[FFmpeg-devel] Sponsoring and generating money in general (IRC meeting follow-up)

[Robert Krüger]

On Sun, Jan 12, 2014 at 8:49 PM, Andrey Utkin
<andrey.krieger.utkin at gmail.com> wrote:
> 2014/1/12 Robert Krüger <krueger at lesspain.de>:
>> Does "the project" want to generate more money?
>>
>> It appears at least some people keep bringing up the topic and there
>> seem to be at least some ideas on how to spend it for the benefit of
>> the project.
>>
>> If the project generates more money, where will it go? (very
>> legitimate question by j-b)
>>
>> Apart from the small things like some hardware, travel expenses or
>> stickers the most obvious thing to me is to sponsor developers for
>> work on ffmpeg so they don't have do all of it in their spare time. If
>> this cannot be done, because of the legal setup, this whole thing hits
>> a wall (currently it is unclear whether SPI can be used for that, as
>> far as I understood Diego this is what ffmtech does today but due to
>> the beef between the two projects this does not seem to be an option
>> atm).
>
>
> Due to uncertainity with spending foundation money to individuals,
> let's better not to try to do that, and let's separate the targets of
> funding ffmpeg and funding ffdevs.
> Regarding spending ffmpeg foundation money. We can spend them for
> advertising, to bring some shine to ffmpeg product name and to ensure
> people that it is the most robust software in its category. This would
> also result in other companies to proudly declare that they use
> ffmpeg. Also earlier i have heard many times that "ffmpeg is eternal
> alpha-version" (meaning low quality and stability); it is not true
> anymore, but who knows, maybe some people think so until now.
> Also spending them for "FFmpeg Summer of Code" is fine idea (forgot
> who suggested it).
> It is not a probem at all to spend money! Even printing a batch of
> T-shirts or coffee cups for ffdevs would be fine and useful.
>
>
>> What ways are there to generate money?
>>
>> 1) Offer development projects as crowd-funding projects
>
>
> If somebody really wants something, they would come to ffmpeg-devel,
> or some developer mentioned on site page, or to freelance portal, or
> just order their hired developer to do it, NOW.
> If they have realized what they want, they aren't going to wait for
> completion of crowd-funding campaign and actual development time. It's
> too long to wait.
> I don't believe in funding one new ffmpeg feature by many backers. Too
> low probablity that several companies realize they need the same
> feature at same time.

I disagree passionately here. But of course I cannot prove this. I run
a small start-up company and in my case every once in a while there
are things that I would like to have improved and we do sponsor some
individual development work already but there are things where I know
we cannot (yet) afford an appropriate bounty and then I don't ask
because I am afraid it would be considered an insult to offer someone
highly qualified 1000 dollars for potentially weeks of work. On the
other hand there are others on the ML expressing interest in the same
thing and in most cases nothing happens. I just think that we are not
alone with that problem. As a concrete example, take a task like
"bring j2kenc in a state where you can encode files compatible with
DCP with good performance (tbd)". I don't know how much work this
would be for someone like Michael but if he advertised this for, say
15000 Euros as a crowd funding project, I am quite confident that
there are quite a few companies out there who would like to use that
in their products and contribute money, but again and maybe no single
one of them would contact him directly about it. Of course this
assumption can be completely wrong. My impression from talking to
other people from companies using ffmpeg is that it is not.

> There is generally low demand for commercial support or enhancement
> for ffmpeg, at last in ffmpeg-devel, as far as i can remember by last
> three years.
> Developers who use ffmpeg in their job get most of their questions
> answered for free in maillists (or StackOverflow etc.). Some people
> request for paid support and get contacted personally by devs reading
> maillist. Everybody is happy.
>
>
>> - Choice of the feature that is implemented because one has to know or
>> be confident there is really enough commercial or private interest in
>> that. Since I believe many companies who work with ffmpeg (either
>> command line or API) read the mailing lists, it would not hurt to post
>> ideas there and see if there is informal feedback by people/companies
>> who would give money for a given cause or help spread the word to
>> lobby for it.
>>
>> - Have a well-defined goal. "Improving filter/codec/command line tool
>> X" is certainly not enough. If someone inside a larger company needs
>> to convince their boss to pledge 1000$ or more for such a project, the
>> boss will most likely ask if the project would make feature X in their
>> software work or not or what the concrete improvement will be
>
>
> Bosses pay only if it gives gain or avoids pain. Both are actual in
> limited time, crowdfunding campaigns take long to finish.

We have the current example where four companies were willing to wait
for almost a year to make yadif relicensing happening. That is an
extreme case but, for example, product release planning is in many
cases long-term and if a crowd funding campaign runs for something
like 6 weeks, this can be easily accommodated in many companies'
release cycles provided the feature is important enough. This, of
course, does not work for desperately needed bug fixes.

>
>
>> With the current legal setup the individual developer(s) who implement
>> the offered feature/improvement for a given price that has to be
>> matched by the pledges of project backers, would be the contractor(s)
>> with the crowd-funding platform and no money would go to ffmpeg itself
>> (unless one would make it a condition that devs doing this would have
>> to donate X% of money generated through this to ffmpeg via SPI if the
>> project was advertised through official project channels like ML or
>> website or something like that).
>
>
> Too much hassle IMO.
>
>
>> So, the next step would be to discuss ideas together with people who
>> offer the actual work on the ML.
>
>
> People who offer the work = people requesting for paid support, right?
> There are so little of them. And most often their requests are not
> quite big. I believe that when people come to ffmpeg-devel for
> commercial support, most of them get contacted by somebody and get to
> their result.
> So i see no such people available for such discussions.
>
>
>> Btw. I forgot to credit Diego for making me aware of the Bountysource
>> platform which at first glance looks better suited than Kickstarter
>> and the like.
>
>
> I like also Gittip. It seems good for developers contributing a lot to
> important projects. Top people make up to $500 per week now.

Will look at it.

>
>
>> 2) Offer a sponsoring program, something along the lines of companies
>> pay a certain annual amount to reach a certain sponsor status and are
>> listed on a page on the web site. I like the idea of treating small
>> companies differently, i.e. to reach bronze, silver, gold status a
>> company of 5 people has to pay a lot less than a company of 10000
>> people (that was the linuxfoundation example Michael gave but hey, I
>> am biased here as I run a small company).
>
>
> Good to FFmpeg. If we find such sponsors.
>
>
>> This is a no-brainer as the only thing needed AFAICS is defining the
>> terms and see if there are interested companies out there. I would
>> volunteer to make a first proposal how that could look and later do
>> some lobbying with companies I am/have been in touch with. Yes,
>> Kieran, there are probably many companies out there where bold
>> assholes work that sell standard ffmpeg features as their super-secret
>> invention and those you won't get. But there are countless others and
>> the few I have talked to have expressed the wish to be seen out there
>> as fair players and supporters.
>
>
> As i have told above, let's make the fact of using FFmpeg marketable
> (through ads), like how it's common to declare that product uses Linux
> inside, because Linux is cool.
>
>
>> 3) More ambitious approaches
>>
>> Stuff like offering a support program or similar things or generally
>> really building a business on top of ffmpeg (like the JBoss example I
>> mentioned in the meeting) is most likely possible but that requires so
>> much work and dedication and is probably almost a full-time job for
>> quite some time to set up and I don't see anyone doing that (I
>> certainly cannot), so I will not make any suggestions regarding that
>> for now.
>
>
> Good idea. Somebody already running legal business could try to open
> new department for this, i believe it would be no problem to hire a
> couple of support engineers of different qualification for affordable
> salaries. The one with not-so-high qualification can provide fast
> reaction, and elevate issue if they cannot solve it by themself.
> In IRC i have mentioned such companies around GStreamer project. I
> don't know how profitable they are, but popularity of GStreamer for
> consumer apps makes me believe they are indeed profitable.
>
>
> --
> Andrey Utkin
> _______________________________________________
> ffmpeg-devel mailing list
> ffmpeg-devel at ffmpeg.org
> http://ffmpeg.org/mailman/listinfo/ffmpeg-devel