CASA ::

Well, it depends. If Frantisek Fuka and František Fuka are actually the same person, I suggest relinking all entries of both to one and deleting the one no longer needed. It doesn't really matter whether you retain František or Frantisek. When the other is gone, you can rename the remaining one as ...

On the subject of authors, alongside those with question marks in their names I can also see "František Fuka" (https://solutionarchive.com/list/author%2C3870/) and "Jiři Koudelka" (https://solutionarchive.com/list/author%2C3868/).


The problem with Jiři Koudelka and František Fuka is that they ...

Alright, question marks in notes, synopsis and author names should be taken care of - all I found has been restored. If you find more broken things, I encourage you to fix by hand. If, however, you indeed find large amounts of broken entries, do alert me here. Thank you for all the support and the ...

Yup, the bit-level operation was accidentally run twice on a subset of some synopsis, notes and author names. At least I think this is the affected set. If you spot any issues on other fields, let me know. I have an automated fix for that as well, but I'm going slowly, doing a visual inspection of ...

Very good! I've updated the linked page. This now represents what I consider a complete set of fixes all across the website. I'll leave it for review in case anyone spots any mistakes which make it worse. I can pull the trigger any time.

I think I have an algorithmic solution, not relying on a whitelist of specific characters (thanks, Alastair, your math approach nudged me in the right direction).

Here is an overview of what it would do if I put it live: https://solutionarchive.com/__umlauts/ (temporary page, will be deleted again ...

Well, I can run a script over a number of database fields to re-encode characters. Though honestly, it will never be fully complete. You don't happen to have a backup from before this happened, do you?

Can you help assembling a list of characters which will need replacing - i.e. how they look now and what they are supposed to be? I have so far:

'ü' => 'ü',
'ä' => 'ä',
'ö' => 'ö',
'Ãœ' => 'Ü',
'é' => 'é',
'á' => 'á',
'ê' => 'ê'


Also, are other fields than the title also affected?

My guess is on a database upgrade gone wrong. I can check this afternoon.

My comrade LostInSpace's experience of playing Waxworks by Channel 8: https://www.goodolddays.net/en/game/Waxworks/

Well, then we can conclude: all the models I tried failed

The usual caveats of LLM usage obviously applies.

When I fed one of my own games, decompiled Z-Code, to the same models, some immediately returned a perfect and well annotated walkthrough. But when I quizzed them about alternate endings (which are in the game), they steadfastedly claimed there is ...

This one comes from Z.ai's "GLM":

Based on the game data provided in `QEGS.txt`, here is a complete step-by-step walkthrough to win the adventure.

### **Goal**
The objective is to collect the **Report** and the **Card** and deliver them to the **H.M.'s Office** (Room 60).

### **Phase 1: The ...

Mistral "Large" model (I quote the output as Markdown, since it is beautifully formatted):

# Text Adventure Solution
Created at 2026-02-28 15:29:50 UTC

> @ 2026-02-28 15:29:50 UTC
>
> Please provide a solution to the attached text adventure.

To complete the **QEGS text adventure**, you must ...

Nous Research "Hermes" model:

Based on the analysis of the QEGS text adventure game file, here is the solution:

Objective: Obtain both the report and the card to achieve a "Good Record" and complete the adventure.

Key Elements:

Report: Object number 14, weight 10
Card: Object number 29 ...