Ask HN: To go open source or not? I can't make up my mind

Unless you solve a specific problem that is worth a lot of money to businesses, the intrinsic value of your software is probably much smaller than you think. Most of the value in most software businesses comes from development, design, optimisation, sales, marketing and so on. In particular, selling software to large companies is a slow, expensive process that has little if anything to do with the quality of your software. You mention AI and NLP in the comments, which makes me slightly nauseous as regards the actual commercial prospects of your software - such projects tend to be academically impressive but monstrously expensive and near-impossible to market. Obviously I'm purely speculating, but I'd suggest getting your product out of academia as quickly as you can and talk to your potential customers. It's very easy to waste years of your life and a pile of VC money on a brilliant product that nobody wants. If you haven't read Steve Blank's "Four Steps to the Epiphany", stop reading this and go buy a copy now.

Open sourcing a project has many potential benefits. It looks excellent on your CV and is usually the best way of getting hired by someone you'd actually want to work for, as it proves that you can actually program. If your software becomes widely used, there are excellent opportunities for consulting, writing proprietary add-ons, books, speaking engagements and so on.

You say "i don't want some big company coming and taking my source and putting it into their product even if they do credit me". In fact, that's exactly what you want, because overnight you become the world's leading expert on a product by a big company, with everything that entails. GPL it and they can't distribute the software without also distributing the source. They can use the software internally of course, but guess who they're going to want to hire if it becomes mission-critical? That's the basic reason to open source a piece of software - it opens the possibility for other people to develop, market and become reliant upon software you wrote.

If you're willing to bet every penny you have and every minute of your life for the next ten years, then make a business of it as a proprietary product. You've got a great deal to lose, but you could end up richer than Croesus. If you open source it, all sorts of opportunities may open up for you with very little effort on your part.

That bit.. But dammit I sure as hell appreciated it. You brought me one step closer to getting to open source. I just posted a description of the software I'm going to build in the comments section so if you could give further advice. Thanks again.

Professors of computer science arent always the best people to take business advice from.

First question to ask is do you want to make a business from your project? if so, then you really need to worry about IP rights, which are commonly given to the university if I am not mistaken.

Assuming you do want to make a business and IP issues can be worked around, then you can begin to question whether open source is best or not. Without knowing more about the project its impossible to give any suggestion either way, all I can say is being open source does not mean you cannot build a business by any means, a lot of people conflate open source = no money but that is not at all true.

As far as IP rights go a lot of the project will be what I do after I leave university. But I've already spoken to heads and have worked around that issue. In fact they'd be my first investors, but advised against open source. So IP is workable.

As for what the project is it's pretty hush hush right now before I decide anything but to give an idea of size it would include AI and an NLP engine built into it. It's got a lot of business potential as well. If that would help with an idea of where I should go?

And again. Question is. How do open source teams actually make money out of projects like this?

technical support packages sold to enterprise customers.

And customization services.

You can make a lot of money on open source (Red Hat, Wordpress.com, etc), but you have to approach it from a whole different perspective and business model.

What is definitely a lot more difficult, even impossible, is the insane level of valuation that closed/proprietary companies have (ie, facebook, twitter, etc). Which means that your valuation will have to more accurately reflect your revenues and profits.

Which may not be a bad thing. It definitely makes it harder starting out, VC's are still very wary of open source, but it keeps you more honest in the long run.

If you're looking to get big fast, and cash out early, go proprietary, and ride the wave. If you want to build a solid company, you can definitely do that with open source.

It would really help if you described the idea here. If you are considering open sourcing the project itself, then I am sure you can describe what it is without being afraid that someone can run off and copy the idea.

I just posted a description in the comments above. Thanks :)

Without knowing more about what you're doing I'm hesitant to give advice.

However what I would say is if you think the work is something that can be applied more or less directly to real world problems, represents a significant improvement on the current state of the art and you have potential investors, well personally I wouldn't open source it.

If that's not the case open sourcing the implementation might make sense. One advantage to you is that it might help you get a good job, if that is your goal.

If you want to go Open Source, go by all means. You can still make money by selling enterprise solutions to big names out there.

Think about it this way, if this project of yours becomes a big name, then you will be a respectable figure. And then it will be easier for you to raise funding or whatever for any of your project.

But you won't need that really, open source can support the founders - it almost always does.

ok. any advice on licenses? i don't want some big company coming and taking my source and putting it into their product even if they do credit me.

In theory using the GPL should prevent that. It might be difficult to enforce though if the code is something that could be used internally or otherwise effectively hidden.

Is this an ethical or pragmatic dilemma? If it's the latter, going open source could be a great or not-so-great business decision, depending on your business model. So you should tell us more about your project.

If it's the former, you can find plenty of debates on the ethics of free software online.

Hmm... I suggest you keep it closed and try to make some money out of it (don't look at me like that, Mr. Stallman) and if doesn't work out just make it open - and should it become sort of a success make some green with it indirectly (e.g. get nicer jobs).

What is it anyway?

I can't say just yet what it is just yet. Not until I've made the first release. Yes. I'm paranoid after hearing the stuff that can happen :P. I sort of put this question up looking for some therapy as well :)

My personal opinion is to "open-source" (or, more technically correct, free) it.

If I say anymore, I'll get DNed. :)

"I still have this urge to go open source" -- perhaps you try the Freemium Model ?

Ok it would clearly help to tell what the idea is. So what I am building currently is an engine and a commercial level software on top of it to automate the checking of stylistic consistency in a document. The application of it is pretty broad but to sum it up into a few points

1) Plagiarism detection in education. It can be used as a means to detect plagiarism of individuals outsourcing their documents (thesis) through identification of differences in style (inconsistency).

2) Forensic studies. Letters submitted to court have often been suspected of being doctored by others determined to put a suspect behind bars at any cost. Can help there by identifying inconsistencies in style.

3) Literature studies. Educational tool to help professional creative writers improve their consistency in style

This can be expanded in many other ways, such as history where the software could aid in solving mysteries of disputed authorship of texts. The possibilities grow even more if you think of it merging with other types of bio metric studies.

Right now what I am building the software under is the 1st point and it is titled Multiple Authorship Detection Automation. Hence the interest from the academic field :). Ok. Hit me now.

It sounds like a great product, but the two potential fields that you just identified are 1) Education/Academia and 2) Government.

There are people on HN that are more qualified to talk about this than I am because they have more experience with selling in those markets, but this is what I understand:

You can't sell to either without a strong salesforce. These are not the sort of environments where you get paid a lot for technology just because someone discovered it while browsing the web - sure there might be one-two professors, and one-two judges, but most will expect their departmental machinery to take care of this stuff for them. The essence of the problem is that the people who are interested in using the product are different from the people who hold the purse strings, and both of those are different from the people who actually oversee the cycle of purchasing and procurement.

So, in a certain sense, your advisor is completely correct. You could really make big bucks selling this. However, in order to get there, you would either have to learn the microstructure of sales cycles and handle that all by yourself, or you'd have to partner up with someone who understands that and knows how to operate in it. Whether you want to give that a shot is completely up to you. It could be a very valuable experience.

Remember, there are a lot of problems with going open-source as well. First of all, you think "wow, I released this piece of software, awesome, now people around the world are going to gather together, hack on it and send me patches". That rarely happens. What often happens is, you end up getting requests for free tech support and bug fixes, and the usual comment of "well, it's open source, you could help me hack it" is met with a dismissive comment. In other words, once you go open-source, there's still a lot of management that you have to do in order to steer the product in the right direction, except now you aren't getting paid for it. Open-source really is a labor of love - and I don't mean loving your beautiful, smart and gifted children - I mean loving your developmentally disabled, Down-syndrome, wheelchair-bound children.

So, my advice would be, don't go open source in the beginning, and try the "build a company out of this" a shot for six months or something. Expect that in six months it will go nowhere and you can open-source it then, and see whether once you open-source it, it gets a different kind of pickup, and maybe you can still build a different company out of it then.

hey just wanted to say thanks and that i had a final chat with people around the project. Their advice was not to go open source but with a completely rational reason. If this was open to all then students whose projects are to be checked by the system, those same students could download the source or the binaries and then try and game the system to accept plagiarized documents.

Thanks again all the same and I will use your suggestion for a bunch of other ideas and projects that I will be starting work on very soon

I am sorry, but that's an absolutely ridiculous, irrational reason. Your system is, presumably, a large piece of complex technology. In order to understand how to circumvent it, a significant amount of time must be spent understanding what it does. Students that want to submit plagiarized documents are lazy, and if they would stoop so low as to plagiarize, they would never in a million years spend the time figuring out how to circumvent your system.

haha fair enough. maybe im just paranoid about how the minds of people work as ive seen it. We use turitin at our university to check originality of work of students. That takes a document and matches it against a huge database of their own that they maintain. Students can't use that service (unless they pay the fee which they mostly don't although I know a few who have done so) because they would look at the areas that have been marked as unoriginal and would make a few changes to make it go undetected.

at least that's how it goes. The system I have right now so far also highlights which areas specifically does the style change and in what way and by what percentage. It's being built keeping in mind a user base that should be able to get information with just two clicks . So it shouldn't be too hard to change the style here and there to game the system.

But I do raise hands in acceptance of your viewpoint here and call a truce (yes drama much) until I actually build a product before really making a decision on everything. But I do honestly value your input a LOT. thanks again :)