Inform 7's proof of concept "admirable game"?

In Why is Fallen London so dominant, Hanon wrote:

I wasn’t around at the time, but this is the I7 early chronology as I understand it:

The first Inform 7 game released was Emily Short’s Mystery House Possessed in March 2005, part of the Mystery House Taken Over project. I’m not positive, but this game’s banner may have been the first public indication that there even was a project called Inform 7.

This, on 2005-04-02, is likely the first public when’s the next release of Inform 7 going to be? question, predating Inform 7 itself by more than a year. (I’m not counting a prior thread by someone who noticed Mystery House Possessed’s provenance and asked whether I7 had been released.)

The next released game was the original SpeedIF Gruff version of Bronze, 2006-01-21.

The next releases were announced 2006-03-01:

Many small and medium-sized games have been written during the development of the forthcoming Inform 7 design system for IF. Some are used for testing, others were simply experiments, but most are intended as illustrative samples. More than 200 of these will appear in the documentation, but they necessarily show off one trick each: so we wanted also to offer a few larger-scale “worked examples”. Though not enormous, these are too long to be included verbatim in any book, and their full source text will instead be published on the Inform website when Inform 7 reaches its public beta.

In that thread, Graham briefly describes I7’s development. (If you read that thread, you’ll see that embittered, entitled complaints about the development of I7, like questions about when it would be out, predate I7 itself.)

At this point, it was public knowledge that I7 would be fundamentally different from I6 toward making it more approachable (and that I7 would compile to I6), but it seems like the natural language-inspired design was still under wraps.

The initial release of the Public Beta of Inform 7 on 2006-04-30 included the source of the three prior games as promised, as well as:

three new works by Emily Short: “Glass”, and the first two episodes from a five-part series called “When in Rome”.

The 2006 IFComp winner was Short’s Floatpoint, written in I7.

So I7 had kind of an embarrassment of riches from the start in the way of demonstration games.

I had thought of Bronze as the closest thing to a single admirable game for Inform 7, but my perspective is inevitably skewed by (my perception of) those games’ endurance.

For those who were there, did any one game outshine the others as being a selling point for I7? Or is the whole question just silly in the context of:

  • the system was developed by Graham Nelson in collaboration with Emily Short
  • when it released it released with a Graham Nelson game and five Emily Short games


Having written this all out, I find myself leaning toward “it’s silly”.


I’m obviously biased, but I think I7 didn’t really have, or need, a founder-effect game in the way you’re talking about.

Inform 6 games (and Infocom games, and other parser-based text adventure games more generally) were already well established. With Inform 7, we aspired to make it possible to do a wide range of experimental, non-standard things too, but a core intention was to make it easier to do what Inform 6 did well. And it felt like everyone in the community basically already knew what that thing was.

From a PR perspective at a time, our chief concern was about proving that I7 wasn’t just a syntactic sugar template that would let you do a tiny subset of I6 programming – that it wasn’t a toy, that it was robust and interesting in its own right. So we asked ourselves: what games do we need to publish (with source) in order to demonstrate that we can do all the standard things, from locks and light sources on? What games would especially showcase things you might expect to be hard with a natural language-based programming system? And then finally, how do we point at some of the aspects of the language that we’ve come to think are interesting, and encourage people to give those a go with their own ideas?

Bronze was in a lot of ways the most conventional of those pieces – the one that comes closest to just replicating what Inform 6 games already tended to do, albeit in a new language. I was at the time also extremely interested in using I7’s improved NPC action handling and rulesets for more automated NPC behaviour. Glass, the WiR games, and Mystery House Possessed were all attempts to suggest the possibilities in that direction.

(Though I am sure we talked about it at the time, I don’t now recall all the details of Graham’s intent with Reliques – though I believe at least some fraction of it was around demonstrating that I7 was flexible enough as a language to handle some fairly math-heavy, non-traditional design features.)

At the time, my impression was that all of those games were understood as example games rather than as real releases: Bronze eventually accrued a decent rating on IFDB, but its only XYZZY nomination was for Best Use of Medium, which at the time was often regarded as basically a technical achievement award; and most of the contemporary conversation was about the language, not about the companion games.

As for the examples more centered on dynamic NPCs, they didn’t really get much traction – perhaps because I hadn’t sufficiently solved the design problem of how to make that kind of experience as polished and compelling to play as people found more traditional puzzle IF, or perhaps because the trad parser IF community was the wrong group of players to get excited about the more textual rogue-lite experiences they offered, or maybe a little of each.

At any rate, it was a very different scenario from Curses + the original Inform, or any of the other examples in that thread, where the game came out and made a big impression and then the tool was also made available.


I was only glancingly engaged with IF back then – was mostly a Comp-only person in those days – so I’m not the best source. I do remember the I7 launch being a big deal, though – maybe covered in some of the mainstream gaming blogs, which is why I noticed it?

Anyway I remember Damnatio Memoriae as being the “example game” – like, that was positioned as the small one whose source code was more accessible to poke around with – and Reliques as the full-fat game (though I don’t think I played very far into it, and as I recall the immediate community response was a bit disappointed).

Somehow I don’t remember Bronze, though when I started plugging back into the community in 2020 and was considering learning I7, I similarly felt like it had become more clearly positioned as the “admirable game”.

Zooming back, though, I definitely felt like it being the next version of Inform was the major selling point; alongside the natural language approach, it felt like it pretty much sold itself.

EDIT: well, by posting immediately after Emily Short, I hope I’ve been useful by establishing a deeply-insightful/hummingbird-brained continuum of responses, here.