Sign up or log in to submit a post.
I have been thinking whether to open-source my project, and why anyone would. Are there any obvious advantages/disadvantages when considering going open-source with your startup?
Unless you have amazing tech, you're highly unlikely to get problems found and fixed by the community. If it's technically challenging, it's likely to be too hard for most of them and if it's easy, it's not sufficiently interesting. It is a big commitment so needs to yield a notable marketing benefit.
Speaking from the experience of having released multiple libs of my own as open source, one of which was used as a key part of Valve's Steam service but they didn't even bother telling me - I found out from a Dr Dobbs article that mentioned me.
More recently (2015-2017) as a member of the Realm mobile database Xamarin SDK team - most of what Realm did was public on github and we still got almost no contributions. Realm mostly did it as a credibility and defensibility tactic so people felt safe adopting their engine.
If you just want the marketing benefit of having open source, go ahead and release demo projects of bits of tech you use, eg:
These three repos are my technical exploration areas. They are deliberately small samples but useful in their own right to exercise a key API or technique. They demonstrate technical proficiency if someone's wondering about our core stuff but are obviously not part of the main codebase.
Extremely helpful, @Andy.
Valve seems like a junk shop - no wonder they don't believe in crediting work.
Also, great tip on demo projects - I suppose that's how most big-tech do it too, for engagement. More recently, OpenAI's GPT-3. OpenAI had a history of being "open" but they decided not to disclose the code, training data or trained result this time.
Have you seen a change in the open source community in recent years? Have malpractices become common?`
Edit: what if you only went open-source with a group of 20 companies (who are all open source themselves) and using similar stack? would that change anything?
I don't think you quite understand how open source works - the open bit is the key. Once released under an open license, you can't stop stuff propagating.
Nothing stops you having a private agreement to share source with other companies. Back when I was doing my C++ dev tools, I sent source code out to licensees because C++ didn't have binary library compatibility. It was common for OO frameworks to ship with source code included - Apple's Cocoa was the first time I encountered a framework that lacked source.
I'm just not sure what you're trying to accomplish here. If you have people who want you to share source for various reasons, then sure go ahead and put some stuff on github. I'd be careful about the Contributor Agreement as you get to define what is expected and good behaviour. Also, if they are pushing you to provide the source, push them to provide the community support people who will jump in and help moderate if things get rowdy. Maybe make it a condition that each company maintains a designated contact who is involved in the open source.
I should have clarified. It is nothing to do with my work yet, but an active area of exploration.
Indeed, the reason I asked the question is because I don't understand the open-source world enough (neither many contributions nor business exposure), so I sincerely appreciate the insight.
Again, an interesting perspective from you - I did not know that shipping source code was common. At present, I quite like the spirit behind open-source but regret that it doesn't work for the creator in most cases unless they have **amazing** tech. I reckon the problem is somewhat analogous to P2P, napster etc and I'd love to see it solved.
For working between a group of companies, the thought is that one could restructure the incentive mechanism so that there is a usage/contributions agreement where businesses either pay a fair amount, or contribute. In practice, it may be difficult to execute.
I've been a professional programmer for nearly 40 years so have seen a lot of the growth of open source from its academic origin where everything was naturally shared (early Unix days) to how corporate licensing kicked in and then the open source movement became a counter-culture.
For your group idea, sharing source or any other digital goods is just something you are licensing with various IP constraints.
The key question, as I've been reminded a few times is simple: what is the value proposition for the other party, vs their risk?
Remember that most people have some kind of bucket representing their amount of mental effort they will put into things - often we have to deliver a clear perceived value that is surprisingly large, to get them to commit some of that reserve. Anyone who is sincere about a relationship going into it has to consider the time and energy they have available to live up to commitments in future. What is their exposure if they want to get out of it?
These kind of social/psych issues often prevent individuals getting involved in open (or shared) source projects.
For organisations (not just companies) there is also the liability exposure. If they are delivering any kind of good or service that relies on this shared source and something goes wrong, how are they protected?
For example, I know one vendor who had to develop their own libraries for processing Unicode conversion and sorting of characters (non-trivial) because the available libraries violated some of their down-stream contracts with clients.
Those are some really insightful points, Andy.
Interesting how the spirit of academia devolved into corporations. Actually, that's one of the things that prompted me to ask this question - I enjoyed my time at university and the environment of collaboration. Getting in the business environment, I can also see why businesses are protective of their IP (that's the only thing making them money).
In the middle-ground scenario, the benefit for companies to commit mental effort would be to leverage a ton of work that they don't need to repeat. [Assumption: companies do a lot of repeat work]
This would be delivered similar to how a bittorrent keeps the network alive by seeding; but through contract/social-obligation. The question is indeed how well that pans out in practice.
Liability exposure is potentially an issue in some narrow cases like the one you mentioned and I am not sure on the mitigation strategy there. It's something to think about.
* not entirely sure on the open source angle for matchmaking talent lending between companies?
* any specific markets can facilitate for early traction?
- nothing to do with talent matchmaking (or my business) yet, but I am exploring this as a curiosity.
The first question in the list is why companies choose open source, and why many don't. Weighing the risks and rewards from different perspectives. I'd appreciate yours.
We open-sourced my last project Dragonchain (https://github.com/dragonchain/dragonchain) under a modified Apache v2 license.
Dragonchain started as the Private Disney Blockchain Project, and was released to our Joe Roets, who was the Architect on the project at Disney. It was released to a Foundation and the code was re-factored in private code. We later decided to open source the code again. A couple of things:
- It help accelerate our developer community
- It builds trust in your code
- It removes barriers during the sales process
There are some well known challenges (Elastic.co) to dealing with the Cloud, where AWS will host the open source for less than the vendor can.
A good SaaS model is Databrick's. It does not include pricing for any required cloud resources (e.g. compute instances). This allows for alignment between the cloud vendors sales staff and the SaaS providers sales staff.
Do you have specific questions?
I started deeplint.com with open source. Not sure why just thought it might help more people. Maybe good, maybe bad.
deeplint looks cool. I have seen great successes in security, infrastructure and cloud tools being open-sourced. Okteto is one example, Dgraph is another - both of which we are using.
I would love to know how it works and contribute. I am currently working on Go, but quite familiar with typescript. Do you have a demo (in-action clip)?
What's your startup? I'd imagine there are some businesses where being first and fast would be advantageous. Especially those that can be easily replicated.
thanks. my startup is wolv.io -> it's a tool (and network) that allow companies to borrow or lend talent from each other. We already have initial customers, investment etc so I was thinking of open-sourcing most things. So far
1. make us more resilient - problems can be found and fixed by the community
2. be transparent
3. move faster - feedback, ideas, and contributions from all sorts of intellectuals would be great
1. no secrets - some things that could be monetised can't be?
2. harder to manage - there are some standards to follow while interacting with the OS community?
Hey Rishabh! We have looked into a few potential open source models. https://blog.timescale.com/blog/how-open-source-software-makes-money-time-series-database-f3e4be409467/ is a really good business model-focused starting point for analysis.
It is easy to underestimate the cost of doing OSS well, especially if the open sourced project starts to succeed. If you don't dedicate adequate resources to community management and developer relations, it's likely to be a business liability more than an advantage. It's also very hard to "take back" if you later discover a theoretically better business model; http://www.defmacro.org/2017/01/18/why-rethinkdb-failed.html is an instructive post-mortem in that regard!
Most businesses that have open-source projects make money off providing support to enterprise customers
Yes, correct. Though it’s hard for me to see why something like Facebook (or most consumer businesses) can’t be open sourced?
Once the network is big enough on anything, it becomes hard to replace (or copy)
Social networks built on top of crypto open source all of their stuff. See mastadon specifically.
In general open sourcing most of your stuff is probably fine if not more advantageous. You will be able to get community help to improve the project and also be able to use it as a recruiting tool. GitLab is another great example of an entirely open source company.
Thanks, apache8080 - mastadon looks interesting. Do they have a critical mass?
Given your points, have you personally considered going open source? What are you working on?
Sign up or log in to upvote.
Sign up or log in to reply.