New Inform 7 - Very Cool, But ...

Authoring Inform 7
41

If you were a “professional software dev” you would recognize that many ideas were put forward here and thus not half-baked at all. Perhaps you think so but you don’t state why and, in that case, that would make your own criticism of the post half-baked, would it not?

All of this is true – or at least can be true. But, again, you have to read some of the specifics to see why I was arguing the way I was. Some bugs will get out yes. The question is what kind of bugs and when could they have be found. If you are a professional software developer, you know that if within moments of a release someone finds a bug that you absolutely could have found – and should have found with a regression suite of tests in the manual – then it should suggest different approaches to your releases. Any professional developer would want to at least suggest these things and try to get them implemented, even if ultimately it ends up not being addressed.

I keep hearing about how it’s a one man show and how that makes it incredibly difficult for Graham. (I don’t hear him say that. I hear others say it on his behalf.) Maybe there are ways to ease that and make quicker releases that allow for more iterative features. Does he not want that? Maybe not. But maybe it’s a discussion that could be fostered.

Feedback from large groups can be contradictory in that people want different things. That’s a given. However, just as few authors will pull down daily builds – which, by the way, no one suggested – in many cases only a segment of your community will actively engage in a user interface session. So there is a self-limiting factor that is usually built in. Granted, that may all not be the case with this community. Each is different to certain degrees. Regarding the daily builds, again, if you are going to comment at least read what you are commenting on. I said “iterative” which, if you are a professional software developer, you know that does not have to imply daily and often does not, particularly in the context of gaming. “Dev builds” (which can be daily) do not equate to “acceptance builds” or even “test builds” (which can be, and usually are, of differing time frames).

Great info. Maybe this was available or obvious somewhere. I wasn’t aware of this. All this being said, it argues even more for iterative type releases of limited feature sets. If you are a professional software developer, you know that the more audiences you listen to and try to satisfy, the harder that gets to do with much larger feature sets that make more systemic changes. It also makes it much harder to course correct if you find you have to normalize a bit between those different audiences. Again, though, great info.

Things like this, assuming your inference is accurate, would be interesting. But you are inferring one thing so I don’t know if “Glulx becoming the default due to outcry” is another inference or a fact. For all I know, Graham may have gone that route even with misgivings and even without outcry. I find the “outcry” odd because many, many threads I read on these forums – even some recently – keep showcasing people who do not want to abandon the z-machine and are worried that Inform is growing beyond it. Not saying you are wrong, just saying this is an example where there are clearly different interpretations possible.

Fair point. This is the first thing you said that I felt was in line with your opening statement of being a professional software developer. That said, commentary and discussion can be “undeniably useful” as well. But it’s a two-way street. Usually.

42

I sense there is no change in the tone at all. And I’m missing the point.

The most important things delivered in whatever market, nowadays, are big upgrades, dropped once every two years, which seem to need a patch the very first day after release. Of course I’m mentioning sub-categories of very unimportant softwares for a succint nerd group of dedicated users like iOS or Windows or any of the 100million players of certain MMORPRGs. So, no, this can’t apply to Inform, a propretary platform that some man decided to release publicly for free. All one has to do to prove it is doubting other - well established - roles in the IT business with sarcasm and ultimate usefulness.

I’m honest (guess you have got, by now, the nature of my temper - for which I humbly apologize): the thread is interesting. But I feel like I’m feeding a troll.

43

Graham is the only person in a position to make decisions about this. He has already discussed this issue multiple times in public, back in the early days of Inform. I am pretty sure he is now done having that conversation. I used to try to play a communications role and shed some more light on the development process, but I found that in practice it didn’t seem to matter how much information I shared: some people were always at least somewhat dissatisfied that the project wasn’t being run the way they wanted, and to reject the explanations they were offered if those explanations differed from their own preferences. So I also mostly stopped participating in broad-spectrum conversations about the development approach, and focused instead on improving communication about specific issues (bugs, feature requests, etc.).

However, it is not true that Graham is ignorant of other development styles, nor that the development team lacks people with experience on commercial software projects, both in game development and in other areas. For that matter, the expectation is that there will be a bugfix release of Inform on the order of weeks to address the issues brought in by the new build – so the distinction between “beta version on a public test server” and this release of Inform is more about how its status was communicated than about how the build will actually be handled in a practical sense. (I accept the point that status-communication is also important, if that’s where we’re about to go. I’m just saying that the distinction between one state and the other is not as vast as you’re making out.)

If I were still doing the broad-spectrum communication thing, I would talk about how this particular build of Inform, unlike most builds, required a dramatic rebuilding of text handling that did not lend itself to partial updates; about how input was solicited at several points, especially from translators and extension-builders, and iteration done around their feedback that would have been less focused had it involved a less experienced group; about the fact that there are a lot of people involved in creating a cross-platform release, which makes really frequent releases more difficult. I’d also say that there are many positive aspects of Inform that could never have been settled on by committee, that its idiosyncratic development style goes along with a unique and personal vision, and that it wouldn’t be nearly so interesting without the latter. But I’m admittedly very biased.

Here’s the part that I think really matters. An all-volunteer project has some considerations that a commercial software project does not: as a condition of its survival, it has to continue to run in a way that the participants enjoy enough to keep doing it. Since Graham is the most vital member of the team, “I don’t think I would enjoy running the project the way you suggest” is a compelling argument that effectively overrides all other possible arguments. I know Graham pretty well*, and I concur with his self-assessment that he is doing it in the way that best suits him and the other obligations in his life, and that he is also not obligated to the IF community to explain what those are as a condition of being allowed to do the project in his preferred way.

You can I suppose feel that he would really like your method if he tried it, etc., but at some point this becomes like trying to argue someone into liking cigars or sashimi. And for the argue-ee, repeatedly defending why you do not and never will like cigars starts to seem like a bad use of time that could be better spent on something you do like.

  • A major part of the reason Inform didn’t get out the door a year earlier is that our getting married and resolving our immigration visa issues turned out to be extremely time-consuming.
44

I notice bukayeva became a member in 2011. I myself became a member in this forum in 2009, was in RAIF for only a few months before that (methinks).

I don’t know whether bukayeva goes to the MUD, or participates in any off-forum discussion or real-life meetings (wish I could go to one…).

My point is: we newcomers probably have a different view on things because we might not know the people involed, or the community that once was. From what I understand, in the early nineties, and with the advent of Inform and Curses, with SpeedIFs, with all sorts of experimentation going on, with the IF scene being relatively limited and people getting to know each other pretty well…

…well, people knew each other. And each other’s way of doing things. I guess it seems natural that a newcomer might look at things (development schedules) in a light which no one had really thought of, and normally that would be a good thing, but in this case it also factors in the human element - or the lack, thereof, for more recent members. I said, jokingly, that Graham was a god-like invisible figure. I was half-joking, but that’s very much true. The zarfs, the Shorts, the Newcombs, the Eves, the Reeds, those are people in this community that I would love to be able to discuss IF with (if they weren’t so completely out of my league - sky high) (I also seem to descend into incoherence or contradiction when I really try to. I suppose I’ll just have to keep dreaming. Anyways, back to the topic). I can’t, but I can still read on their discussion and posts. You guys are all here, are all present.

Graham Nelson, for whom I have nothing but respect, admiration and awe, could be a pseudonym for an amalgam of authors and developers, for all I know. There is no human substance to the name, except for a winning, charming writing style. And I don’t frequent the circles where that name might HAVE some substance, or failing that where the spirit of the nineties-IF-scene lives on (I’m thinking very specifically about the MUD).

That might be why it’s easier for a newcomer to come up with great ideas to do it better, and why someone who’s been around longer has to explain, all over again, why things are optimal the way they are (and eventually get fed up with explaining altogether. I think, in fact, Short’s post above should be copy-pasted and made prominent so it can be easily quoted if the issue ever arises again).

45

Emily already stated this cogently enough but let me offer a case in point that I think backs up her example.

I created a tool called Lucid (github.com/jnyman/lucid). It’s a clone of the popular Cucumber tool which uses a constrained natural language structuring element (Gherkin) to allow people to use entirely unconstrained natural language to state requirements that are manual tests that can also be converted down to automated tests.

Lots and lots of people had (and still have) lots and lots of opinions of how I should do this or that feature. That’s cool and all. But you know what? I sort of had this vision for what I believed the product needed to be.

Lucid is an opinionated tool and it does support a high level of convention over configuration. Some people don’t like that at all. Yet Lucid also allows every convention to be overturned by configuration. Some people don’t like that at all either. However, it was necessary for the tool to be and allow just what I described. Necessary to whom? To me! I was the one that was driving it. That necessity, at least in my mind, required that I keep many things close to the vest, that I make myself the sole repository of design decisions (thus denying pull requests), that I incorporate features some thought were of the devil (sequenced steps), and that I not adhere to any specific release cycle that others may perceive as being better or more in line with the community. In my case, that would be the BDD test tool community.

Do I have a point here? Probably not. I rarely do. But if I did have one, I would probably try to tie what I just said above into a possible way of understanding how Graham might view his own particular creation.

Speaking to Peter’s last point, it is certainly true that long-standing members of a community can automatically assume that relative newcomers have all of the operative context around discussions that happened months or even years ago. Rarely is that the case, of course. This happens often with my own tools. I get people who come in and start talking about a lot of stuff and how it “should” be done, often seemingly assuming that these discussions had never taken place before and with the further assumption that I’m doing it the “wrong way” out of complete ignorance of all the better ways out there. Needless to say, said discussions had taken place and I was aware of my alternatives. I did periodically have to step back and realize not everyone knew that and figure out how to incorporate “newcomers” into a pre-existing context.