Issue sponsorship

Several people have started to add sponsorships to various outstanding issues (see the “Popular Issues” over on the right). I just want to clarify a couple of rather important points about this system. First of all, everyone participating should be aware that it is a pure honor system - we expect and hope that you will pay up once a bug that you have sponsored is fixed for you. The money you pay will be routed (minus the usual PayPal fees) to the developer(s) who actually fixes the bug.

Secondly, and more generally, I’d like people to understand the scale of things. Right now, 6 people are sponsoring “MIDI editing/sequencing functionality in Ardour”. The current total is $220. The current implementation of MIDI editing and sequencing in Ardour 3.0 was basically sponsored by Google’s Summer of Code over 2 summers. That means that Google has already put more than US$10000 into this functionality. Dave Robillard, the primary developer working on MIDI stuff, has continued working on this and related matters between the two summers that Google was paying him.The work that remains, while probably not on the same scale as the work Dave has already put in, is still substantial. I don’t want to disappoint people, and I don’t want to sound ungrateful or dismissive of the sponsorship associated with this issue. But people need to realize that amounts on the order of a few hundred dollars in connection with such broad issues just isn’t realistic. Now fortunately, Dave and the rest of us seem fairly motivated to get this task done regardless of the specific financial renumeration involved. But people need to be clear about realistic levels of financial support. We are talking about at least a couple of man-months of work, and possibly a lot more than that, and for sponsorship to accelerate that it would need to be much more substantial.

This applies to smaller bugs too. Most bugs in the Ardour bug tracker require a minimum of two hours of work. Often the majority of that time is spent investigating and debugging the problem - the actual fix is often much quicker. We do have some developers who work in parts of the world where US$50 for at least a couple of hours work is an attractive offer. But in general, people need to understand that, as with MIDI, the scale of financial incentive necessary to get people to specifically work on a bug is likely to be larger than most issues are currently listed with.

Please continue sponsoring bugs and features. Its very welcome and sometimes motivates developers when they might otherwise watch a movie. But also please understand the scale of the work involved, and the way incentives are likely to work.

Hello Paul,
I’m relatively new to this forum, but i have been playing arround with Ardour for quit a while.
First of all, the moment i started playing with Ardour and Jack some 3 years ago, it blew my mind away, in a positive sense. As I was then (and still now) using Cubase+Reason on windows,but i 'm amazed by the stability and flexibility Ardour and Jack offers. Thumbs up!
In the following I would like to point out some problems that I see, which may seem uncorrelated to this current issue. In my opinion (consider me a layman in audioLinux), there’s a strong connection.
The reason why I have not yet accomplished the complete switch to audio Linux is the fact, that for a decent Production one needs Ardour, Hydrogen and Rosegarden or Muse for MIDI, and depending on the style some synths (ZynAdd,FluidSynth…) and/or samplers.
Count them up and you get at least 5 progies for one project.
You know where i’m heading to, right? I’m aware of the LASH Project, also figuring out how to utilze LASH-aware apps. But the level of integration is by FAR not optimal like Cubase+Reason+VSTs from the point of handling the project. BTW, i am NOT talking about the integration level on the audio side (It is DAMN superior to Windooz and apples :slight_smile: ).
On Windooz I really just use Cubase+Reason+VSTs. That is: to work on my projects, I open Cubase then Reason and thats it.

And on I go to LADSPA. The plugins sound good, but the workflow is hazardous as one has no decent individual GUI per plugin. I know to it’s dew to the LADSPA spec.
But Individual GUIs for plugins are ESSENTIAL (What about LV2 in Ardour? Native Linux VST?) .

Now comes the bridge to this issue: On the long run Ardour could only survive, if the project had a commercial Branch (such as Trolltech with their dual license, or Ubuntu/Canonicle, OpenSUSE/Novel …).
But if we are honest, commercial Ardour alone has NO CHANCE against Cubase, Logic… mainly because of the 2 problems above, no matter how stable and low latecy it is!
It takes a commercial firm that

  1. pushes the LASH Project on carefully chosen apps, say Ardour+Hydrogen+Rosegarden (until midi is stable in Ardour) and QSampler, and only focus on these.
  2. pushes a standard for plugins and synths (Native Linux VST again?) to near-maturity.

I have tried 64Studio, but idont see the point of maximizing the amount of audio apps in a (well configured )distro, and selling it. (64Studio guys, pls dont take this as an insult!).

What the audio user wants is a product with a streamlined production philosophy,
that is:

  1. Ardour for audio
  2. Hydrogen for drumming
  3. Rosegarden with synths as plugins
    then
  4. QSampler for sampling (optional)

And THATS IT, no freewheeling, Freqtweak or other apps just built to do that tiny fine job,
as these must be done by the plugins .

I am a Software developer in the image analysis realm, so I am pretty aware of the complexity of the task. Nevertheless I do see a chance for a commercial streamlined audio distro, because musicians like me just like a small set of programms with ultimate integration, and we would run to the stores for such a distro.

This brings me to my last point:
Paul, you are the GOD of audioLinux :wink: At your word, all the programmers of those neat fine small audio apps like Freqtweak, CALF, the dssi Plugins, ZynAdd, would port their code to the plugin standard of your desire.
Wouldn’t it be viable to gather some people of your trust to focus on a distro with the 4 major apps above? You can manage it!! Why not a joint venture with canonicle to produce a commercial audio distro? After all Ubuntu is well known. I think, if you follow the points above you could get considerable positive response, especially from major plugin vendors (like NI…).

As a plugin standard I therefore suggest VST, since it would make life for major plugin vendors easier, when porting their code to linux.

But heh, what do I know? I’m a layman in audioLinux :slight_smile: !

Best regards

thecrazyguitargod

EDIT: Sorry, I forgot LV2 is on the way in Ardour. Personally I think Native Linux VST would also be a good feature for the above reason.

“Popular Issues” is on the right for me…

Perhaps there could be a time estimate on bugs, saying how many hours it probably will take to fix?

@djuato: the difficulty with that is that, as noted, it often takes longer to decide how long the bug might take to fix than to actually fix it. Identifying the likely “depth” of the problem and thinking about ways to re-implement the broken design is often twice as much work as actually doing the re-implementation.

I knew, when starting the sponsorship about the “Midi editing and sequencer functionality” my 50$ would never ever cover the efforts which are behind that feature.

My original idea behind that was more to motivate other people which wish/want/need that feature in ardour to sponsor something for it.

For the last two years watching the ardour development process one of the most often mentioned “missing features” has been MIDI editing/sequencing. Therefore the idea was, if all of those people mentioning that features as missing would also sponsor a related amount of money, the sum of all sponsorships can reach a sum which could be a real motivation to finish that feature.

Kind regards,
Harald

A couple of things: first off all, I’m more than willing to pay a fairly reasonable amount of money to sponsor the MIDI editing feature (between €50 and €100). I consider this upgrade money, since I’m already a paying subscriber. However, if only 6 people are sponsoring an issue, you will never receive enough to motivate the coders in question. You’re obviously trying to make a living out of an open source project (and apparently not just you, but Dave as well), so it seems obvious to me that you should be aiming for a larger paying user base instead of hoping for larger payments from a subset.

Maybe you should consider effectively selling versions of Ardour for a certain amount of money (in the EnergyXT range perhaps)? That will also require a more commercial approach to things, not so “grass roots” as it is now, but I’m willing to bet that some form of advertisement will greatly improve Ardour’s image within the recording community. As a matter of fact, I know certain people (who buy Logic, Pro Tools and what not) who like Ardour but aren’t willing to use it (not even for free) because of it’s lack of standing and usage in the recording community - silly, I know, but it’s an untapped market for you. From where I’m sitting, it looks like Ardour 3.0 will have all the features necessary to compete with the aforementioned DAW’s, so maybe “professionalizing” your business might not be a bad idea. Instead of looking at your work as a service, you should look at the result of your work as a product.

As it stands now, Ardour is probably one the best sponsored smaller open source projects. Hope of improving this seems rather futile and based on dreams more than anything else. Not to discreted your current Modus Operandus of course ;). Furthermore, I’m a big supporter of open source, so I’m not preaching to have svn access shut down or anything like that. But binaries can definitely be sold without too much hassle. It’s not like one can’t pirate Logic or Cubase with more ease than checking out an SVN copy of Ardour and building the damn thing yourself…

One final thing: I have not commited to sponsoring the MIDI bug simply because my subscription has not been linked to my drupal account yet. I have sent you two mails about this, but you’re probably busy doing other things. As soon as this is fixed, I’ll commit myself to the bug as well. Oh, maybe a release date, milestone plan or anything like that would be nice though, but having been a developer myself, I know how annoying a request like this is ;).

All fine points, but somewhat in the opposite direction than the ones I was trying to make.

As far as making a living from “selling” Ardour, that’s underway already. Perhaps you’re not aware of this because you’re using it on Linux, where its much more difficult to generate a single binary image for all distributions. But look at http://ardour.org/download to see the “pay tunnel” that already exists, and is generating more than half of the income associated with Ardour.

Bug sponsorship is not likely to generate much revenue for any particular bug, ever. Its a way for someone who really, really wants a particular issue implemented or fixed to provide some incentive. My point was that even with a small bug, people need to understand the real work involved - in general most people won’t want to pay for it all themselves. When you move on to overly-broad tasks like “MIDI editing and sequencing”, the whole idea breaks down in my eyes.

Yeah, I actually did see that page. Tried to download for MacOS X (the reason I’m hoping you’ll link my subscription with my account here). But you’re right, I mainly use it on Linux. Anyway, like I said, I didn’t mean to critique the way you’re currently handling things, so I’ll leave it at this.

Back when the idea of sponsored features & bug fixes was first mooted I argued very strongly that feature requests should be in their own separate database, rather than being lumped into Mantis. These were the main reasons:-

  1. Users, rightly or wrongly, percieve Mantis as a bug tracking system. One of the reasons that feature sponsorship is so slow is that users probably aren’t even looking in Mantis to keep themselves up to date (see also, point 4).

  2. Having feature requests and bugfix requests in the same database was bound to be detrimental to the feature requests. This is because psychologically, users - however unwittingly - would perceive a “going rate” for the work. I stated at the time that this would most probably end up in the low hundreds of dollars (which indeed it has). A few hundred dollars might be an acceptable reward for fixing a bug but it’s never going to compensate anyone for developing a new feature from scratch.

  3. The honor system is fine for sponsoring bug fixes but it’s wholly inappropriate for feature development. Here’s why… let’s suppose that 6 users sponsor $50 each for a bug to be fixed. I spend 2 hours fixing the bug and check in the fix. But 2 of the sponsors renege on their promise and don’t pay up. Okay, that’s disappointing but I’ve only lost $100 and I’ve only given up about 40 minutes of my time for free. Contrast that with feature development… 300 users sponsor $50 each for a brand new feature. This takes me 6 months to write and I check it in. One third of the sponsors again renege on their promise. That leaves me $5,000 out of pocket and having done two whole months’ worth of work for no reward.

  4. Mantis has no mechanism whereby ‘unsubscribed’ users can be periodically reminded about feature popularity.

  5. Mantis has no easy mechanism whereby developers can propose ‘quotations’ for the work.

Mantis is not a suitable vehicle for promoting sponsored features and never has been. Sponsored features are likely to cost in the low tens of thousands of dollars - not in the low hundreds of dollars like bug fixes. A separate database is needed. It should allow devs to quote for jobs that interest them (and even to collaborate). It should also work on a ‘commitment’ basis - i.e. once the dev(s) commit to delivery within a reasonable timescale, this obligates sponsors not to renege on their sponsorship. Most importantly, it should deliver a monthly report to ALL registered users, keeping them up to date about which features are attracting sponsors.

I don’t know if such a system already exists (or if a suitably modified version of Mantis could be run separately). But for sponsored features to have any chance of being successful, those are the minimum requirements IMHO.

Although I broadly agree with you John, I would just quibble with the claim that “a few hundred dollars is never going to compensate anyone for developing a new feature from scratch”. This might apply to major features - it does not apply to minor things. I have frequently been involved in discussions on IRC where a knowledgeable user has been able to concisely and accurately describe a new editing feature, and I’ve been able to implement it within a couple of hours. I think you are mistakenly imagining that all new features are large. One good example that lies somewhere between the two right now is “cut preview”, requested in these forums within the last 5 days. I haven’t thought carefully about it yet - it could be 2 hour task or a 4 day one.

Features that take 6 months to write are not going to be handled by any “automated” system of sponsorship, IMHO. For them to be capitalized in that way will require a lot more “marketing” and so forth. This is what the main, non-sponsor income is intended to cover, since I don’t believe that enough software users really want to pay up front specifically for features that are not implemented yet. This is how proprietary software is developed too: the company receives a stream of income from sales which pays programmers to do the work and at some point in the future, a new version is released with new features. I find it increasingly likely that a lot of FOSS development that receives any funding at all will move in this direction too (as Ardour has).

Well, for my part I just wanted to support the development of midi with money. I had no delusions that it would speed the process just reward the developers for their hard work and as there is no other mechanism to do this specifically I tacked on to the idea to show my support for the project. I’ve upped my contribution to $100 and hope others will too. I understand that this isn’t what the bounty system was intended for but it’s all we got. So while it may not be a way to provide one for one financial support to the feature it does help me show my appreciation for the developers working on midi with my money.

Just to clarify, Paul - I was referring to the scale of feature development that you described in your opening header (i.e. features that will take a couple of man months or more to complete). It’ll be great if non-sponsorship income reaches the stage of being able to cover this but I think it should be possible for conventional sponsorship to play a significant role too. However, I don’t think sufficient sponsorship can be achieved using the existing bounty system (via Mantis). IMHO Mantis hasn’t got what it takes to be effective for this kind of task but for smaller features / sponsored bug fixes etc it’s fine.

don’t know, but imo this is a bug market and no more an open source project! if developer lose there motivation to spend livetime (or free time) into projects like this, it’s time to stop. if frustration is the main wage you get… don’t waste your time (live time can not be payed by mony). ardour is a great software. but, since several time I miss the innovative open source feeling here.

the same thing for all users in the feeling waste time with open source software. pay for commercial software and support…

@is_the_project_dead_ohh: perhaps thats because you are not on our IRC channels. We (the main developers) do not use web forums to develop, discuss or design Ardour. We don’t even use email for that, most of the time. Our IRC channels are active, and more days that not, feature long discussions about new features, new designs and more. If you sign up for the commit notification mailing list, you can also watch check-in’s as they happen. There are huge changes taking place right now in the 3.0 branch, but you won’t find these discussed anywhere online except IRC (and even then, a lot of that is on a private channel).

i know, at moment there is a lot of check-in’s (WORK) into the 3.0 branch. and yes that’s right, i get my ???_feeling from this project page and forum.