This topic seems to have two threads running in parallel. Let’s deal with them separately:
The idea with SpeedIF is that you are given a theme and a short time frame. You must write a game using that theme and finish it within that time frame. Mini-comps are exactly the same thing, except that the time frame tends to be a bit longer than SpeedIF. This principle is no different to modern-day game jams.
The theme can be a few keywords or an idea. For an example of keywords, the keywords could (hypothetically) be pumpkin, bobby pin, green and tarantula. You must write a game that incorporates those words (or concepts) in some way. The story cubes concept works very well for this.
For a theme, this could be just about anything. Past themes have included dinosaurs, Douglas Adams tribute, New Years Eve and so on.
Both approaches would work equally well for parser-based games and choice-based games, so no one would feel left out.
For the time frame, the very first SpeedIF organised by David A Cornelson in October 1998 gave you only 15 minutes. This was immediately extended to 30 minutes, then 1 hour and finally 2 hours. I think this is a ridiculously short time frame and authors clearly agreed, as there was only one entry.
The very first mini-comp organised by Lucian Paul Smith in May 1998 had a time frame of 4 weeks. I feel that two weeks is the shortest acceptable time frame. For recent examples, the first Adventuron game jam was two weeks and PunyJam #1 was 3 weeks. I entered both and, believe me, it was a real struggle to get something decent done in that time.
The whole point of ‘Cloak of Darkness’ is that it allows potential authors to compare different authoring systems when producing exactly the same game. The game itself is not a very good example, as it tests very few features of each system and, let’s be honest, the original game is a bit buggy. ‘Cloak of Darkness’ has already been done for all the major systems. Doing it again serves no purpose whatsoever.
So, let’s hear no more talk of ‘Cloak of Darkness’, unless the theme is to extend ‘Cloak of Darkness’ into a full-featured game. This could then include choice-based games.
Any game jam should be fun. If it’s not fun for both the author and the player, then don’t bother.
I think that PunyJam #1 was the perfect example of SpeedIF or mini-comp. It was promoted beforehand, but the theme was not revealed until a set time and date. The theme itself was very inspiring and resulted in a lot of very different games. The time frame was just about right. The itch.io platform worked very well, except for the voting, which was done offline.