How do I startup if I don't own the code? | Hacker News Reader

How do I startup if I don't own the code?

17 pointstricky25 comments

I was recently approached about commercializing some software I created while working at an academic research lab. I was paid with research grants so I'd assume the school owns the code.

I'd love to turn this into a startup... What's the best way to do it considering I don't know who owns the source?

View on HN

If you can rewrite it from scratch, that may be the safest plan. You'll probably need to anyway. You can also go talk to the university about it, but many (perhaps most) universities are cluelessly overreaching about this sort of thing, and take forever as well.

I'm afraid of the overreaching part... I'd love to rewrite, but i'm having trouble justifying a rewrite of something that works really well as-is. Translating it to another language doesn't seem like enough.

i'm having trouble justifying a rewrite of something that works really well as-is

You're different from a lot of hackers I know.

Seriously, though, you probably will eventually rewrite most of it if you turn it into a startup. And often when people rewrite things they're glad afterwards. It will probably take less time than working out a deal with the university, and be much more interesting for you.

Justify it by thinking of the dollars in your pocket. Right now, none, after rewrite, (possibly) many.

Sounds like a justification to me. :-)

Just translating would probably leave it as a "derivative work". However, note that licenses like the BSD and MIT licenses came specifically from universities releasing their code under very permissive terms. Perhaps your university can be convinced to do the same.

How many universities have ever given a startup trouble for making software that was a "derivative work," though? Not one that I know of. In practice it's enough to rewrite under conditions where no one has claim to your IP.

Sure -- a real rewrite would be fine. Porting from, say, C++ to Java probably wouldn't be. (At least I wouldn't found a company on something that potentially shakey.)

I think the point that you're getting at, and I wholeheartedly agree, is that if you understand how something works in and out that rewriting it isn't a big challenge.

It doesn't make sense to sue a startup--it's just not profitable. However, there are instances where universities have appeared to wait for the company or technology to mature before taking legal action. Two examples of this:

University of Wisconsin sues Intel for Core 2 Duo patent infringement: http://www.engadget.com/2008/02/07/intel-sued-for-core-2-duo...

Northeastern University sues Google over Database Architecture: http://www.techcrunch.com/2007/11/10/google-being-sued-over-...

This goes to show that universities aren't always benevolent when it comes to IP. Even if the software is rewritten, there's no guarantee that university lawyers won't accuse you of infringement down the line.

Isn't licensing specifically designed to address this issue of IP? What would be a reason not to license the software from the university?

Neither of these were universities suing their own students.

What would be a reason not to license the software from the university?

(a) That the university might have insanely unrealistic ideas about terms, because they're so inexperienced, and (b) that it could take so long to arrange the deal that it would literally be faster to rewrite it.

Neither of these were universities suing their own students.

That's an interesting point, but 1) would the university not have sued if it were a student (or professor) that provided the IP? and 2) then why shouldn't he just use the code outright, especially since he claims to have written it and the university is not likely to care?

Licensing is there for a purpose--someone paid for him to develop this software, so he has an obligation to that employer to at least ask permission before using the software.

Imagine if a startup took seed money from you to develop their software, then turned around to rewrite it for use in another project and thereby sidestepping your 2-10% share. Wouldn't that be unethical? They could even claim both (a) and (b) above.

(Based on the OP's comment I assume that he is not starting his own company, but rather partnering with someone.)

A rewrite is probably best, but it might hinder negotiation of terms.

-Pro: Wrote the original, lessons learned and experience. -Con: Isn't bringing existing IP so might not be worth as much. -Con: Maybe someone else can write a better clone (UI expert). -Con: Clean room rewrites are tainted if you have seen the original code. Willful infringement is much more expensive.

If it was me I would try and convince the university to release under an open source license. BSD would help, but the GPL could be good too depending on sector. Then I would be free to go off and commercialize it. Open source code requires a different business model, but there are plenty of examples of success, especially for companies that own the original thinking behind the project.

Some schools are really good about this. CMU for example is extremely flexible, even with patents.

But I'd also recommend looking at the terms of the contract that gave you funding.

Government funded projects usually include government purpose rights. It depends on the contract how bad the rights are for a company.

In most cases, government funding means that in a pinch, they can take all your tech and give it to another company to mass produce. The reason would probably be some wartime situation where you needed to produce a million X to fight. I think this happened to B52s in WWII.

Other recommendations about tech transfer are spot on.

You should also take this as an opportunity to network with potential investors. Go to your school's entrepreneurship center (if they have one), and talk to people there about your plans.

I know for CMU, you can meet lots of Angels that way. Lots of the associated companies have probably gone through the same process.

I was paid with research grants so I'd assume the school owns the code.

I don't know what school you went to, but many (most?) large engineering/research institutions have a technology licensing office (e.g. http://web.mit.edu/tlo/www/) that can check that assumption definitively for you.

I checked this out since I wanted to release code I wrote for my research under the GPL: Where I work (UCSC), the employee owns the copyright to code he writes, unless the code was part of a deliverable for some grant (in which case it generally belongs to the grant-awarding institution). For general research grants, that's not the case. If there are potential patents, the university does claim right to those, but if there aren't, there was no problem.

When I worked for a university, I asked my boss if I could open-source my code, which was largely funded by NSF. I showed him a license, and summarized it for him. His response surprised me:

"By all means -- anything 'open' is music to NSF's ears, because it means they don't have to pay for several labs to do the same work. I'll mention that in our annual report."

Talk to the University's tech transfer office.

But don't tell them that someone is interested in your technology. If I were you, I would find out about the status of the IP (who owns, how old it is, etc.) and try to find out if the university is interested in it or not. If they are not interested in it, ask them to sign over the IP to you outright. After you personally own the IP, you can do whatever you want with it, and it will be the universities tough luck that they weren't able to capitalise on the IP when they had it - they've had their chance.

I know of three university spin-out companies that have had IP they developed assigned over to them because the university wasn't interested, upon which successful companies have been formed.

Also, I don't know about tech transfer offices in the states, but in the UK, they are completely incompetent. I wouldn't waste my time or effort in trying to negotiate a licensing deal. You probably could have re-written it by then!

It depends on the school policies. You'll need to just find out what they are. We can't tell you that.

Some schools are quite friendly to business arrangements, while others are grabby. There are plenty of products on the market that were developed at universities and later commercialized.

Talk to your universities office of technology licensing, and ask them to give you an exclusive license to your software. That's what Google did with Stanford.

How long would it take you to redo what you did at the research lab?

not too long, but the concepts would be exactly the same... wouldn't it then become, "who owns the idea?"

There is no such legal concept, at least not in the United States. One can own an implementation of an idea, but not an idea. If you can implement the concepts in a novel way, you're fine.

(In agreement, just to express it another way): Patents on implementations can and do effectively limit idea reuse.

E.g. if the idea is "a web 2.0 flower box" making a second implementation is fine.

If the idea is a "web 2.0 flower box using this specific patented method of procedural generation" then you won't be able to recreate the same product again - at least not without moving to Europe where there are no software patents :)

I don't think that is the case. If you are willing to spend a few hundred dollars you can get legal counselling from a lawyer.

Also, if you are attending startup school, they might have half an hour talk from someone on legal issues and you could ask that person.

have someone else write the code then buy it from them.

Reply on news.ycombinator.com