FYI to all players: The two “dishes” from Season One Episode 1 of Iron ChIF, produced by Challenger lpsmith and defending Iron Chef Inform 7 Afterward, have been posted and are available for online and offline play.
It was a real nailbiter finish this time, with both chefs working until only a few minutes were left on the clock. The vote for the winner will start in just 47 hours – don’t miss it!
There’s just a little over 12 hours to go until voting closes for this episode of Iron ChIF. The results will be revealed and the winner declared at noon UTC Tuesday March 3rd.
The archive with the games has been uploaded to the ifarchive, but I notice that there are two game files for “course correction”, one with just the name’s title and one with a filename that implies a later release - but both files are, in fact, the same release, serial 260227. They have the same filesize and everything. Was this intentional?
The difference is described (briefly) in the README file.
Challenger @lpsmith corrected a typo in the >CREDITS response (the 20260303 version), but for the sake of completeness the original version as seen in the competition is also included.
The files are the same size but should differ if the binaries are compared, and should show different responses to >CREDITS.
Ok, I was misled by them both showing the same serial number in response to ABOUT (EDIT - VERSION, actually), which they do… I just looked at the serial number and didn’t check CREDITS. I would have thought the serial number would have been different. Cheers.
They usually are, but the convention for parser IF serial numbers is that they record the date of compilation, so multiple builds in the same day will get the same number. It’s best to also increment the release number when that happens.
Right, that’s why I expected the second file to have the serial 260303. When it didn’t, I thought it was the same as the previous one.
Speaking as someone who just spent two weeks downloading everything they could (fabularium playable) from the archive and then from ifdb and then checking individual games and their external-available only latest versions to ensure I have the latest… this convention is good, and worth sticking to. 
Should I do a different release? It was just a single ‘l’ in Sarah’s name, so I thought I could sort of slip it in without causing problems. But if we need a whole ‘release 2’ for it, I can do that.
The date code in the episode archive reflects the date that I received the updated version. I probably should have used the suffix “-updated” instead of a date code. Sorry for any confusion caused!
At the moment, the IFDB page’s “download” and “play online” links are still pointing to the official competition version, without the updates; the typo-corrected version was included in the episode archive by request. I think @lpsmith is planning to release a post-competition version (with an incremented release number), and once that is ready and uploaded to IF Archive, the download and play links on the IFDB page can be changed to point to the post-comp release.
(I hope this post is still sufficiently on-topic. If it’s verging away too much, I pre-emptively apologise and would not object to it being split.)
For curiosity, and in a more general manner, do the “release” numbers have to be whole numbers? Could there be releases 1.01 for itsy bitsy tiny things like this? Or does Inform not like that idea?
The truth is, in this case, it’s simply inconsequential. But I would argue, hopefully constructively, that it’s by sticking to a few guidelines and conventions in small, inconsequential times that we ensure that they also get used in other times when it matters a lot more. It may seem a bit black-and-white - “has it been released? If so, then anything that comes next is the incrementally next release, or at least has some other obvious identifier, like a serial number that reflects a posterior date”. It’s kinda like a judge passing sentence, a phrase like that, and I see where it is off-putting.
I’m just saying, however, that these little things matter not at all to anyone shortly after the game is released, but in a few years’ time - and, unlike twenty years ago, when we could not care much because it was uncertain whether this hobby would ever get past its circle of hobbyists, we have found that IF insists on surviving and there are still actually people who go back to games of twenty years ago - what we will have is effectively two different releases, both internally indistinguishable except for one of them preserving an incorrect spelling of a tester. Let’s assume that that is the version which stands the test of time! Because it might happen! The version of the game with the typo may be the one preserved, partly because the other one has a release date in its filename that doesn’t match the serial. I assume that, Lucian, if I may address you directly, you would not be comfortable with the version with the typo being the one that survived.
My long-winded point is that even for small stuff it’s best, I find, to keep to these small things. Inconsequential as they are in the present, there is likely to come a time when it’s important. I had only the serial numbers to navigate between older game releases this past few weeks. There were a small handful of cases where the release number of the latest compiled version was actually inferior to the release number of a version that had been compiled with a previous-dated serial number! Having to choose, I chose the one with the latest serial number.
This was posted while I was typing, and yep, it pretty much overshadows everything and blots it over and resolves it because the post-comp version will have incremented release and serial number. Still, I have found that it’s best to keep to these conventions even in these cases. Things happen; post-comp versions may not materialise; authors may decide some changes are too small to be considered an incremented release. I can understand that! But when the players have more than one version available to them, they do need to be able to tell which is the latest; that is all. (having a filename saying “nameofthegame-updated” doesn’t help much in that regard, I’m afraid, because what if there is a version that was updated after that one? If I see a filename saying “nameofthegame” and “nameofthegame-updated” or “nameofthegame-date”, is the “nameofthegame” version the original, unupdated version? Or is it the latest, finalised version after all the updates? Anyway, filenames are easily altered…)
The Z-machine was invented well before semantic versioning, so it only has a single 16-bit field for release number—and as a result, the same is true for Glulx and Å-machine. I’ve tried to establish a convention for storing major and minor version in that field, but unfortunately it’s never caught on. 
You can use any release or version number you like, it just can’t be embedded into the storyfile’s metadata. But by all means print a more detailed version number in the banner etc.
It would be a little confusing, but you could increment the metadata version every release, so that release 15 would be version 2.4.