Hacker News new | ask | show | jobs
by hf 4148 days ago
I freely confess to being flabbergasted by these displays of less-than-rigorous thought processes.

How would a free software project 'pay it forward'?

They are in a very similar position, aren't they?

Edit: For some reason, I can't reply to child comments (probably a cool-off time-out at work?).

Just a short note here, then: $1.25e6 for the FSF translates to 10 developers like Koch being paid (the donation page quotes "120000 EUR").

That's ten. For the whole FSF. As an example of a well-funded project. I'm not going to comment on that. HN would rightly give me months of cool-off time.

5 comments

There are many free software projects that are decently or well funded. They have no problem meeting their donation requests, and having a good budget year over year. These projects are usually end user facing, in a way that their dependencies aren't.

It seems reasonable that these projects should consider adding items to their budget to redistribute funds to projects that they depend on.

Some probably do this, however I think the GP was suggesting that something like this become more common.

> There are many free software projects that are decently or well funded. They have no problem meeting their donation requests, and having a good budget year over year.

Too bad OpenSSL wasn't one of them until after the big "heartbleed" incident.

The core infrastructure projects don't seem to get as much funding as they ought to, especially given almost everyone relies on them (even if they don't realize it).

Prior to Heartbleed and the industry rallying to fund critical projects, OpenSSL only received an average of $2,000 USD a year[1]... that's pathetic.

[1] http://arstechnica.com/information-technology/2014/04/tech-g...

Arguably, OpenBSD deserves that funding more for the ongoing process of fixing the OpenSSL codebase than OpenSSL does for making it as awful as it is.
Over the next three years, the Linux Foundation will receive a combined total of $3.9 million from Google, Intel, Amazon and others to fund core infrastructure projects such as OpenSSL. Sounds good until you take a step back...

> Intel will invest "$300 million to help improve the pipeline for women and minorities, actively support the hiring and retention of diverse candidates, and fund programs that support the positive representation of women and minorities in technology and gaming industries."

http://www.wired.com/2015/01/intel-diversity/

> "Google Gives $775,000 to Nonprofit for Tech Diversity CODE2040 said Monday it received $775,000 in grants from the tech giant to support the launch of free training programs for more than 5,000 black and Latino college engineering students over the next two years."

http://blogs.wsj.com/digits/2015/02/02/google-gives-775000-t...

I bet they also spent crazy money on non-diversity related things like employee perks or lobbying.
It's interesting with the Outreach programme in GNOME (I think that's what it's called), because if you periodically look at planet.gnome.org, there are interesting things going on with developers within that outreach programme.

But there are also justified backlashes to the programme, given that there is a perceived priority given to the programme in some areas instead of writing software. The argument is that not everyone and their dog needs to be involved with writing software, so why should we encourage them to? You don't see such pushes in dentistry, the car industry or anything like that; "Are you a WOMAN? Then join the car industry!".

Strangely we do in IT though, where it is the belief that we should make EVERYONE code!

The "I will fight you and I will win" response from Emmanuelle Bassi is a particularly horrible/strange/passionate reaction from one of the guys involved with the programme: see http://blogs.gnome.org/tvb/2014/09/12/im-looking-at-you/comm...

> You don't see such pushes in dentistry, the car industry or anything like that; "Are you a WOMAN? Then join the car industry!".

Yes you do. Every single fucking time this comes up someone says "you don't see this in construction". It's rebutted every time it comes up and it's really fucking easy to do a simple web search to find examples of programmes to get men into teaching or nursing or to get women or minorities into construction.

Example links have been posted to HN many times.

I wiped your profanity and spittle off my screen.

I haven't seen anything like this here in the UK. Is it a US thing?

BTW, I wasn't advocating for or against in this GNOME argument. I am entirely an observer and was highlighting the arguments made by both.

Replying to myself to mention that I am neither for or against the GNOME outreach programme!! - I am entirely an observer! I haven't used GNOME since GNOME2 due to disliking the new interface; I periodically check on there to see what's going on GNOME world and it is quite interesting, both for developments and also for what's going on internally with disagreements and discussions between people.

I thought I ought to add that because I suspect people are thinking that I dislike the programme? Either that or people like down voting with no reply.

With regard to the encouragement for everyone and their dog to take up coding, I see it a lot but in truth I do not see the same things in other professions - I have never seen a push to make youngsters take an interest in banking or journalism yet over here in the UK there is a push to make programming/coding a part of the national curriculum for youngsters, hence the introduction of the Raspberry Pi to encourage that.

Didn't GNOME almost go bust last year and hold an emergency donation drive, due entirely to granting too many paid internships (or similar) to people who aren't established GNOME developers, or even programmers?
No. The GNOME foundation is managing the funds of the outreach program (which doesn't have anything to do with GNOME, really, except it started there), so when some of the sponsors (think big corps) of the outreach programme didn't pay their agreed upon share on time, the foundation ended up in trouble. It was simply a liquidity problem which was solved when they received the sponsorship money.

That's at least how I think it went, you can go look it up, the details are online.

> There are many free software projects that are decently or well funded.

Which ones are those?

Debian, Ubuntu, OpenSSH, Firefox, Apache, etc.

In addition, some projects that are not well-funded as a project are "funded" in the sense that companies pay people to work on them, for example Microsoft paying Simon Peyton Jones to work on GHC.

Firefox is kinda special because they get a lot money from whoever pays them to be the default search engine, doesn't matter if it's Google or Microsoft or Yahoo. There's only so much software that can get away with that, GPG certainly can't.

All of the Debian developers are volunteers, am I wrong? Slackware can barely support one employee, the founder. OpenSSH falls under OpenBSD, which also supports just the founder, everyone else volunteers, and they DO volunteer some serious time and do important things. They also had problems raising funds, there were discussions on HN about that here, and I'm sure there will be more in a year or two.

From: http://www.openbsdfoundation.org/campaign2014.html : * If $10 were given for every installation of OpenBSD in the last year from the master site (ignoring the mirrors) we would be at our goal. * If $2 were given for every download of the OpenSSH source code in the last year from the master site (ignoring the mirrors) we would be at our goal. * If a penny was donated for every pf or OpenSSH installed with a mainstream operating system or phone in the last year we would be at our goal.

This is kinda depressing.

Is there someone from the Debian project here? I'm wondering if they could afford to run their own mirrors around the world if they had to. Could they cover hardware, colo and bandwidth costs, if they had to? I'm just curious.

> There's only so much software that can get away with that, GPG certainly can't.

That's the point -- the ones that can should support the ones that can't, which are often foundational components of the reason the ones that can, can get money in the first place.

Perhaps GPG should annually auction off which nation-state security service or online advertising company's public key gets automatically added to the recipients list for each encrypted message? </cynical>
It's a complex topic, so I think it's best to look at different funding models differently.

Ubuntu is funded by Canonical. http://en.wikipedia.org/wiki/Mark_Shuttleworth in other words. So one big donor.

Debian... seems to mostly get by on volunteer labor and be ok with it, or it did when I was involved with them. Has that changed? I suspect that in some ways Debian is underfunded given the amount of work they do. Perhaps money would make some things happen faster there.

The Apache Software Foundation does a decent job at fundraising, and even employs a few people to do stuff like administration. Most of the projects get by with companies that pay people to work on them, which seems to work out pretty well.

Firefox/Mozilla make most of their money with deals: Google and Yahoo, last I remember.

I don't know about OpenSSH.

It's possible, but mostly donations don't seem to work unless it's big chunks of money from companies.

    > The Apache Software Foundation does a decent job at
    > fundraising, and even employs a few people to do stuff
    > like administration. Most of the projects get by with
    > companies that pay people to work on them, which seems
    > to work out pretty well.
The ASF budget is roughly 1.2 million a year, the bulk of which goes to maintaining infrastructure for 200+ projects.

https://www.apache.org/foundation/records/minutes/2014/board...

As a 501c(3) non-profit, it is constrained as to what it can do with donations.

AFAIK Ubuntu is profitable for Canonical, so I wouldn't call that a donation.
> AFAIK Ubuntu is profitable for Canonical, so I wouldn't call that a donation.

Unfortunately, that is very far from true.

Ubuntu was negative $21 million USD in 2013[1]. Canonical would literally make money by just not doing Ubuntu anymore.[2]

(Every time Canonical is on the verge of bankruptcy, Shuttleworth re-seeds back into the company from his personal checkbook)

[1] http://www.scribd.com/doc/199373896/Canonical-Group-Limited-...

(the numbers are represented in thousands, so 21,343 is 21,343,000)

[2] http://www.wired.co.uk/news/archive/2013-08/13/mark-shuttlew...

The significant losses due to Ubuntu development and related expenses are why Canonical as-of-late has been turning focus away from Ubuntu towards other markets such as Mobile and especially Enterprise (a la Red Hat's turf).

I've never heard one way or the other whether it is or not; I'd be curious to see some data.
Name a few. I know very very few and the funding goals are very very modest most of the time (they don't really convey the idea of full time devs)
https://www.fsf.org/about/financial

The FSF had revenue of 1.25 million in 2013. I'm not trying to comment on where it came from or where it went to. I'm only pointing out that they are not in a very similar position.

Those are some rather interesting documents, thanks!

In 2013 FSF paid $ 689,239 in salaries and, astoundingly!, $ 48,995 in credit card fees.

FIY: That's 3.8% assuming everybody donated by credit card.

Anybody got an idea why they pay so much?

Typical credit card fees are 2.9% + 30c. Assuming that they have regular fees at not non-profit rates(which tend to be lower), it would make their average donation amount to be around $7.14

Source: 48995/689239 = (x*.029+.30)/x

With that many transactions they should be able to negotiate a lower fee than that. My company did a bit over 4 million in CC transactions last year and our rate is 1.9% and I believe the flat rate per transaction is 25 cents.
That's in the ballpark for monthly membership fees. ~$10/month.
As someone else pointed out 2.75%-2.9% is common, often thre is a charge per transaction too (on the order of 25cents after it's all said and done). The fee can change based on the card type (the merchant pays a higher fee on rewards cards normally...someone has to pay for rewards!) and international purchases can have additional fees. Charges backs can also bump up fees, ditto for outsourced fraud protection.
2.9% is fairly standard with Stripe et. al., but fees from things like chargebacks could probably add up to another 1%.
Maybe it includes unfavorable (normal for credit card xactions) exchange rate conversions?
My understanding is that credit cards give you the best exchange rates available at retail.

Some banks do tack on a foreign transaction fee though, but that goes on the payer.

Aren't these generally dumped on the person paying?
This is partly why I stopped donating to the FSF. They're dumping some amount of that money into misguided PR campaigns rather than helping out the developers trying to make free software better.
1.25 million USD is really not a lot of money at all... especially given all of the projects the FSF supports under the GNU umbrella.

If they had zero expenses other than staffers, at a very modest 65,000 USD a year that would not even cover 20 people.

The EFF, FSF are the only real "good guys" out there fighting for your techie rights every day... They could really use your donations and support (even if you don't agree 100% with all of their message).

So… enough pay for about 20 developers?
In India? Do you know how expensive developers are? (Especially fully loaded, ie with all the overhead that a company has to pay.)
More like 8 or 10. There are costs beside that: hardware, hosting, office, internet, etc. etc. The math doesn't work that way ;)
> How would a free software project 'pay it forward'?

As mentioned by the grandparent comment, GPG is in use by Debian, Ubuntu and RedHat package managers. Whether or not you count those three as free software they have plenty of money to pay forward to a piece of software that underpins their entire stacks.

Unless I've missed something big in the past 2 years even the Debian Project Leader is still a volunteer[0]. If most/all money's going to operating costs it's hard to make the case they're holding out on somebody. They also have a record of treating their upstream quite well so I'd need some evidence to believe they're dropping the ball.

[0] - https://lists.debian.org/debian-vote/2013/03/msg00095.html

The point isn't easily settled, it seems:

I am having a hard time to find financial statements from Debian.

Ubuntu, or rather Canonical, being a private company, doesn't seem to release financial information. The Ubuntu main page doesn't even provide a 'donate' link anymore.

Which leaves RedHat, at last. A public company, of course[0]:

   Operating profit 2014: $ 1.3e9
   Net total income 2014: $ 178.3e6
[0] http://investors.redhat.com/financials-statements.cfm
Software in Public Interest handles debian's donations. The latest treasurer report from SPI can be found here: http://lists.spi-inc.org/pipermail/spi-general/2014-November...

For some reason SPI has not put out an annual report since 2012: http://www.spi-inc.org/corporate/annual-reports/2012.pdf

This article claims that $222k was the budget for 2012 for Debian, down at the bottom. It also claims the budget should be $19B, if paid at market rates.

http://www.pc-freak.net/blog/what-is-the-development-costs-o...

> The Ubuntu main page doesn't even provide a 'donate' link anymore

Try to download it: http://www.ubuntu.com/download/desktop/contribute/?version=1...

Here's Canonical's 2013 numbers:

http://www.scribd.com/doc/199373896/Canonical-Group-Limited-...

21 million USD negative, largely due to Ubuntu development and related expenses.

Canonical has never been profitable, which is why Shuttleworth constantly re-invests his own personal capital in the company.

You're off on your numbers. The 120k are for him plus one developer. By that estimate that's 20 people for the FSF.

However, a lot of commercial entities use pgp as core of their business: all software packaged for the linux world is signed with gpg one way or another. All commercial distributions depend on it at their very core. I'm amazed that they don't fund gpg at least partially.

I'm not surprised. For the same reason OpenSSH should be very well founded, and if I recall correctly it's not the case.

One of the points of open source is that software becomes a commodity, and that will always hurt OSS founding.

The thing is, the underfunded dependencies that are the most vital are those that are widely used by many projects. If it was normal to chuck a small percentage of available funds to dependencies, the money would start to add up.