case, you might want to look for a system that's easy to install and maintain. WordPress & Moodle have got you covered there, since they're based on LAMP, for which a lot of documentation and external expertise is available.

- Enterprise features: do you require features like advanced customizable reporting, integrations with HR systems, and certification programs? Then you might want to look into Totara. I wouldn't call it open source, though you can host and customize the source code yourself once you buy their per seat licensing (still cheaper than most enterprise LMS solutions out there). Totara's based on a Moodle fork, although they're slowly moving away from the Moodle core code base.

[Edited for layout; added clarification on Totara]

The short version:

* Education is socialized. That is, people don't pay for it, instead the state buys it. At university level, this is slightly less true but the state still provides loans which may never be paid back. Even for the rich, who buy their own education, the sector is still extremely inefficient since the main value added is not from the teachers but from the other students. That is, universities are fundamentally clubs (in the economic sense of the term) rather than businesses.

* As a result, there is no payoff for innovating.

* As a result, there is no innovation worth mentioning. This applies to software, teaching, course design and everything else.

* At some point the existing system may become so bad that individual consumers seek alternatives. Then, there may be money to be made in teaching people. Until then, you are condemning yourself to a precarious existence. Good luck!

> As a result, there is no payoff for innovating

I mean thats not true. It means there are different pressures for innovating.

Buying a system in that will reduce the number of staff hours needed to do admin by 5% clear sale.

Buying in staff who cost 3-5x what a normal admin would cost, for a open ended, unsupported software project that has a high chance of failure? not a clear sale.

Schools are not tech companies, they do not have the capacity to spin up a software stack. They want software they can buy, and buy support at the same time.

Buying a system in for marginal gains isn't innovation. I think you agree with my substantive point.

I find statements like this incredibly arrogant. As if education could be revolutionised by some webapp.

Moreover it's also provably false. Education is big business in most countries and a significant portion of gdp (and often not "socialise" either, universities, private schools, private certificate/education providers etc.). Despite of that nobody managed to "disrupt" education yet, because education is difficult and most "small, agile" providers in this space are awful, mainly preying on the desperate.

Regarding good lms, to the OP it really depends on what you want. Systems like Canvas, Moodle are catering for universities/school. The requirements for such large organisations might be very different. Why do you find canvas/Moodle bad for example?

What's arrogant is assuming that we are educating our kids as best as we can already. For anyone working in the sector, that's not just arrogant, it's absurd.

Most developed countries have state-run primary and secondary education systems, and most of them also have highly or entirely state-funded tertiary education. This is a statement of the obvious, I think. I also mentioned reasons why private providers often had little incentive to innovate.

I haven't experienced Canvas or Moodle, but I've experienced Blackboard. Blackboard is beyond dire. See my other comment and the notorious review it links.

I never said that we already are educating our kids as best as we can, so that's a strawman. I actually agree that there's still a lot we don't know about learning. However, I don't believe we we will see a revolution of how we learn, and that revolution will certainly not come from businesses or startups. It requires way too much long term investment with uncertain outcomes.

Regarding LMS, so you know one proprietary system (which is admitably aweful) and based on that you make general judgements of the two open source alternatives?

Right, we agree that (a) there's not enough innovation in education and (b) under the current structures, business won't provide it. The only thing I add is a specific explanation for (b): state control, and sluggish private organizations at the top, provide little incentive to innovate.

The OP suggested that LMS were not good. I provided a reason why that might be so, which fits my experience. I can confirm that the one LMS I know is terrible.

Reply on news.ycombinator.com