Reflections on PGIP phase 1

I first wrote about the Philip Greenspun illustration project (PGIP), which I was/am coordinator of, in November 2007. PGIP was made possible by a US$20k grant by Mr Greenspun to the Wikimedia Foundation, “for the purpose of creating and improving illustrations on Wikimedia.” I said:

I’m going to be co-ordinating the project, which is both exciting and scary. If it goes well, it will likely open the door to future “targeted donations” and content creation projects. If it goes well, it will get new people involved in a really global SVG editing community that is open, growing and self-supporting. If it goes well, complex and fundamental topics will gain world-class illustrations to rival any “visual dictionary” or “children’s encyclopedia” and the like. What’s more, those illustrations will be able to be translated with nothing more than a text editor. And they will be free to the world to use however they like.

There are two ways it could go poorly. One way is due to lack of interest, which would be disappointing but not disastrous. The other way is spectacular failure, where the introduction of money into a previously volunteer-only cycle reduces or ruins the motivation of those contributors.

It seems I forgot a third way it could go poorly, “disappointing but not disastrous”, which is something like setting itself up to fail by asking to reach standards that required a significantly higher amount of human effort (compared to existing FP processes). Combined with operator failure (creeping perfectionism + decreasing time availability + changing volunteer role on my part). Combined with a manifestly less-than-perfect technical solution.

The latter, an obviously less-than-optimal part of the current project, ended up being wiki pages + overly complex bug tracker software + uploading at Commons + mailing list. That undoubtedly put pretty much everyone off, but I am grateful to the illustrators that did persevere.

The resulting images are here. These 20 images are really superb. As ever I am impressed at what our community is capable of. But they are only 20. At this point, I can’t help thinking that I failed to learn a lesson from Nupedia.

As my original comment makes clear, I had some reservations about “paying volunteers”, but I wanted to take the opportunity to try something. I was familiar with some ideas about how introducing money into a volunteer system could actively kill off volunteer motivation. Naturally that was the last thing I wanted.

I think the original idea for the project was pretty simple — Problem: Wikipedia needs more illustrations. Solution: Pay people to create them. Keeping in mind the known problems, I conceived of it at a different angle; I wanted to solve the meta-problem. Meta-problem: Wikipedia needs more illustrators. Meta-solution: Create a project that helps to grow the international illustrator community, by using payment as the spark of interest. I still think that is ultimately the right answer, if Wikipedia has shown us anything at all. But it was just a tad ambitious to take on as a volunteer project myself, given that around this time I started my first full time job. This ambition and perfectionism also led to a dangerous tendency for too many “bucks” to stop with me.

If money is involved, a project must be (somewhat) professional. It must have standards. Someone must be accountable. I still think these things are true, but where I failed was in failing to hand over bundles of control to trusted people. I failed to gather those people, in part. I let too many people who might have been involved fall through the cracks by taking a REALLY. LONG. TIME. to get the project moving.

I was, it can safely be said in hindsight, overly concerned about answering all the critics before they could open their mouths. Given that it was a ‘paying volunteers’ project. Given that it was the first project like this on WMF’s behalf. Given that Wikipedia already gets plenty of stick for not meeting Standards of any particular description, and errors may already be seized as news stories whenever a writer sees fit. Given all these things, I was a bit frozen in fear of making a wrong move, delaying action by making extra plans, extra guidelines, extra rules, to the point that the criticism that this money was not being used for its purpose became valid. Ack. Didn’t see that one coming.

I wanted to use the money to pay for illustrations to be created that did not compete with existing volunteer motivation. So I thought the best way is to use it for “worthy” topics that are too boring or of little interest to naturally attract volunteer illustrators. It would be a bad idea, I believe, to simply pay illustrators for images they are making already. And as with Wikipedia, where addicted editors find themselves searching for references on topics they have no particular interest in, merely because another editor has requested it, many illustrators also take part in this kind of cooperation, as can be seen at the graphic labs and image request pages. But these open ended request processes are like bug reports: people will fill them if they want to (maybe it piques their interest, or they like you, or you give a usefully detailed request). But there are plenty of requests made this way that won’t go filled either. And I believe, although I hardly have enough data to prove it, that providing nominal payments is not enough reward either. Like a nicely worded request from an appropriately appreciative person, it is an incentive, but that’s it: a carrot, not a stick. As one paper puts it, pay enough, or don’t pay at all.

It is possible that the system I set up could have worked. If I had acted differently. If someone else had run it. If if if if. Basing a project on one person’s force of will is pretty risky, as I have manifestly demonstrated. A better system has more redundancy built in and can work even if one or two bits fail. Especially a volunteer system.

But the thing is, in thinking about a possible phase 2 PGIP, I am having trouble envisioning a working system might look like.

So I am kind of wondering if the idea of paying volunteers is just way more trouble than it’s worth. I don’t think there is no answer to that question, just that this answer is not quite the right one. And it’s good to be clear, is there money for things produced by creators (nice tangible things, like illustrations), or for a creator community (something that might look suspiciously like a community manager salary)? Which is generally less tangible and sexy, and riskier on the part of the donor. But much safer on the part of the community and, if successful, ultimately productive.

Thus endeth PGIP phase 1.

10 June, 2009 •



did you consider adding together the money and have a lottery amongst the submittors? To have a competition for the best image? That way the amount is higher, which would solve some problems, and especially with the competition adds a component of pride.

However, paying per edit seems to have worked for Betawiki..

Lodewijk · 18. June 2009, 02:20


Yes, we considered a competition. But there are two issues with that.
i)Are people competing on the same image? If so, you are wasting valuable people’s effort.
ii) Are they competing on different images? If so on what basis are they compared? And what is an acceptable submission?

This doesn’t avoid the problem of paying volunteer for what they already do, unless you draw up a list of suggestions which they choose from, in which case that doesn’t avoid the problem of needing to create that list.

2) If you have winners, then you also have losers. How does making people feel like losers help?

As for paying per edit, that is only for languages which haven’t found “natural” volunteers, right? You don’t get paid to do any big European language, IIRC. But the acceptable submissions are already defined, by the list of MediaWiki messages. What could the equivalent in illustrations be?

pfctdayelise · 18. June 2009, 18:49


You sound distraught. Why does it seem that anyone who takes up a visible, demanding, volunteer position with/for the Wikimedia projects, almost always eventually seems to regret it, worn out, disillusioned?

And, why the wringing of hands about “You must not compete with volunteer motivation. (Payment must be for things volunteers are not showing an inclination to do on their own bat.)”? I would imagine that many Wikimedia volunteers would love to be paid $75K for a speaking engagement, talking in a mixed manner about Wikipedia and Wikia. Jimbo, and Jimbo alone, gets paid that way, but nobody else seems to up and quit the project because they’re not making the money he makes from the labor and dedication of others.

I think you’ve been had. Had by an irresponsible board of trustees and a staff that deliberately tries to offload as much of the actual “work” onto the most starry-eyed of volunteers… while they, and they alone, collect their paychecks.

There’s a reason only about 34% of what the WMF takes in actually gets paid back out to program services. It’s called a “confidence game”.

Gregory Kohs · 21. June 2009, 13:44


Your concern for my welfare is so touching, Greg. I know no one has done more than you to make sure Wikimedia volunteers feel part of a safe, supportive community.

In your rush to introduce an irrelevant comment about Jimbo, you seemed to miss my comment about being aware of research that found that “introducing money into a volunteer system could actively kill off volunteer motivation”.

And you know as well as I do that it’s not the board or staff’s place to do “work” that involves editorial control.

pfctdayelise · 21. June 2009, 19:36


You have it exactly backwards, Ms. Laugher. But you’re in too deep to realize it. I am indeed concerned about your welfare, while the WMF couldn’t give bupkis about it. Ask yourself… what happened to the welfare of the following volunteers: Danny, Durin, Alex756, Essjay, Alison, FT2, Sam Blacketer, etc.? Did the safe, supportive community come to their aid?

You’re being chewed up, and you don’t even realize it. The only question is how long before you’ll be spit out ( ) by the machine that cares not one whit about your welfare.

Gregory Kohs · 21. June 2009, 22:35


I wonder if PGIP might be better suited to creating highly technical illustrations that our contributors probably couldn’t ever create on their own. For example, I need specific micrographs of the inner ears of a cat cochlea (and so on)… the chances of me ever properly dissecting the cat and then having access to a scanning electron microscope is slim to none, yet our fair use policy prevents me from lifting images from scientific journals since they would technically be “replaceable”. Here, PGIP could fund the creation of very useful illustrations without infringing on areas where volunteer contributors would otherwise fill the need – since they basically can’t. No erosion of volunteer motivation on this front.

I wonder if that would be in line with Mr. Greenspun’s vision for the project.

Mike.lifeguard · 24. June 2009, 09:07


@Mike, so if our own contributors “basically can’t” create these images, then who is creating them?

Paying outsiders to create images is certainly possible, but you still have to create the list of images you want and evaluate them. But if you are paying professional rates, it’s more just like a straightforward business transaction and you won’t get as much bang for buck. It also won’t help grow the illustrator community.

pfctdayelise · 24. June 2009, 11:16


1. what do we really need new illustrations of? what are current priorities and how can we better gather collective priorities? special:wantedimages doesn’t exist and wantedpages only gets you so far…

2. what are barriers to getting illustrations? skill among available community? time of community members? other external needs for tools and access that require time or funds?

Once people truly become wikimedians, the pleasure of having a certain audience see and use their work tends to dominate their desire to fulfill a meaningful need.

But most talented people, artists or writers, aren’t yet Wikimedians and have quite a barrier to cross before they becomeone, or feel welcome or part of the community.

Money helps overcome certain barriers (reimbursing personal costs, encouraging putting in that first effort to try something new).

I don’t agree with most of the specifics in your post here — from the sense that a monied project has different accountability and quality standards than any other [these standards should be high, regardless! imagine telling the Main Page coordinators that they shouldn’t worry so much about the quality of its daily updates and layout.] to half of your last four points — make all volunteers part of the prioritization, selection, and distribution process, and you now can’t possible compete with them; you’re just helping them find another way to realize the community priorities. combine points 2 and 4. and if the community thinks an illus is really important, and can’t find it on their own, and someone finds and uploads one that meets standards, that’s providing just as useful a service as if it were created from scratch. Do what the mathematicians do and reward the finder :)

But maybe constructive ideas are better suited for a full post of its own. I like Mike’s comment about new images that can’t be produced without certain access, prioritizing them and finding people to make them so.

SJ · 26. June 2009, 12:27



“Rewarding the finder” may well be a good idea but that was not what the money was specified for.

Maybe I didn’t explain this “competing with volunteer motivation” thing well. It’s not about people-seeking-money directly “competing” with volunteers. It’s about money-motivation entering the scene and in a single person, competing/crowding out the volunteer-motivation that they previously had. And I think that is a factor regardless of whether or not a paid event thingy is entirely operated by volunteers or not.

pfctdayelise · 30. June 2009, 00:52

Elsewhere on the web...

Commenting is closed for this article.

list of all posts, ever

find articles by tag

monthly archive

most popular articles

  1. NLA Innovative Ideas Forum audio/video now available
  2. [guest] Rethinking the Top Ten
  3. How to use Gmail to manage high-traffic mailing lists
  4. Top 10 software extensions Wikimedia Commons needs in 2008
  5. Write API enabled on Wikimedia sites!
  6. Free MediaWiki hosting offered by Dreamhost Apps
  7. An alternative term for "User-generated content"
  8. GLAM-WIKI, day one
  9. NaBloPoMo
  10. Reflections on PGIP phase 1

(from the last 30 days)