Good to see this change, but I wish people wouldn't go down this road to begin with.

Speaking as someone who sits in industry now, I have found it almost impossible to successfully conclude negotiations between not-very-interested academics who don't really understand how licensing works (and the staff who supposedly help them) and a big, paranoid company.

When you put some rando commercial license or complicated terms on your (likely tax-payer funded) data or software, what it essentially requires from big companies is a extremely expensive legal review to see what we would be signing up for. This review might be considerably more expensive than the amount being asked for the data/software. What's more, if a small company signs up for this, and ever wants to be acquired by a big company, the review may not happen (small companies and startups are usually more relaxed about this stuff) but will have to take place during acquisition.

Bonus points for when the academics or their representatives wander away mid-discussion.

As a final bonus, I have seen our commercial competitors in this area just essentially fund university researchers to do thinly veiled advertising by accessing the same dataset for an "academic" paper. Meanwhile I sit on my hands respecting "the latter and spirit" of the law.

As someone also on the industry side of things, what would be your recommended solution to the problem? On the one hand we don't want the legal mess that licensing entails, on the other hand I'm absolutely in favor of reimbursing the academics in some way. But how?

Some new, standardized, industry accepted commercial license (who will invest the required legal effort to come up with such a license that's accepted by both sides?)? A concept based on voluntary donations (good luck getting that by the finance department)?

What are these not-very-interested academics expecting? For you to agree to whatever commercial license or...?

I don't get why you'd expect academic institutions to offer you permissive licenses though, when in all likelihood you have your own complicated partnership/IP/licensing processes for industry partners. Why is it ok for them but not for you?

And what difference does it make that their research is tax-payer funded? That doesn't make it free. That means that the institute has an obligation to make returns and, academic research being somewhat open, lock down IP rights with certainty.

Regarding "tax-payer funded data/software" From the blog:

> I believed then, and still do now, that it’s right that institutions that support software development should benefit from its commercial use (UC Berkeley receives 2/3 of the royalties for commercially licensed products), that students are entitled to renumeration for software engineering work that does not directly support of their own research goals, and that funds are needed to support specialized personnel who can maintain/improve code and service user requests.

Also i believe, most students in academia are under-paid.

It's off-topic but a very similar argument exists for why the Facebook license + patent type arrangements(1) can cause problems for large companies or small startups that may want to be acquired later.

And I too wish that the Facebook-of-the-future don't go down this road. Number two on devs wishlist should be transparency about exactly which patents cover the code at hand. IMHO.

I draw the analogy due to the similarities of "open source" and government funded findings. Another area where these issues arise is charitable contributions that come with strings attached (like naming a building or agreeing to continue a certain program).

People like getting full recognition for the act of generosity without the cost of their strings being acknowledged or subtracted from the total.

(1) The reality of Facebook is that it's unlikely to ever be an issue, but if the arrangement becomes more commonplace the open source community could become a minefield of unquantifiable risks. I realize this is a slippery-slope argument and is thus easy to overstate.

oops "the letter and spirit" of the law.

Follow up to my other comment...I just noticed from your bio that you work at Intel, and now have to lol at you complaining about academics trying to license commercial products. I presume you don't mind when tax-funds are directed towards academic R&D when it's conducted for Intel (and plenty of it is in Europe) :)

I don't understand why you think you were wrong, based on the article. Surely it depends upon your objectives. If your goal is for researchers to use the software then one license might be preferable, if your goal is commercialisation then another license might be.

In the same way that some researchers wont use software which doesn't have permissive licenses, some companies wont use software which does.

I don't get it. I also don't get why you expected license agreements to be signed and completed in a single day by commercial entities. You couldn't sign and complete a 5 minute coffee date with most company's IP staff within a single day, nevermind have licensing arrangements agreed upon.

The wrongness I read from it is a certain level of arrogance, that there can't really be anything too complex in licensing and copyright issues, and all those lawyers just must be posturing or engaging in pointless make-work. The discovery that well-understood standard license terms are preferred for a reason, and that there are very real issues lurking in non-standard license agreements, and corresponding realization that maybe all the people saying "you're wrong" had a point... is a step toward correcting that arrogance.

That makes total sense but it sounds as if his realisation tended towards open licenses being preferable in all academic scenarios, due to a now-more firmly held belief that lawyers etc are a waste of time. This obviously isn't a rule which applies generally, and I'm not sure why it even applies here when his goal was to commercialise, not to please researchers.

I'd like to commend the author for realizing that a commercial license was not achieving the goals he had for his software and being willing to own up publicly about why and how he changed directions.

However, while I have used MIT/BSD-style licenses in the past, I can no longer recommend them, as they potentially open up software users to malicious submarine patent attacks. That is, a contributor can intentionally contribute code that (purports to) read on a (not-even-issued-yet) patent, and then later threaten to sue users of the software.

That's why I instead recommend Apache-2.0. My arguments are more fully described here: https://www.cncf.io/blog/2017/02/01/cncf-recommends-aslv2/

Interesting, has this actually happened?

Yes, a patent-owning entity has tried to contribute code reading on that patent. Apologies that I can't give the details.

I've been about half a block down this road. Thank God, the university office determined it wasn't worth the universities time to invite the lawyers -- and my project is happily open source.

One of my least favorite experiences in grad school was dealing with our school's tech transfer office. They fought hard to avoid using the GPL for licensing anything. Just getting them to agree that a simple (to be published) tool should have an open source license was difficult. (And took many meetings)

Fast forward to two different Bay area universities and it was a completely different experience. Open source licenses were agreed upon over a quick email.

I had a fun time with the library copyright office. 6 months before my thesis defense I emailed them with "hi what's policy and procedure for creative commons licenses on theses". Nobody had ever asked before...

edit: my thesis is cc-by licensed.

Is it traditional for students to sign away copyright to their thesis to the University? That seems outrageous to me, but perhaps I am missing something?

One of the reasons I never continued on to do an Honours degree is because the University would have retained the copyright in my work, and the rights to license it commercially. I thought it was outrageous as well. But that was over 15 years ago, so maybe the situation has changed since.

traditionally you register a personal copyright. it's implicit if you register nothing.

some students may choose to request an embargo on their thesis, in those cases where their thesis work will be published elsewhere (in a peer reviewed journal, usually), or if their thesis would constitute disclosure of something for which a patent application is underway.

my thesis was boring enough to avoid those cases.

If you retained the copyright personally, why would you need permission from the university to publish your work under a creative commons license? Was it just a courtesy email to let them know your plans?

It was a courtesy, and also a formatting question.

Because it's going to wind up in the library archives (and proquest) for ever and ever, the university is quite strict about formatting. I checked about 6 months ahead of time to make sure that replacing the (C) symbol with the (CC) symbol on the copyright page wasn't going to cause any issues. If a thesis is not properly submitted to proquest, you don't graduate, so... wanted to tick that box off pretty early.

I can't speak for OP, but in my experience it's very rare for a thesis to be just your work.

It's based on data and research often paid for by the university, and includes the work of many different people. indirectly: advisors, lab partners, professors, other students. Directly: Co-authors of papers, research assistants, and other contributors.

It's a messy bundle of work.

But copyright isn't based on ideas and doesn't cover raw data/facts- it covers the expression of those ideas and the compilation of the data.

Lucky you, I signed over a document assigning copyright to my university I believe. :sad face:

I was an onlooker -- for four months -- as a university IP office struggled to license source code developed at the university, to one of its own researchers. When the drama was concluded, the IP office couldn't find the source code.

BSD does not have a patent license clause so if Lior, et al. have a patent on any part of the algorithm then won’t commercial use will still be ambiguous? And even if they don’t it could still be using (infringing) on a patent of a third party. Basically software patents make using any software a risk but possibly more so when you’re using something based on novel ideas.

Edit: GPL and ASL have explicit patent clauses (obviously only applies to patents the software licensor has authority to license).

Some context:

The blog author just wrote another post which was essentially a hatchet job on a 'rival' software package. One of the major differences between his software kallisto and the rival Salmon was the license - Salmon has a GNU General Public License.

'I was wrong' part 1 was a really a discussion about how wrong / fraudulent some authors were who published in Nature in 2004. It looks like it turned into a bit of flame-fest[1], with the original authors chipping in.

[1] https://liorpachter.wordpress.com/2015/05/26/pachters-p-valu...

No I mean this blog post: https://liorpachter.wordpress.com/2017/08/02/how-not-to-perf...

Thanks for sharing your painful journey!

I can't take credit, I just follow Lior's blog :)

I'm not sure that's so true. Everyone enjoys a good argument over what license the author should have used. )

I do get what you're saying but I think people discuss positive things all the time on even the worst sites. Admitting you're wrong isn't all that novel, but the changing of license is.

What are you trying to argue? Can't people on the Internet both enjoy telling someone they're wrong, _and_ be interested in someone saying they were wrong, at the same time?

I Was Wrong (Part 3)

So you're saying this was a trap?