I don't hate the font so much but the stroke is too light to read easily and the "a" characters look like "o" characters. So it's more a couple of implementation details that make it hard for my old eyes to read.
So the idea is that they are so confident in their human capital, clientele and business processes that even with all of the code no one could potentially compete with them? Or that they want to create more competitors?
This is actually a very good move for them. The type of person who could take the full code dump and stand up the service is likely the type of person who would just write it themselves.
They're selling an operational service that their customers can't easily replicate, even with the code.
Their product is a globally-distributed monitoring system. Running these systems all over the world, and keeping them running, is a serious operational headache. It's probably at least as complex as operating the service the customer wants to monitor - perhaps a lot more, given that most services are not globally distributed.
As long as they price below "double your ops budget", they'll remain value for money - even if the code they use is free.
I have mixed feelings about it, it seems to me giving away your code for no return indicates you don't really value it that much. Whether it will have an affect on the industry I don't know.
There was an interesting paper linked on HN a couple of years ago that indicated a lot of the money is disappearing from software sales. The inference was the availability of mysql depresses the revenue from MSSQL, the availability of Linux depresses the revenue from Windows and so on.
Software isn't a brick though, it is changing and updating and is a brick that has to be maintained. When you put software bricks together you get a construction that is more powerful than the sum of its parts. It is something inherently valuable in and of itself.
If I need 1000 software bricks I need for my house, you can’t really convince me to pay a premium for each. Especially since I often have alternative bricks.
Psychology plays an important part in how markets work.
I think a lot of capacity disappeared after the financial crisis and demand has shot up in the last couple of years due to relaxation of the planning laws.
No - building materials in the UK have gone up in price because house prices have gone up and interest rates are low. It's easy to borrow money and the ROI on building is high, so the merchants can charge more for their goods without making a significant dent in the buyers budget.
> it seems to me giving away your code for no return indicates you don't really value it that much
Well of course it seems that way to you, because you have defined it circularly. Open sourcing code is in no way "giving it away for no return". There is massive return on open sourcing code, from money to improvements in the code and all other manner of things.
I have the opposite perspective from you. I think that closed source code indicates you don't value you it very much. But I understand that perspectives are different and that's okay. But you have made your own leap to not valuing code very much but you should know that the truth for a lot of people is the exact opposite of what you wrote.
I'm somewhat dubious about the financial aspects, sure you can start offering support services but who wants to do technical support for a living, or advertising, or begging via Patreon.
Money is a good proxy for value, I know the customers who paid for my software today are getting some benefit from it, they wouldn't have put their hand in their wallet otherwise. Every time I get an order I get a warm feeling that goes well beyond the financial reward.
On the other hand some of my worst and most needy "customers" have been those using open source or free tools. Their expectations are high and gratitude low.
As I said I do have some open source projects and they have received contributions, when I add it up though it hasn't paid off.
Two reasons to open source: either you're doing it as a hobby and want the world to see it/don't mind sharing it, or you believe/know that open source contributions will be better than what you or your company can do.
Otherwise, open source is terrible for making money. Most clients don't care, but you're giving your competition everything in exchange for nothing.
It's great for society, though.
Just my opinion, and I know a lot of people share it.
I work as a software developer who's pushed for open sourcing code at our company (unsuccessfully). Premise is that the company's value isn't code, & the idea is that making it open source will produce a more successful product. Generally a software developer's job security correlates with a product's success. It may mean we don't need to hire more people, but I'm not looking to work at a large company, & less people means less mouths to feed
In my own experience I was trying to demonstrate that we didn't need to bastardize the codebase into a million pieces in order to give remote developers only the small part their working on; that we might as well make the code publically available
Certainly, but not in the direction I think you're implying. Open source means that you know how to solve all the problems MySQL has solved, even if you don't actually know how to implement a good database engine.
I'm going to hijack this thread to mention that I'm thinking of open sourcing a SLAM system I've been working on under BSD or MIT.
Ive been working on this because I'm interested in underwater drones (uuvs), and hope to commercialise this in the form of services.
My current understanding is as follows:
- most open source systems come from universities and are GPL, with the option for licensing. (PTAM, ORB SLAM, DSO, SVO, LSD)
- Optimistically, it would require $1-3 million to build a compedative system privately. The sky is the limit here.
- No one, neither the universities nor private entities, is making serious money by licensing just a SLAM system. The well performing startups are somehow coupled to a product or a service.
- All systems basically solve the same problems, but mix and match different solutions. For example for initial motion estimation, you could rely on constant motion (DSO, ORB SLAM), image alignment (SVO), or solving directly for rotation between cameras (GSLAM).
- As a researcher, to implement a new idea, you either need to implement an entire system from scratch, or modify a GPL'd system. This is great in theory, but since everyone is playing the GPL or pay game they are reluctant to accept contributions.
I'm imagining something like opencv, with implementations of common algorithms accompanied by wiki pages describing high level details. Components could be mixed and matched into different reference SLAM systems targeting different use cases, do you want something that will run on a MAV or a self driving car?
From a commercial point of view, companies could worry more about tweaking things to work for a particular use case without needing to start from scratch. I believe that universities, with their many departments, have more opportunity than anyone else.
My belief is that it's time for a liberally licensed SLAM system. Has anyone come to a similar conclusion or care to comment to the contrary?
Can companies really get away for not licensing properly with GPL? I mean selling your product without buying the commercial license. How to get yourself caught? Different SLAM solves the same problem anyway.
It sounds like they own the copyright and are dual licensing. However they don't accept contributions (without a copyright assignment) since they wouldn't be able to dual license the contribution.
I was thinking something similar. OpenCV has lots of pieces of the puzzle, but the quality is... inconsistent. For instance, I've had repeatable segfaults from doing RANSAC (which of course disappear in the debugger and Valgrind). There have also been some significant feature regressions since the 2.x --> 3.x migration, and the build system breaks every time MSVC upgrades.
If you'd like to release something like this I'd be happy to contribute. I have a background in SLAM, embedded and photogrammetry, so this is something I could really get behind.
I was working with Opencv's python wrapper a few months back and I didn't find a lot of options in it. It maybe because the wrapper didn't include most of the things from native C++ code. Is there any proper way to start with Opencv?
If you have a robust SLAM method which is relatively computationally efficient, and you release it under an unhindered or attribution license, I think a whole bunch of people would use that. I know I would.
I've been wanting to spend some time in that field myself but workload hasn't allowed it. Just out of curiosity, have you looked into DTAM (PTAM's successor) or RATSLAM? Both were on my shortlist for a future project.
59 comments
[ 8.9 ms ] story [ 140 ms ] threadOn the other hand, man that custom font is tough on the eyes.
Proofreading works, everyone!
Their product is a globally-distributed monitoring system. Running these systems all over the world, and keeping them running, is a serious operational headache. It's probably at least as complex as operating the service the customer wants to monitor - perhaps a lot more, given that most services are not globally distributed.
As long as they price below "double your ops budget", they'll remain value for money - even if the code they use is free.
There was an interesting paper linked on HN a couple of years ago that indicated a lot of the money is disappearing from software sales. The inference was the availability of mysql depresses the revenue from MSSQL, the availability of Linux depresses the revenue from Windows and so on.
Psychology plays an important part in how markets work.
Well of course it seems that way to you, because you have defined it circularly. Open sourcing code is in no way "giving it away for no return". There is massive return on open sourcing code, from money to improvements in the code and all other manner of things.
I have the opposite perspective from you. I think that closed source code indicates you don't value you it very much. But I understand that perspectives are different and that's okay. But you have made your own leap to not valuing code very much but you should know that the truth for a lot of people is the exact opposite of what you wrote.
Money is a good proxy for value, I know the customers who paid for my software today are getting some benefit from it, they wouldn't have put their hand in their wallet otherwise. Every time I get an order I get a warm feeling that goes well beyond the financial reward.
On the other hand some of my worst and most needy "customers" have been those using open source or free tools. Their expectations are high and gratitude low.
As I said I do have some open source projects and they have received contributions, when I add it up though it hasn't paid off.
Otherwise, open source is terrible for making money. Most clients don't care, but you're giving your competition everything in exchange for nothing.
It's great for society, though.
Just my opinion, and I know a lot of people share it.
Open source is not actually giving away your code for no money, let alone no return. It's not free as in beer.
In my own experience I was trying to demonstrate that we didn't need to bastardize the codebase into a million pieces in order to give remote developers only the small part their working on; that we might as well make the code publically available
Don't forget, the vast majority of open source is paid for by companies.
Ive been working on this because I'm interested in underwater drones (uuvs), and hope to commercialise this in the form of services.
My current understanding is as follows:
- most open source systems come from universities and are GPL, with the option for licensing. (PTAM, ORB SLAM, DSO, SVO, LSD)
- Optimistically, it would require $1-3 million to build a compedative system privately. The sky is the limit here.
- No one, neither the universities nor private entities, is making serious money by licensing just a SLAM system. The well performing startups are somehow coupled to a product or a service.
- All systems basically solve the same problems, but mix and match different solutions. For example for initial motion estimation, you could rely on constant motion (DSO, ORB SLAM), image alignment (SVO), or solving directly for rotation between cameras (GSLAM).
- As a researcher, to implement a new idea, you either need to implement an entire system from scratch, or modify a GPL'd system. This is great in theory, but since everyone is playing the GPL or pay game they are reluctant to accept contributions.
I'm imagining something like opencv, with implementations of common algorithms accompanied by wiki pages describing high level details. Components could be mixed and matched into different reference SLAM systems targeting different use cases, do you want something that will run on a MAV or a self driving car?
From a commercial point of view, companies could worry more about tweaking things to work for a particular use case without needing to start from scratch. I believe that universities, with their many departments, have more opportunity than anyone else.
My belief is that it's time for a liberally licensed SLAM system. Has anyone come to a similar conclusion or care to comment to the contrary?
Thanks HN
I'm not really familiar with Python yet and try to find some example of these code.
Thank you.
If you'd like to release something like this I'd be happy to contribute. I have a background in SLAM, embedded and photogrammetry, so this is something I could really get behind.
I've been wanting to spend some time in that field myself but workload hasn't allowed it. Just out of curiosity, have you looked into DTAM (PTAM's successor) or RATSLAM? Both were on my shortlist for a future project.
I'm the guy behind latency.at. So far it's really just a little side project.
Since few people asked: I mainly open sourced it because I believe it's useful to have a complete, real life service as an example/inspiration.
https://gitlab.com/latency.at/latencyAt/issues/36
:)