I wonder if there is a real benefit on just expanding the userbase, if
that doesn't come with an expansion in the developer-base.

In terms of donation capital, yes. We MUST support popular operating systems to gain the popularity we need to be a successful software project *of any type*. The end-game does not matter in terms of the capital raised for development, except of course in the assurance that the software will always be free and usable by everyone.

A large userbase can only benefit the project if there's a way to
convert that userbase in some sort of cash flow to support full-time
paid development.

Agreed. That's where kick-starter and other crowd funding campaigns have been a success. We need a crowd for those to work, and that crowd is going to be composed almost entirely of users. The reason crowd funding works so well is that a small donation from a lot of people is far easier to get than large donations from a small user-base.

If somebody asks for successful free software projects, examples like
Blender or Krita come up. However, those projects are successful getting
funds for paid development, successful getting the media and the
"industry" attention, and succesful integrating to workflows that are
not necessarily based on free software.

We have to present something that works with existing industry machinery. Yes, Blender supports CUDA for rendering and CUDA is a proprietary technology owned by nVidia. Without paralel multi-threaded graphics hardware rendering our scenes, we would not have Big Buck Bunny, we would not have Sintel, we would not have Tears of Steal, and we would not have as many people interested in Blender as an industry-grade alternative for Autodesk Maya, or 3DS Max, or Cinema 4D, or Lightworks, etc. etc.

So I think it's a valid question to ask if Blender and Krita are really
good examples of successful free software projects.
They are indeed sucessful, but as "open source products", in the context
of commercial software, within the rules of the market.

If we want to expand a software project beyond the realm of the hobbyist, we must first provide competitive software. We presently can't do that without making some compromises. The industry is too advanced, and too integral to cinema, gaming, and commercial video to be ignored, or waved off as unimportant. Blender knows this, and supports Unity 3D, as well as file formats for CAD programs, and various proprietary closed source technology. This is why they are popular, and why they are getting ample funding and media attention (and throngs of extension developers) right now. They are a *perfect* example of successful open source software projects. They stay true to their own open-source roots by providing all their code, while not expecting the world to conform to their personal software creedo. That is the only thing that moves Open Source software forward, and the only companies that are making a success of it are the ones that understand that. We stick to our FOSS morals, and we invite people to join, we do not demand it.

It's a tricky subject, because having programs as technically advanced
as Blender or Krita makes free software platforms more appealing to
users, but they are not necessarily successful bringing more people to
free software operating systems.

Agreed, but their popularity and funding is a product of their willingness to compromise on what professionals need in terms of integration with existing systems, and new technology that is developed by companies whose bread and butter relies on keeping their drivers proprietary (for the time being). It is not reasonable to expect nVidia to open source their drivers just because we say all software should be open source. Not when we are a trivial minority. It is thus presently good-enough that nVidia is even willing to write drivers for Linux. When we buy nVidia, we are supporting Linux, because it is showing these large companies that Linux users mater as customers.

For that reason, I don't think that having a large userbase should be
among the goals of any free software project. The goal should be
achieving technical excellence and giving users powerful tools alone. If
you have that and the program happens to be both free of charge as free
as in speech, putting no restrictions about how you can use and modify
it, it won't be difficult to attract a large userbase.

I think that's a bit like saying: "We want to open a coffee shop, but our goal is not to have lots of customers, or make money."

It sounds nice, but it's self-delusion at best. Of course we want lots of customers, and of course we want money. It does not have to be a stated goal, but it is a goal of any business, just as a large user base is a goal of any software project (especially those that want things like crowd-funding).

To achieve the goal of technical excellence, power and flexibility,
projects need the right people contributing. Good developers, good
interaction architects AND good users.

Define "good user".

Not just lots of users. Advanced users, people who really know what
they're doing and what they want to achieve when they use the tools.

So, professional graphics people... fine. But we have to cater to what they need for work. In the case of Blender 3D that means supporting closed-source drivers, and file formats that existing proprietary software uses.

We could immediately up the excellence and quality of Inkscape, as well as professional user base, by being able to read and write the file formats that work with industry-leading tools like Illustrator. That's not necessarily in-line with the end-game of having only FOSS, but it's a necessary first step to gain the professional user base we are after.

Now, let's go back to Blender and Krita and let's see what really made a
difference, apart from their apparent success getting funds and
attention: They listened to the right people, and made decisions that
made the software better.

Yep, and made compromises. Lots of them, in-fact.

David Revoy switched from GIMP to Krita, and when he did he found a
welcoming project that paid attention to his needs as an artist.

Like supporting .psd Photoshop format to be able to work with other industry artists without requiring them to learn a new software package.

They are adding layer-effect support too this time around I believe.

He has both the skills and the experience to be considered a valuable
asset in a project, even if he doesn't write a single line of code. A
person like him gives a project credibility, since other artists who
respect him as an artist will pay attention to the software.

Exactly. If Krita didn't welcome him instead of demanding that he toss out all his Photoshop projects and start over using open formats, would he would have still been interested? Probably not. Why? Well it's not very welcoming to evangelize new users, and demand they change, and go FOSS cold-turkey.

That's the userbase a project has to grow.
The guys from Blender did the same, and you can tell by the impressive
quality of the art created with Blender lately that serious artists have
chosen it.

And again, it's because Blender made necessary compromises. If Cycles were run on just FOSS software, most of that art would still be rendering on CPUs so hot you could cook your dinner on them.

So, is it really necessary to grow a huge user base for that? I don't
think so.

Yes, it is. Blender had a fairly substantial user base before professional artists got wind that it was a decent replacement for Maya and 3DS Max.

14 years ago, I remember playing with it along side Maya (when it was still Alias Wavefront). It's reached a maturity lately that can compete with the big-boys, and in a lot of ways surpass them. If they didn't have the initial user base, it would never have gotten this big. If they didn't make compromises to get big, they would still be hobbyist. It takes small steps to climb a mountain, and it's never a straight path.

Why not hunting the people doing awesome art with Inkscape and ask them
about their needs to take decisions about the direction of the project
instead?
Defining an audience, trying to create the best tool for that audience.

It's a good idea.

I would be surprised if Inkscape devs have not already been doing this. They are actually doing a fantastic job as well.

I do luxury branding with Inkscape. I would be happy to help with direction for professional graphic designers. To be honest though, they surprise me with all the awesome things they add to the project with each release.

The main things that are lacking in terms of vector tools (from my perspective) are:

1. Proper real-time perspective transformation (like GIMP's perspective transform tool)

2. A better method for transforming a shape into another shape (like deforming a title like "FOOTBALL CLUB" into a football shape).

3. Output to CMYK FOGRA39 (or even 27) for print.

4. Export to JPEG, TIFF, and other formats (currently just PNG).

5. Multi-page support and output to multi-page PDFs.

6. Import and export to the latest Illustrator formats (at least CS3) for use with printing companies.

7. Layer preservation in PDFs

8. Proper export for blur to a raster version for PDFs (if a vector solution is not possible)

9. Mesh gradients

There are workarounds for some of these, but they are quite ugly and time consuming.

That said, there are many many things in Inkscape that actually work much better than Illustrator. For example, I was unable to produce full-scale convention booth wall graphics in Illustrator because it limits the scale of what you are making (which is ridiculous for a professional vector program to do). Also, you could only output raster samples at a minimum of 72 dpi, and for an entire booth wall that measures meters in length and height, that is still an insanely large file. I have designed three convention booths with Inkscape to date (working on the nest one now), and it's easy and simple to do so because they DO actually listen to their users. I can even output a 20dpi version of the whole booth to wrap around a blender 3d model of the actual booth, and my life is way way simpler as a graphic designer. If I had to use Adobe software, life would be much harder, and I say that as a 14+ year user of Adobe software products.

If the tool is compelling enough for serious artists, they might
consider switching operating systems if the software is awesome and they
can't get it in their OS.

Not unless they have viable pre-installed Linux machines to buy along side the alternatives, or are hard-core at experimenting with other OSs (like me), there is unfortunately no way most professionals are going to be able to switch OS just to use free software. It's neither worth the time nor the trouble, and unless we have industry integration built-in, they are not going to even consider it a replacement for what they have to work with for their daily bread. Their primary objective is to use what works, what makes them money in the industry. Switching what they know to venture into the unknown, and possibly break their working computer (and source of income) to do so is just not going to happen, and unfortunately, that's the reality of the current Linux desktop/laptop situation. It is getting better though, so it's not to say that the play would not work in the future, but most of us are just trying to get by financially, in the first place at the moment. I'm fortunate enough to have a job that pays me well enough to be able to lend a hand with Open Source software projects, and write huge dissertations like this reply to try and communicate what people in my profession are looking for. The vast majority are not in my position, and I've spoken to a lot of designers from all walks of life. Most of them are told from the beginning (just as I was all through university) that Macs are the best graphic design tools. Yes, the most expensive computers out there are the "best". That's what we are up against. We can't win unless we make compromises and expand user base first, and Linux needs to come on retail computers first for the vast majority to switch.

Seems unlikely? Think again. That's exactly what people do when they
move from Windows to Apple. They make the switch for a number of
reasons, but one is definitely that they can get something they can't
get in the other platform.

It ain't because it's free, that's for damn-sure. ;)

And how popular would Macs be as an alternative if the user had to install the OS themselves? And the sound card doesn't work? And the wireless is broken? The video card makes the computer crash on resume from suspend? They would move to Windows faster than if we filled a kiddy pool in the front yard with water, locked all the doors, and set them on fire...

So, imho, there are a number of things that come before growing a
massive userbase. And if we are going to use loose terms as
"popularity", I'd say "credibility" should come first.

"credibility" is just as nebulous a term. At least "popularity" is quantifiable by number of users = user base.

Inkscape has a large and growing user base. That's based on an active community of people posting tutorials on how to use it to make stuff. Professionals don't need coddling. They (including myself) watch tutorials from other people (whether they are professionals of not), and deduce for ourselves whether or not the tool will replace the expensive ones we've been using for ages. Designers are not usually celebrities either. If you ask any given designer to name famous designers, they will probably give you names like "Jony Ives" (who is a fake designer in the first place). I'd be surprised if they could name 3 in total.

Now, that's not to say there isn't value in producing professional materials ourselves that cater to the professional market, and show them what can be done with our tools. There are plenty of excellent illustrators and designers that currently use Inkscape on their respective platforms to make jaw-dropingly stunning masterpieces that you have to squint at to see if it's really vector or not. (follow the Inkscape G+ feed to see some really wicked stuff). We could put together Inkscape showreels of this and post it on social media platforms and Youtube. That will get people of all types interested in Inkscape. We don't have to lose our current user base to expand further into the professional market, we just need to have more exposure, and know where we must go feature-wise to be able to compete. I can help that and with showreels as well, so if anyone is interested in doing that, let me know.

Get credibility from top notch users giving them the right tools to
satisfy their creative needs, and the rest would come effortlessly.

Yes and no... we have to increase exposure, and it will mean making compromises like sticking with coding the Windows version and not requiring users to switch their entire OS just to use Inkscape.

-C