To add to Ben’s list:
Jim Munroe has had four novels published (the first by HarperCollins, the rest self - published). He wrote Everybody Dies (3rd place – IF Comp '08) and Roofed in Inform 7.
Uh, no of course not. Being attractive to authors of static fiction doesn’t mean attracting authors to write static fiction!
As to web stats, I can say that in the last three weeks Parchment had just over 2000 visitors, though there may be some returning visitors counted in that. When a popular blog links to Parchment then several thousand easily will use it in a day.
My two cents…
I decided to use Inform because of z-machine support, and web browser playability. My programming experience is mostly for embedded processors and very application specific needs (and languages) so I have always done most of my work in assembler or in proprietary languages designed for specific industries or individual processors. In fact it was a z-machine interpreter (z3 only) for a chip called the Propeller that led me here. I loved Zork and all of its offspring. And the people I see as my audience (friends) are more likely to try it if it can be run in a web browser.
So blame Parallax (and its creation the Propeller) if you get tired of my posts. 
Right. I just more meant that if I wanted to write “static fiction” – I notice people here use that term quite a bit – I would spend my time doing that. If I wanted to write “static fiction” and that was my primary goal, I wouldn’t be doing it via tools like Inform 7 or TADS 3. If I was someone who wanted to write a game that had a story and I didn’t want it to be a game with lots of graphics, then I’d probably use something like Inform 7 or TADS 3. I realize the the two aren’t mutually exclusive but people were saying Inform 7 was meant to appeal to writers and I was just indicating that only would work if those writers didn’t want to just write “static fiction.”
So if I want to write a book or short story, then even if Inform 7 was designed to “appeal to writers,” I don’t think it would appeal to me because, after all, I want to write a book or a short story. Of course if I was an author that was interested in trying out interactive fiction, then the situation is different. The assumption seems to be that Inform 7’s natural language approach is what makes it something that would target or at least appeal to writers. I would be more curious if interactive fiction as a whole appealed to writers of “static fiction” rather than worrying about whether Inform 7 or TADS 3 (or whatever else) was designed to appeal to them.
It was interesting to hear how many writers have been interested in interactive fiction, though. And it does bring into question if some of the writers mentioned above were or are part of the development team, such as subject matter experts or the equivalent of business analysts. If I’m making a tool to appeal to writers, I’d want as much design input from them as possible.
What does this mean? I had to look Parchment so I see it’s a web-based interpreter for z-machine games. Do these 2000 visitors actually play a game all the way through? Do they just try things for a bit and then move on? (I’m thinking of the old “hits” vs. “page views” distinction.) 2000 visitors could be fantastic but without a usage pattern I have no idea what to make of it.
No one has been suggesting that! When people say I7 is supposed to be attractive to writers they mean that it’s supposed to be fairly easily approachable for writers who don’t have programming experience and would like to try writing interactive fiction. The jury is still out on whether or not I7’s natural language subset is really easier to learn than a traditional language like TADS3.
Well it supports Glulx now too. Hits is the same as page views, and the visitors visit an average of 2.25 pages/visit currently. There currently isn’t any way to tell how long they play for, but for comparing web terps to traditional desktop terps that doesn’t matter because you can always stop playing in a desktop terp before the story is finished too!
And some of us just visited the Parchment site to download the tools so we can host it ourselves for our own creations. I am not sure why this is even in question. Parchment is a fantastic (and growing) tool for all of us to use if we choose to… or not if we choose not to. It opens up other venues for our play/work to be seen in. Remember, it isn’t how many play on Parchment’s servers… it is how many of us can use it for our own sites too.
But is this because they’re Windows-only games, or because they’re done in homebrewed systems? Browsing a smattering of the reviews of these games, I got the idea that it was the latter – either they lacked the niceties that other systems give authors automatically, or they weren’t much like traditional IF at all.
For an anecdote (not the singular of “data”), I prefer playing IF in the browser if I can. I started playing a little before Parchment rolled out; I played several games like Galatea, Photopia, The Dreamhold, and Spider and Web using ZMPP, I think, before I decided I needed to download an interpreter.)
Or to get the URL we’d use if we wanted to host it on the Parchment server.
All of this is cool: but I still don’t understand why it was originally said that Inform’s implementation makes success and failure not so simple. 
Is it just because it’s not implemented that way or because it can’t be implemented that way? I gave some counter examples as to what I was talking about to show why it seems success and failure could be modeled perfectly well.
Something like the implementation of success and failure for actions seems like an important point of consideration and it does seem important to understand if this was a deliberate design choice of Inform or a limitation imposed by Inform.
To do something on an action’s success, you want an After doing something rule. Be sure to end it with continue the action; so the report rules will happen.
Is this in response to what I posted here? If so, I have absolutely no idea how that answers the question. Is this a response to the original poster? If so, why did everyone say this was trickier than it sounded?